Receive a weekly summary and discussion of the top papers of the week by leading researchers in the field.

Pathology Pathology

A systematic evaluation of deep learning methods for the prediction of drug synergy in cancer.

In PLoS computational biology

One of the main obstacles to the successful treatment of cancer is the phenomenon of drug resistance. A common strategy to overcome resistance is the use of combination therapies. However, the space of possibilities is huge and efficient search strategies are required. Machine Learning (ML) can be a useful tool for the discovery of novel, clinically relevant anti-cancer drug combinations. In particular, deep learning (DL) has become a popular choice for modeling drug combination effects. Here, we set out to examine the impact of different methodological choices on the performance of multimodal DL-based drug synergy prediction methods, including the use of different input data types, preprocessing steps and model architectures. Focusing on the NCI ALMANAC dataset, we found that feature selection based on prior biological knowledge has a positive impact-limiting gene expression data to cancer or drug response-specific genes improved performance. Drug features appeared to be more predictive of drug response, with a 41% increase in coefficient of determination (R2) and 26% increase in Spearman correlation relative to a baseline model that used only cell line and drug identifiers. Molecular fingerprint-based drug representations performed slightly better than learned representations-ECFP4 fingerprints increased R2 by 5.3% and Spearman correlation by 2.8% w.r.t the best learned representations. In general, fully connected feature-encoding subnetworks outperformed other architectures. DL outperformed other ML methods by more than 35% (R2) and 14% (Spearman). Additionally, an ensemble combining the top DL and ML models improved performance by about 6.5% (R2) and 4% (Spearman). Using a state-of-the-art interpretability method, we showed that DL models can learn to associate drug and cell line features with drug response in a biologically meaningful way. The strategies explored in this study will help to improve the development of computational methods for the rational design of effective drug combinations for cancer therapy.

Baptista Delora, Ferreira Pedro G, Rocha Miguel

2023-Mar-23

Radiology Radiology

Interoperable slide microscopy viewer and annotation tool for imaging data science and computational pathology.

In Nature communications ; h5-index 260.0

The exchange of large and complex slide microscopy imaging data in biomedical research and pathology practice is impeded by a lack of data standardization and interoperability, which is detrimental to the reproducibility of scientific findings and clinical integration of technological innovations. We introduce Slim, an open-source, web-based slide microscopy viewer that implements the internationally accepted Digital Imaging and Communications in Medicine (DICOM) standard to achieve interoperability with a multitude of existing medical imaging systems. We showcase the capabilities of Slim as the slide microscopy viewer of the NCI Imaging Data Commons and demonstrate how the viewer enables interactive visualization of traditional brightfield microscopy and highly-multiplexed immunofluorescence microscopy images from The Cancer Genome Atlas and Human Tissue Atlas Network, respectively, using standard DICOMweb services. We further show how Slim enables the collection of standardized image annotations for the development or validation of machine learning models and the visual interpretation of model inference results in the form of segmentation masks, spatial heat maps, or image-derived measurements.

Gorman Chris, Punzo Davide, Octaviano Igor, Pieper Steven, Longabaugh William J R, Clunie David A, Kikinis Ron, Fedorov Andrey Y, Herrmann Markus D

2023-Mar-22

General General

Dense reinforcement learning for safety validation of autonomous vehicles.

In Nature ; h5-index 368.0

One critical bottleneck that impedes the development and deployment of autonomous vehicles is the prohibitively high economic and time costs required to validate their safety in a naturalistic driving environment, owing to the rarity of safety-critical events1. Here we report the development of an intelligent testing environment, where artificial-intelligence-based background agents are trained to validate the safety performances of autonomous vehicles in an accelerated mode, without loss of unbiasedness. From naturalistic driving data, the background agents learn what adversarial manoeuvre to execute through a dense deep-reinforcement-learning (D2RL) approach, in which Markov decision processes are edited by removing non-safety-critical states and reconnecting critical ones so that the information in the training data is densified. D2RL enables neural networks to learn from densified information with safety-critical events and achieves tasks that are intractable for traditional deep-reinforcement-learning approaches. We demonstrate the effectiveness of our approach by testing a highly automated vehicle in both highway and urban test tracks with an augmented-reality environment, combining simulated background vehicles with physical road infrastructure and a real autonomous test vehicle. Our results show that the D2RL-trained agents can accelerate the evaluation process by multiple orders of magnitude (103 to 105 times faster). In addition, D2RL will enable accelerated testing and training with other safety-critical autonomous systems.

Feng Shuo, Sun Haowei, Yan Xintao, Zhu Haojie, Zou Zhengxia, Shen Shengyin, Liu Henry X

2023-Mar

General General

Characterization of RNA polymerase II trigger loop mutations using molecular dynamics simulations and machine learning.

In PLoS computational biology

Catalysis and fidelity of multisubunit RNA polymerases rely on a highly conserved active site domain called the trigger loop (TL), which achieves roles in transcription through conformational changes and interaction with NTP substrates. The mutations of TL residues cause distinct effects on catalysis including hypo- and hyperactivity and altered fidelity. We applied molecular dynamics simulation (MD) and machine learning (ML) techniques to characterize TL mutations in the Saccharomyces cerevisiae RNA Polymerase II (Pol II) system. We did so to determine relationships between individual mutations and phenotypes and to associate phenotypes with MD simulated structural alterations. Using fitness values of mutants under various stress conditions, we modeled phenotypes along a spectrum of continual values. We found that ML could predict the phenotypes with 0.68 R2 correlation from amino acid sequences alone. It was more difficult to incorporate MD data to improve predictions from machine learning, presumably because MD data is too noisy and possibly incomplete to directly infer functional phenotypes. However, a variational auto-encoder model based on the MD data allowed the clustering of mutants with different phenotypes based on structural details. Overall, we found that a subset of loss-of-function (LOF) and lethal mutations tended to increase distances of TL residues to the NTP substrate, while another subset of LOF and lethal substitutions tended to confer an increase in distances between TL and bridge helix (BH). In contrast, some of the gain-of-function (GOF) mutants appear to cause disruption of hydrophobic contacts among TL and nearby helices.

Dutagaci Bercem, Duan Bingbing, Qiu Chenxi, Kaplan Craig D, Feig Michael

2023-Mar-22

Surgery Surgery

Comparison of Online-Onboard Adaptive Intensity-Modulated Radiation Therapy or Volumetric-Modulated Arc Radiotherapy With Image-Guided Radiotherapy for Patients With Gynecologic Tumors in Dependence on Fractionation and the Planning Target Volume Margin.

In JAMA network open

IMPORTANCE : Patients with newly diagnosed locally advanced cervical carcinomas or recurrences after surgery undergoing radiochemotherapy whose tumor is unsuited for a brachytherapy boost need high-dose percutaneous radiotherapy with small margins to compensate for clinical target volume deformations and set-up errors. Cone-beam computed tomography-based online adaptive radiotherapy (ART) has the potential to reduce planning target volume (PTV) margins below 5 mm for these tumors.

OBJECTIVE : To compare online ART technologies with image-guided radiotherapy (IGRT) for gynecologic tumors.

DESIGN, SETTING, AND PARTICIPANTS : This comparative effectiveness study comprised all 7 consecutive patients with gynecologic tumors who were treated with ART with artificial intelligence segmentation from January to May 2022 at the West German Cancer Center. All adapted treatment plans were reviewed for the new scenario of organs at risk and target volume. Dose distributions of adapted and scheduled plans optimized on the initial planning computed tomography scan were compared.

EXPOSURE : Online ART for gynecologic tumors.

MAIN OUTCOMES AND MEASURES : Target dose coverage with ART compared with IGRT for PTV margins of 5 mm or less in terms of the generalized equivalent uniform dose (gEUD) without increasing the gEUD for the organs at risk (bladder and rectum).

RESULTS : The first 10 treatment series among 7 patients (mean [SD] age, 65.7 [16.5] years) with gynecologic tumors from a prospective observational trial performed with ART were compared with IGRT. For a clinical PTV margin of 5 mm, IGRT was associated with a median gEUD decrease in the interfractional clinical target volume of -1.5% (90% CI, -31.8% to 2.9%) for all fractions in comparison with the planned dose distribution. Online ART was associated with a decrease of -0.02% (90% CI, -3.2% to 1.5%), which was less than the decrease with IGRT (P < .001). This was not associated with an increase in the gEUD for the bladder or rectum. For a PTV margin of 0 mm, the median gEUD deviation with IGRT was -13.1% (90% CI, -47.9% to 1.6%) compared with 0.1% (90% CI, -2.3% to 6.6%) with ART (P < .001). The benefit associated with ART was larger for a PTV margin of 0 mm than of 5 mm (P = .004) due to spreading of the cold spot at the clinical target volume margin from fraction to fraction with a median SD of 2.4 cm (90% CI, 1.9-3.4 cm) for all patients.

CONCLUSIONS AND RELEVANCE : This study suggests that ART is associated with an improvement in the percentage deviation of gEUD for the interfractional clinical target volume compared with IGRT. As the gain of ART depends on fractionation and PTV margin, a strategy is proposed here to switch from IGRT to ART, if the delivered gEUD distribution becomes unfavorable in comparison with the expected distribution during the course of treatment.

Guberina Maja, Santiago Garcia Alina, Khouya Aymane, Pöttgen Christoph, Holubyev Kostyantyn, Ringbaek Toke Printz, Lachmuth Manfred, Alberti Yasemin, Hoffmann Christian, Hlouschek Julian, Gauler Thomas, Lübcke Wolfgang, Indenkämpen Frank, Stuschke Martin, Guberina Nika

2023-Mar-01

oncology Oncology

An Unsupervised Machine Learning Approach to Evaluating the Association of Symptom Clusters With Adverse Outcomes Among Older Adults With Advanced Cancer: A Secondary Analysis of a Randomized Clinical Trial.

In JAMA network open

IMPORTANCE : Older adults with advanced cancer who have high pretreatment symptom severity often experience adverse events during cancer treatments. Unsupervised machine learning may help stratify patients into different risk groups.

OBJECTIVE : To evaluate whether clusters identified from baseline patient-reported symptom severity were associated with adverse outcomes.

DESIGN, SETTING, AND PARTICIPANTS : This secondary analysis of the Geriatric Assessment Intervention for Reducing Toxicity in Older Patients With Advanced Cancer (GAP70+) Trial (2014-2019) included patients who completed the National Cancer Institute Patient-Reported Outcomes version of the Common Terminology Criteria for Adverse Events (PRO-CTCAE) before starting a new cancer treatment regimen and received care at community oncology sites across the United States. An unsupervised machine learning algorithm (k-means with Euclidean distance) clustered patients based on similarities of baseline symptom severities. Clustering variables included severity items of 24 PRO-CTCAE symptoms (range, 0-4; corresponding to none, mild, moderate, severe, and very severe). Total severity score was calculated as the sum of 24 items (range, 0-96). Whether the clusters were associated with unplanned hospitalization, death, and toxic effects was then examined. Analyses were conducted in January and February 2022.

EXPOSURES : Symptom severity.

MAIN OUTCOMES AND MEASURES : Unplanned hospitalization over 3 months (primary), all-cause mortality over 1 year, and any clinician-rated grade 3 to 5 toxic effect over 3 months.

RESULTS : Of 718 enrolled patients, 706 completed baseline PRO-CTCAE and were included (mean [SD] age, 77.2 [5.5] years, 401 [56.8%] male patients; 51 [7.2%] Black and 619 [87.8%] non-Hispanic White patients; 245 [34.7%] with gastrointestinal cancer; 175 [24.8%] with lung cancer; mean [SD] impaired Geriatric Assessment domains, 4.5 [1.6]). The algorithm classified 310 (43.9%), 295 (41.8%), and 101 (14.3%) into low-, medium-, and high-severity clusters (within-cluster mean [SD] severity scores: low, 6.3 [3.4]; moderate, 16.6 [4.3]; high, 29.8 [7.8]; P < .001). Controlling for sociodemographic variables, clinical factors, study group, and practice site, compared with patients in the low-severity cluster, those in the moderate-severity cluster were more likely to experience hospitalization (risk ratio, 1.36; 95% CI, 1.01-1.84; P = .046). Moderate- and high-severity clusters were associated with a higher risk of death (moderate: hazard ratio, 1.31; 95% CI, 1.01-1.69; P = .04; high: hazard ratio, 2.00; 95% CI, 1.43-2.78; P < .001), but not toxic effects.

CONCLUSIONS AND RELEVANCE : In this study, unsupervised machine learning partitioned patients into distinct symptom severity clusters; patients with higher pretreatment severity were more likely to experience hospitalization and death.

TRIAL REGISTRATION : ClinicalTrials.gov Identifier: NCT02054741.

Xu Huiwen, Mohamed Mostafa, Flannery Marie, Peppone Luke, Ramsdale Erika, Loh Kah Poh, Wells Megan, Jamieson Leah, Vogel Victor G, Hall Bianca Alexandra, Mustian Karen, Mohile Supriya, Culakova Eva

2023-Mar-01

General General

A self-adaptive hardware with resistive switching synapses for experience-based neurocomputing.

In Nature communications ; h5-index 260.0

Neurobiological systems continually interact with the surrounding environment to refine their behaviour toward the best possible reward. Achieving such learning by experience is one of the main challenges of artificial intelligence, but currently it is hindered by the lack of hardware capable of plastic adaptation. Here, we propose a bio-inspired recurrent neural network, mastered by a digital system on chip with resistive-switching synaptic arrays of memory devices, which exploits homeostatic Hebbian learning for improved efficiency. All the results are discussed experimentally and theoretically, proposing a conceptual framework for benchmarking the main outcomes in terms of accuracy and resilience. To test the proposed architecture for reinforcement learning tasks, we study the autonomous exploration of continually evolving environments and verify the results for the Mars rover navigation. We also show that, compared to conventional deep learning techniques, our in-memory hardware has the potential to achieve a significant boost in speed and power-saving.

Bianchi S, Muñoz-Martin I, Covi E, Bricalli A, Piccolboni G, Regev A, Molas G, Nodin J F, Andrieu F, Ielmini D

2023-Mar-21

General General

Accelerating network layouts using graph neural networks.

In Nature communications ; h5-index 260.0

Graph layout algorithms used in network visualization represent the first and the most widely used tool to unveil the inner structure and the behavior of complex networks. Current network visualization software relies on the force-directed layout (FDL) algorithm, whose high computational complexity makes the visualization of large real networks computationally prohibitive and traps large graphs into high energy configurations, resulting in hard-to-interpret "hairball" layouts. Here we use Graph Neural Networks (GNN) to accelerate FDL, showing that deep learning can address both limitations of FDL: it offers a 10 to 100 fold improvement in speed while also yielding layouts which are more informative. We analytically derive the speedup offered by GNN, relating it to the number of outliers in the eigenspectrum of the adjacency matrix, predicting that GNNs are particularly effective for networks with communities and local regularities. Finally, we use GNN to generate a three-dimensional layout of the Internet, and introduce additional measures to assess the layout quality and its interpretability, exploring the algorithm's ability to separate communities and the link-length distribution. The novel use of deep neural networks can help accelerate other network-based optimization problems as well, with applications from reaction-diffusion systems to epidemics.

Both Csaba, Dehmamy Nima, Yu Rose, Barabási Albert-László

2023-Mar-21

Cardiology Cardiology

Clinical and genetic associations of deep learning-derived cardiac magnetic resonance-based left ventricular mass.

In Nature communications ; h5-index 260.0

Left ventricular mass is a risk marker for cardiovascular events, and may indicate an underlying cardiomyopathy. Cardiac magnetic resonance is the gold-standard for left ventricular mass estimation, but is challenging to obtain at scale. Here, we use deep learning to enable genome-wide association study of cardiac magnetic resonance-derived left ventricular mass indexed to body surface area within 43,230 UK Biobank participants. We identify 12 genome-wide associations (1 known at TTN and 11 novel for left ventricular mass), implicating genes previously associated with cardiac contractility and cardiomyopathy. Cardiac magnetic resonance-derived indexed left ventricular mass is associated with incident dilated and hypertrophic cardiomyopathies, and implantable cardioverter-defibrillator implant. An indexed left ventricular mass polygenic risk score ≥90th percentile is also associated with incident implantable cardioverter-defibrillator implant in separate UK Biobank (hazard ratio 1.22, 95% CI 1.05-1.44) and Mass General Brigham (hazard ratio 1.75, 95% CI 1.12-2.74) samples. Here, we perform a genome-wide association study of cardiac magnetic resonance-derived indexed left ventricular mass to identify 11 novel variants and demonstrate that cardiac magnetic resonance-derived and genetically predicted indexed left ventricular mass are associated with incident cardiomyopathy.

Khurshid Shaan, Lazarte Julieta, Pirruccello James P, Weng Lu-Chen, Choi Seung Hoan, Hall Amelia W, Wang Xin, Friedman Samuel F, Nauffal Victor, Biddinger Kiran J, Aragam Krishna G, Batra Puneet, Ho Jennifer E, Philippakis Anthony A, Ellinor Patrick T, Lubitz Steven A

2023-Mar-21

Public Health Public Health

Multiomic signatures of body mass index identify heterogeneous health phenotypes and responses to a lifestyle intervention.

In Nature medicine ; h5-index 170.0

Multiomic profiling can reveal population heterogeneity for both health and disease states. Obesity drives a myriad of metabolic perturbations and is a risk factor for multiple chronic diseases. Here we report an atlas of cross-sectional and longitudinal changes in 1,111 blood analytes associated with variation in body mass index (BMI), as well as multiomic associations with host polygenic risk scores and gut microbiome composition, from a cohort of 1,277 individuals enrolled in a wellness program (Arivale). Machine learning model predictions of BMI from blood multiomics captured heterogeneous phenotypic states of host metabolism and gut microbiome composition better than BMI, which was also validated in an external cohort (TwinsUK). Moreover, longitudinal analyses identified variable BMI trajectories for different omics measures in response to a healthy lifestyle intervention; metabolomics-inferred BMI decreased to a greater extent than actual BMI, whereas proteomics-inferred BMI exhibited greater resistance to change. Our analyses further identified blood analyte-analyte associations that were modified by metabolomics-inferred BMI and partially reversed in individuals with metabolic obesity during the intervention. Taken together, our findings provide a blood atlas of the molecular perturbations associated with changes in obesity status, serving as a resource to quantify metabolic health for predictive and preventive medicine.

Watanabe Kengo, Wilmanski Tomasz, Diener Christian, Earls John C, Zimmer Anat, Lincoln Briana, Hadlock Jennifer J, Lovejoy Jennifer C, Gibbons Sean M, Magis Andrew T, Hood Leroy, Price Nathan D, Rappaport Noa

2023-Mar-20

Radiology Radiology

AI Cardiac MRI Scar Analysis Aids Prediction of Major Arrhythmic Events in the Multicenter DERIVATE Registry.

In Radiology ; h5-index 91.0

Background Scar burden with late gadolinium enhancement (LGE) cardiac MRI (CMR) predicts arrhythmic events in patients with postinfarction in single-center studies. However, LGE analysis requires experienced human observers, is time consuming, and introduces variability. Purpose To test whether postinfarct scar with LGE CMR can be quantified fully automatically by machines and to compare the ability of LGE CMR scar analyzed by humans and machines to predict arrhythmic events. Materials and Methods This study is a retrospective analysis of the multicenter, multivendor CarDiac MagnEtic Resonance for Primary Prevention Implantable CardioVerter DebrillAtor ThErapy (DERIVATE) registry. Patients with chronic heart failure, echocardiographic left ventricular ejection fraction (LVEF) of less than 50%, and LGE CMR were recruited (from January 2015 through December 2020). In the current study, only patients with ischemic cardiomyopathy were included. Quantification of total, dense, and nondense scars was carried out by two experienced readers or a Ternaus network, trained and tested with LGE images of 515 and 246 patients, respectively. Univariable and multivariable Cox analyses were used to assess patient and cardiac characteristics associated with a major adverse cardiac event (MACE). Area under the receiver operating characteristic curve (AUC) was used to compare model performances. Results In 761 patients (mean age, 65 years ± 11, 671 men), 83 MACEs occurred. With use of the testing group, univariable Cox-analysis found New York Heart Association class, left ventricle volume and/or function parameters (by echocardiography or CMR), guideline criterion (LVEF of ≤35% and New York Heart Association class II or III), and LGE scar analyzed by humans or the machine-learning algorithm as predictors of MACE. Machine-based dense or total scar conferred incremental value over the guideline criterion for the association with MACE (AUC: 0.68 vs 0.63, P = .02 and AUC: 0.67 vs 0.63, P = .01, respectively). Modeling with competing risks yielded for dense and total scar (AUC: 0.67 vs 0.61, P = .01 and AUC: 0.66 vs 0.61, P = .005, respectively). Conclusion In this analysis of the multicenter CarDiac MagnEtic Resonance for Primary Prevention Implantable CardioVerter DebrillAtor ThErapy (DERIVATE) registry, fully automatic machine learning-based late gadolinium enhancement analysis reliably quantifies myocardial scar mass and improves the current prediction model that uses guideline-based risk criteria for implantable cardioverter defibrillator implantation. ClinicalTrials.gov registration no.: NCT03352648 Published under a CC BY 4.0 license. Supplemental material is available for this article.

Ghanbari Fahime, Joyce Thomas, Lorenzoni Valentina, Guaricci Andrea I, Pavon Anna-Giulia, Fusini Laura, Andreini Daniele, Rabbat Mark G, Aquaro Giovanni Donato, Abete Raffaele, Bogaert Jan, Camastra Giovanni, Carigi Samuela, Carrabba Nazario, Casavecchia Grazia, Censi Stefano, Cicala Gloria, De Cecco Carlo N, De Lazzari Manuel, Di Giovine Gabriella, Di Roma Mauro, Focardi Marta, Gaibazzi Nicola, Gismondi Annalaura, Gravina Matteo, Lanzillo Chiara, Lombardi Massimo, Lozano-Torres Jordi, Masi Ambra, Moro Claudio, Muscogiuri Giuseppe, Nese Alberto, Pradella Silvia, Sbarbati Stefano, Schoepf U Joseph, Valentini Adele, Crelier Gérard, Masci Pier Giorgio, Pontone Gianluca, Kozerke Sebastian, Schwitter Juerg

2023-Mar-21

General General

Metheor: Ultrafast DNA methylation heterogeneity calculation from bisulfite read alignments.

In PLoS computational biology

Phased DNA methylation states within bisulfite sequencing reads are valuable source of information that can be used to estimate epigenetic diversity across cells as well as epigenomic instability in individual cells. Various measures capturing the heterogeneity of DNA methylation states have been proposed for a decade. However, in routine analyses on DNA methylation, this heterogeneity is often ignored by computing average methylation levels at CpG sites, even though such information exists in bisulfite sequencing data in the form of phased methylation states, or methylation patterns. In this study, to facilitate the application of the DNA methylation heterogeneity measures in downstream epigenomic analyses, we present a Rust-based, extremely fast and lightweight bioinformatics toolkit called Metheor. As the analysis of DNA methylation heterogeneity requires the examination of pairs or groups of CpGs throughout the genome, existing softwares suffer from high computational burden, which almost make a large-scale DNA methylation heterogeneity studies intractable for researchers with limited resources. In this study, we benchmark the performance of Metheor against existing code implementations for DNA methylation heterogeneity measures in three different scenarios of simulated bisulfite sequencing datasets. Metheor was shown to dramatically reduce the execution time up to 300-fold and memory footprint up to 60-fold, while producing identical results with the original implementation, thereby facilitating a large-scale study of DNA methylation heterogeneity profiles. To demonstrate the utility of the low computational burden of Metheor, we show that the methylation heterogeneity profiles of 928 cancer cell lines can be computed with standard computing resources. With those profiles, we reveal the association between DNA methylation heterogeneity and various omics features. Source code for Metheor is at https://github.com/dohlee/metheor and is freely available under the GPL-3.0 license.

Lee Dohoon, Koo Bonil, Yang Jeewon, Kim Sun

2023-Mar-20

Public Health Public Health

PCR-like performance of rapid test with permselective tunable nanotrap.

In Nature communications ; h5-index 260.0

Highly sensitive rapid testing for COVID-19 is essential for minimizing virus transmission, especially before the onset of symptoms and in asymptomatic cases. Here, we report bioengineered enrichment tools for lateral flow assays (LFAs) with enhanced sensitivity and specificity (BEETLES2), achieving enrichment of SARS-CoV-2 viruses, nucleocapsid (N) proteins and immunoglobulin G (IgG) with 3-minute operation. The limit of detection is improved up to 20-fold. We apply this method to clinical samples, including 83% with either intermediate (35%) or low viral loads (48%), collected from 62 individuals (n = 42 for positive and n = 20 for healthy controls). We observe diagnostic sensitivity, specificity, and accuracy of 88.1%, 100%, and 91.9%, respectively, compared with commercial LFAs alone achieving 14.29%, 100%, and 41.94%, respectively. BEETLES2, with permselectivity and tunability, can enrich the SARS-CoV-2 virus, N proteins, and IgG in the nasopharyngeal/oropharyngeal swab, saliva, and blood serum, enabling reliable and sensitive point-of-care testing, facilitating fast early diagnosis.

Park Seong Jun, Lee Seungmin, Lee Dongtak, Lee Na Eun, Park Jeong Soo, Hong Ji Hye, Jang Jae Won, Kim Hyunji, Roh Seokbeom, Lee Gyudo, Lee Dongho, Cho Sung-Yeon, Park Chulmin, Lee Dong-Gun, Lee Raeseok, Nho Dukhee, Yoon Dae Sung, Yoo Yong Kyoung, Lee Jeong Hoon

2023-Mar-18

Public Health Public Health

PCR-like performance of rapid test with permselective tunable nanotrap.

In Nature communications ; h5-index 260.0

Highly sensitive rapid testing for COVID-19 is essential for minimizing virus transmission, especially before the onset of symptoms and in asymptomatic cases. Here, we report bioengineered enrichment tools for lateral flow assays (LFAs) with enhanced sensitivity and specificity (BEETLES2), achieving enrichment of SARS-CoV-2 viruses, nucleocapsid (N) proteins and immunoglobulin G (IgG) with 3-minute operation. The limit of detection is improved up to 20-fold. We apply this method to clinical samples, including 83% with either intermediate (35%) or low viral loads (48%), collected from 62 individuals (n = 42 for positive and n = 20 for healthy controls). We observe diagnostic sensitivity, specificity, and accuracy of 88.1%, 100%, and 91.9%, respectively, compared with commercial LFAs alone achieving 14.29%, 100%, and 41.94%, respectively. BEETLES2, with permselectivity and tunability, can enrich the SARS-CoV-2 virus, N proteins, and IgG in the nasopharyngeal/oropharyngeal swab, saliva, and blood serum, enabling reliable and sensitive point-of-care testing, facilitating fast early diagnosis.

Park Seong Jun, Lee Seungmin, Lee Dongtak, Lee Na Eun, Park Jeong Soo, Hong Ji Hye, Jang Jae Won, Kim Hyunji, Roh Seokbeom, Lee Gyudo, Lee Dongho, Cho Sung-Yeon, Park Chulmin, Lee Dong-Gun, Lee Raeseok, Nho Dukhee, Yoon Dae Sung, Yoo Yong Kyoung, Lee Jeong Hoon

2023-Mar-18

Radiology Radiology

Recalibration of a Deep Learning Model for Low-Dose Computed Tomographic Images to Inform Lung Cancer Screening Intervals.

In JAMA network open

IMPORTANCE : Annual low-dose computed tomographic (LDCT) screening reduces lung cancer mortality, but harms could be reduced and cost-effectiveness improved by reusing the LDCT image in conjunction with deep learning or statistical models to identify low-risk individuals for biennial screening.

OBJECTIVE : To identify low-risk individuals in the National Lung Screening Trial (NLST) and estimate, had they been assigned a biennial screening, how many lung cancers would have been delayed 1 year in diagnosis.

DESIGN, SETTING, AND PARTICIPANTS : This diagnostic study included participants with a presumed nonmalignant lung nodule in the NLST between January 1, 2002, and December 31, 2004, with follow-up completed on December 31, 2009. Data were analyzed for this study from September 11, 2019, to March 15, 2022.

EXPOSURES : An externally validated deep learning algorithm that predicts malignancy in current lung nodules using LDCT images (Lung Cancer Prediction Convolutional Neural Network [LCP-CNN]; Optellum Ltd) was recalibrated to predict 1-year lung cancer detection by LDCT for presumed nonmalignant nodules. Individuals with presumed nonmalignant lung nodules were hypothetically assigned annual vs biennial screening based on the recalibrated LCP-CNN model, Lung Cancer Risk Assessment Tool (LCRAT + CT [a statistical model combining individual risk factors and LDCT image features]), and the American College of Radiology recommendations for lung nodules, version 1.1 (Lung-RADS).

MAIN OUTCOMES AND MEASURES : Primary outcomes included model prediction performance, the absolute risk of a 1-year delay in cancer diagnosis, and the proportion of people without lung cancer assigned a biennial screening interval vs the proportion of cancer diagnoses delayed.

RESULTS : The study included 10 831 LDCT images from patients with presumed nonmalignant lung nodules (58.7% men; mean [SD] age, 61.9 [5.0] years), of whom 195 were diagnosed with lung cancer from the subsequent screen. The recalibrated LCP-CNN had substantially higher area under the curve (0.87) than LCRAT + CT (0.79) or Lung-RADS (0.69) to predict 1-year lung cancer risk (P < .001). If 66% of screens with nodules were assigned to biennial screening, the absolute risk of a 1-year delay in cancer diagnosis would have been lower for recalibrated LCP-CNN (0.28%) than LCRAT + CT (0.60%; P = .001) or Lung-RADS (0.97%; P < .001). To delay only 10% of cancer diagnoses at 1 year, more people would have been safely assigned biennial screening under LCP-CNN than LCRAT + CT (66.4% vs 40.3%; P < .001).

CONCLUSIONS AND RELEVANCE : In this diagnostic study evaluating models of lung cancer risk, a recalibrated deep learning algorithm was most predictive of 1-year lung cancer risk and had least risk of 1-year delay in cancer diagnosis among people assigned biennial screening. Deep learning algorithms could prioritize people for workup of suspicious nodules and decrease screening intensity for people with low-risk nodules, which may be vital for implementation in health care systems.

Landy Rebecca, Wang Vivian L, Baldwin David R, Pinsky Paul F, Cheung Li C, Castle Philip E, Skarzynski Martin, Robbins Hilary A, Katki Hormuzd A

2023-Mar-01

oncology Oncology

A longitudinal circulating tumor DNA-based model associated with survival in metastatic non-small-cell lung cancer.

In Nature medicine ; h5-index 170.0

One of the great challenges in therapeutic oncology is determining who might achieve survival benefits from a particular therapy. Studies on longitudinal circulating tumor DNA (ctDNA) dynamics for the prediction of survival have generally been small or nonrandomized. We assessed ctDNA across 5 time points in 466 non-small-cell lung cancer (NSCLC) patients from the randomized phase 3 IMpower150 study comparing chemotherapy-immune checkpoint inhibitor (chemo-ICI) combinations and used machine learning to jointly model multiple ctDNA metrics to predict overall survival (OS). ctDNA assessments through cycle 3 day 1 of treatment enabled risk stratification of patients with stable disease (hazard ratio (HR) = 3.2 (2.0-5.3), P < 0.001; median 7.1 versus 22.3 months for high- versus low-intermediate risk) and with partial response (HR = 3.3 (1.7-6.4), P < 0.001; median 8.8 versus 28.6 months). The model also identified high-risk patients in an external validation cohort from the randomized phase 3 OAK study of ICI versus chemo in NSCLC (OS HR = 3.73 (1.83-7.60), P = 0.00012). Simulations of clinical trial scenarios employing our ctDNA model suggested that early ctDNA testing outperforms early radiographic imaging for predicting trial outcomes. Overall, measuring ctDNA dynamics during treatment can improve patient risk stratification and may allow early differentiation between competing therapies during clinical trials.

Assaf Zoe June F, Zou Wei, Fine Alexander D, Socinski Mark A, Young Amanda, Lipson Doron, Freidin Jonathan F, Kennedy Mark, Polisecki Eliana, Nishio Makoto, Fabrizio David, Oxnard Geoffrey R, Cummings Craig, Rode Anja, Reck Martin, Patil Namrata S, Lee Mark, Shames David S, Schulze Katja

2023-Mar-16

General General

Heterogeneous intercalated metal-organic framework active materials for fast-charging non-aqueous Li-ion capacitors.

In Nature communications ; h5-index 260.0

Intercalated metal-organic frameworks (iMOFs) based on aromatic dicarboxylate are appealing negative electrode active materials for Li-based electrochemical energy storage devices. They store Li ions at approximately 0.8 V vs. Li/Li+ and, thus, avoid Li metal plating during cell operation. However, their fast-charging capability is limited. Here, to circumvent this issue, we propose iMOFs with multi-aromatic units selected using machine learning and synthesized via solution spray drying. A naphthalene-based multivariate material with nanometric thickness allows the reversible storage of Li-ions in non-aqueous Li metal cell configuration reaching 85% capacity retention at 400 mA g-1 (i.e., 30 min for full charge) and 20 °C compared to cycling at 20 mA g-1 (i.e., 10 h for full charge). The same material, tested in combination with an activated carbon-based positive electrode, enables a discharge capacity retention of about 91% after 1000 cycles at 0.15 mA cm-2 (i.e., 2 h for full charge) and 20 °C. We elucidate the charge storage mechanism and demonstrate that during Li intercalation, the distorted crystal structure promotes electron delocalization by controlling the frame vibration. As a result, a phase transition suppresses phase separation, thus, benefitting the electrode's fast charging behavior.

Ogihara Nobuhiro, Hasegawa Masaki, Kumagai Hitoshi, Mikita Riho, Nagasako Naoyuki

2023-Mar-16

Ophthalmology Ophthalmology

Visual electrophysiology and "the potential of the potentials".

In Eye (London, England) ; h5-index 41.0

Visual electrophysiology affords direct, quantitative, objective assessment of visual pathway function at different levels, and thus yields information complementary to, and not necessarily obtainable from, imaging or psychophysical testing. The tests available, and their indications, have evolved, with many advances, both in technology and in our understanding of the neural basis of the waveforms, now facilitating more precise evaluation of physiology and pathophysiology. After summarising the visual pathway and current standard clinical testing methods, this review discusses, non-exhaustively, several developments, focusing particularly on human electroretinogram recordings. These include new devices (portable, non-mydiatric, multimodal), novel testing protocols (including those aiming to separate rod-driven and cone-driven responses, and to monitor retinal adaptation), and developments in methods of analysis, including use of modelling and machine learning. It is likely that several tests will become more accessible and useful in both clinical and research settings. In future, these methods will further aid our understanding of common and rare eye disease, will help in assessing novel therapies, and will potentially yield information relevant to neurological and neuro-psychiatric conditions.

Mahroo Omar A

2023-Mar-16

Surgery Surgery

Artificial Intelligence for Hip Fracture Detection and Outcome Prediction: A Systematic Review and Meta-analysis.

In JAMA network open

IMPORTANCE : Artificial intelligence (AI) enables powerful models for establishment of clinical diagnostic and prognostic tools for hip fractures; however the performance and potential impact of these newly developed algorithms are currently unknown.

OBJECTIVE : To evaluate the performance of AI algorithms designed to diagnose hip fractures on radiographs and predict postoperative clinical outcomes following hip fracture surgery relative to current practices.

DATA SOURCES : A systematic review of the literature was performed using the MEDLINE, Embase, and Cochrane Library databases for all articles published from database inception to January 23, 2023. A manual reference search of included articles was also undertaken to identify any additional relevant articles.

STUDY SELECTION : Studies developing machine learning (ML) models for the diagnosis of hip fractures from hip or pelvic radiographs or to predict any postoperative patient outcome following hip fracture surgery were included.

DATA EXTRACTION AND SYNTHESIS : This study followed the Preferred Reporting Items for Systematic Reviews and Meta-analyses and was registered with PROSPERO. Eligible full-text articles were evaluated and relevant data extracted independently using a template data extraction form. For studies that predicted postoperative outcomes, the performance of traditional predictive statistical models, either multivariable logistic or linear regression, was recorded and compared with the performance of the best ML model on the same out-of-sample data set.

MAIN OUTCOMES AND MEASURES : Diagnostic accuracy of AI models was compared with the diagnostic accuracy of expert clinicians using odds ratios (ORs) with 95% CIs. Areas under the curve for postoperative outcome prediction between traditional statistical models (multivariable linear or logistic regression) and ML models were compared.

RESULTS : Of 39 studies that met all criteria and were included in this analysis, 18 (46.2%) used AI models to diagnose hip fractures on plain radiographs and 21 (53.8%) used AI models to predict patient outcomes following hip fracture surgery. A total of 39 598 plain radiographs and 714 939 hip fractures were used for training, validating, and testing ML models specific to diagnosis and postoperative outcome prediction, respectively. Mortality and length of hospital stay were the most predicted outcomes. On pooled data analysis, compared with clinicians, the OR for diagnostic error of ML models was 0.79 (95% CI, 0.48-1.31; P = .36; I2 = 60%) for hip fracture radiographs. For the ML models, the mean (SD) sensitivity was 89.3% (8.5%), specificity was 87.5% (9.9%), and F1 score was 0.90 (0.06). The mean area under the curve for mortality prediction was 0.84 with ML models compared with 0.79 for alternative controls (P = .09).

CONCLUSIONS AND RELEVANCE : The findings of this systematic review and meta-analysis suggest that the potential applications of AI to aid with diagnosis from hip radiographs are promising. The performance of AI in diagnosing hip fractures was comparable with that of expert radiologists and surgeons. However, current implementations of AI for outcome prediction do not seem to provide substantial benefit over traditional multivariable predictive statistics.

Lex Johnathan R, Di Michele Joseph, Koucheki Robert, Pincus Daniel, Whyne Cari, Ravi Bheeshma

2023-Mar-01

Public Health Public Health

Development of a Machine Learning Model to Estimate US Firearm Homicides in Near Real Time.

In JAMA network open

IMPORTANCE : Firearm homicides are a major public health concern; lack of timely mortality data presents considerable challenges to effective response. Near real-time data sources offer potential for more timely estimation of firearm homicides.

OBJECTIVE : To estimate near real-time burden of weekly and annual firearm homicides in the US.

DESIGN, SETTING, AND PARTICIPANTS : In this prognostic study, anonymous, longitudinal time series data were obtained from multiple data sources, including Google and YouTube search trends related to firearms (2014-2019), emergency department visits for firearm injuries (National Syndromic Surveillance Program, 2014-2019), emergency medical service activations for firearm-related injuries (biospatial, 2014-2019), and National Domestic Violence Hotline contacts flagged with the keyword firearm (2016-2019). Data analysis was performed from September 2021 to September 2022.

MAIN OUTCOMES AND MEASURES : Weekly estimates of US firearm homicides were calculated using a 2-phase pipeline, first fitting optimal machine learning models for each data stream and then combining the best individual models into a stacked ensemble model. Model accuracy was assessed by comparing predictions of firearm homicides in 2019 to actual firearm homicides identified by National Vital Statistics System death certificates. Results were also compared with a SARIMA (seasonal autoregressive integrated moving average) model, a common method to forecast injury mortality.

RESULTS : Both individual and ensemble models yielded highly accurate estimates of firearm homicides. Individual models' mean error for weekly estimates of firearm homicides (root mean square error) varied from 24.95 for emergency department visits to 31.29 for SARIMA forecasting. Ensemble models combining data sources had lower weekly mean error and higher annual accuracy than individual data sources: the all-source ensemble model had a weekly root mean square error of 24.46 deaths and full-year accuracy of 99.74%, predicting the total number of firearm homicides in 2019 within 38 deaths for the entire year (compared with 95.48% accuracy and 652 deaths for the SARIMA model). The model decreased the time lag of reporting weekly firearm homicides from 7 to 8 months to approximately 6 weeks.

CONCLUSIONS AND RELEVANCE : In this prognostic study of diverse secondary data on machine learning, ensemble modeling produced accurate near real-time estimates of weekly and annual firearm homicides and substantially decreased data source time lags. Ensemble model forecasts can accelerate public health practitioners' and policy makers' ability to respond to unanticipated shifts in firearm homicides.

Swedo Elizabeth A, Alic Alen, Law Royal K, Sumner Steven A, Chen May S, Zwald Marissa L, Van Dyke Miriam E, Bowen Daniel A, Mercy James A

2023-Mar-01

General General

Machine Learning-Based Prediction of Attention-Deficit/Hyperactivity Disorder and Sleep Problems With Wearable Data in Children.

In JAMA network open

IMPORTANCE : Early detection of attention-deficit/hyperactivity disorder (ADHD) and sleep problems is paramount for children's mental health. Interview-based diagnostic approaches have drawbacks, necessitating the development of an evaluation method that uses digital phenotypes in daily life.

OBJECTIVE : To evaluate the predictive performance of machine learning (ML) models by setting the data obtained from personal digital devices comprising training features (ie, wearable data) and diagnostic results of ADHD and sleep problems by the Kiddie Schedule for Affective Disorders and Schizophrenia Present and Lifetime Version for Diagnostic and Statistical Manual of Mental Disorders, 5th edition (K-SADS) as a prediction class from the Adolescent Brain Cognitive Development (ABCD) study.

DESIGN, SETTING, AND PARTICIPANTS : In this diagnostic study, wearable data and K-SADS data were collected at 21 sites in the US in the ABCD study (release 3.0, November 2, 2020, analyzed October 11, 2021). Screening data from 6571 patients and 21 days of wearable data from 5725 patients collected at the 2-year follow-up were used, and circadian rhythm-based features were generated for each participant. A total of 12 348 wearable data for ADHD and 39 160 for sleep problems were merged for developing ML models.

MAIN OUTCOMES AND MEASURES : The average performance of the ML models was measured using an area under the receiver operating characteristics curve (AUC), sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV). In addition, the Shapley Additive Explanations value was used to calculate the importance of features.

RESULTS : The final population consisted of 79 children with ADHD problems (mean [SD] age, 144.5 [8.1] months; 55 [69.6%] males) vs 1011 controls and 68 with sleep problems (mean [SD] age, 143.5 [7.5] months; 38 [55.9%] males) vs 3346 controls. The ML models showed reasonable predictive performance for ADHD (AUC, 0.798; sensitivity, 0.756; specificity, 0.716; PPV, 0.159; and NPV, 0.976) and sleep problems (AUC, 0.737; sensitivity, 0.743; specificity, 0.632; PPV, 0.036; and NPV, 0.992).

CONCLUSIONS AND RELEVANCE : In this diagnostic study, an ML method for early detection or screening using digital phenotypes in children's daily lives was developed. The results support facilitating early detection in children; however, additional follow-up studies can improve its performance.

Kim Won-Pyo, Kim Hyun-Jin, Pack Seung Pil, Lim Jae-Hyun, Cho Chul-Hyun, Lee Heon-Jeong

2023-Mar-01

General General

Attention is not all you need: the complicated case of ethically using large language models in healthcare and medicine.

In EBioMedicine

Large Language Models (LLMs) are a key component of generative artificial intelligence (AI) applications for creating new content including text, imagery, audio, code, and videos in response to textual instructions. Without human oversight, guidance and responsible design and operation, such generative AI applications will remain a party trick with substantial potential for creating and spreading misinformation or harmful and inaccurate content at unprecedented scale. However, if positioned and developed responsibly as companions to humans augmenting but not replacing their role in decision making, knowledge retrieval and other cognitive processes, they could evolve into highly efficient, trustworthy, assistive tools for information management. This perspective describes how such tools could transform data management workflows in healthcare and medicine, explains how the underlying technology works, provides an assessment of risks and limitations, and proposes an ethical, technical, and cultural framework for responsible design, development, and deployment. It seeks to incentivise users, developers, providers, and regulators of generative AI that utilises LLMs to collectively prepare for the transformational role this technology could play in evidence-based sectors.

Harrer Stefan

2023-Mar-14

AI ethics, AI trustworthiness, Augmented human intelligence, Foundation models, Generative artificial intelligence, Information management, Large language models

General General

Revealing influencing factors on global waste distribution via deep-learning based dumpsite detection from satellite imagery.

In Nature communications ; h5-index 260.0

With the advancement of global civilisation, monitoring and managing dumpsites have become essential parts of environmental governance in various countries. Dumpsite locations are difficult to obtain in a timely manner by local government agencies and environmental groups. The World Bank shows that governments need to spend massive labour and economic costs to collect illegal dumpsites to implement management. Here we show that applying novel deep convolutional networks to high-resolution satellite images can provide an effective, efficient, and low-cost method to detect dumpsites. In sampled areas of 28 cities around the world, our model detects nearly 1000 dumpsites that appeared around 2021. This approach reduces the investigation time by more than 96.8% compared with the manual method. With this novel and powerful methodology, it is now capable of analysing the relationship between dumpsites and various social attributes on a global scale, temporally and spatially.

Sun Xian, Yin Dongshuo, Qin Fei, Yu Hongfeng, Lu Wanxuan, Yao Fanglong, He Qibin, Huang Xingliang, Yan Zhiyuan, Wang Peijin, Deng Chubo, Liu Nayu, Yang Yiran, Liang Wei, Wang Ruiping, Wang Cheng, Yokoya Naoto, Hänsch Ronny, Fu Kun

2023-Mar-15

oncology Oncology

Investigation of liquid biopsy analytes in peripheral blood of individuals after SARS-CoV-2 infection.

In EBioMedicine

BACKGROUND : Post-acute COVID-19 syndrome (PACS) is linked to severe organ damage. The identification and stratification of at-risk SARS-CoV-2 infected individuals is vital to providing appropriate care. This exploratory study looks for a potential liquid biopsy signal for PACS using both manual and machine learning approaches.

METHODS : Using a high definition single cell assay (HDSCA) workflow for liquid biopsy, we analysed 100 Post-COVID patients and 19 pre-pandemic normal donor (ND) controls. Within our patient cohort, 73 had received at least 1 dose of vaccination prior to SARS-CoV-2 infection. We stratified the COVID patients into 25 asymptomatic, 22 symptomatic COVID-19 but not suspected for PACS and 53 PACS suspected. All COVID-19 patients investigated in this study were diagnosed between April 2020 and January 2022 with a median 243 days (range 16-669) from diagnosis to their blood draw. We did a histopathological examination of rare events in the peripheral blood and used a machine learning model to evaluate predictors of PACS.

FINDINGS : The manual classification found rare cellular and acellular events consistent with features of endothelial cells and platelet structures in the PACS-suspected cohort. The three categories encompassing the hypothesised events were observed at a significantly higher incidence in the PACS-suspected cohort compared to the ND (p-value < 0.05). The machine learning classifier performed well when separating the NDs from Post-COVID with an accuracy of 90.1%, but poorly when separating the patients suspected and not suspected of PACS with an accuracy of 58.7%.

INTERPRETATION : Both the manual and the machine learning model found differences in the Post-COVID cohort and the NDs, suggesting the existence of a liquid biopsy signal after active SARS-CoV-2 infection. More research is needed to stratify PACS and its subsyndromes.

FUNDING : This work was funded in whole or in part by Fulgent Genetics, Kathy and Richard Leventhal and Vassiliadis Research Fund. This work was also supported by the National Cancer InstituteU54CA260591.

Qi Elizabeth, Courcoubetis George, Liljegren Emmett, Herrera Ergueen, Nguyen Nathalie, Nadri Maimoona, Ghandehari Sara, Kazemian Elham, Reckamp Karen L, Merin Noah M, Merchant Akil, Mason Jeremy, Figueiredo Jane C, Shishido Stephanie N, Kuhn Peter

2023-Mar-13

COVID-19, Liquid biopsy, Long COVID, Post-COVID sequelae, Post-acute COVID-19 syndrome (PACS), SARS-CoV-2

oncology Oncology

Investigation of liquid biopsy analytes in peripheral blood of individuals after SARS-CoV-2 infection.

In EBioMedicine

BACKGROUND : Post-acute COVID-19 syndrome (PACS) is linked to severe organ damage. The identification and stratification of at-risk SARS-CoV-2 infected individuals is vital to providing appropriate care. This exploratory study looks for a potential liquid biopsy signal for PACS using both manual and machine learning approaches.

METHODS : Using a high definition single cell assay (HDSCA) workflow for liquid biopsy, we analysed 100 Post-COVID patients and 19 pre-pandemic normal donor (ND) controls. Within our patient cohort, 73 had received at least 1 dose of vaccination prior to SARS-CoV-2 infection. We stratified the COVID patients into 25 asymptomatic, 22 symptomatic COVID-19 but not suspected for PACS and 53 PACS suspected. All COVID-19 patients investigated in this study were diagnosed between April 2020 and January 2022 with a median 243 days (range 16-669) from diagnosis to their blood draw. We did a histopathological examination of rare events in the peripheral blood and used a machine learning model to evaluate predictors of PACS.

FINDINGS : The manual classification found rare cellular and acellular events consistent with features of endothelial cells and platelet structures in the PACS-suspected cohort. The three categories encompassing the hypothesised events were observed at a significantly higher incidence in the PACS-suspected cohort compared to the ND (p-value < 0.05). The machine learning classifier performed well when separating the NDs from Post-COVID with an accuracy of 90.1%, but poorly when separating the patients suspected and not suspected of PACS with an accuracy of 58.7%.

INTERPRETATION : Both the manual and the machine learning model found differences in the Post-COVID cohort and the NDs, suggesting the existence of a liquid biopsy signal after active SARS-CoV-2 infection. More research is needed to stratify PACS and its subsyndromes.

FUNDING : This work was funded in whole or in part by Fulgent Genetics, Kathy and Richard Leventhal and Vassiliadis Research Fund. This work was also supported by the National Cancer InstituteU54CA260591.

Qi Elizabeth, Courcoubetis George, Liljegren Emmett, Herrera Ergueen, Nguyen Nathalie, Nadri Maimoona, Ghandehari Sara, Kazemian Elham, Reckamp Karen L, Merin Noah M, Merchant Akil, Mason Jeremy, Figueiredo Jane C, Shishido Stephanie N, Kuhn Peter

2023-Mar-13

COVID-19, Liquid biopsy, Long COVID, Post-COVID sequelae, Post-acute COVID-19 syndrome (PACS), SARS-CoV-2

Pathology Pathology

1,2,4,5-Tetrazine-tethered probes for fluorogenically imaging superoxide in live cells with ultrahigh specificity.

In Nature communications ; h5-index 260.0

Superoxide (O2·-) is the primary reactive oxygen species in mammal cells. Detecting superoxide is crucial for understanding redox signaling but remains challenging. Herein, we introduce a class of activity-based sensing probes. The probes utilize 1,2,4,5-tetrazine as a superoxide-responsive trigger, which can be modularly tethered to various fluorophores to tune probe sensitivity and emission color. These probes afford ultra-specific and ultra-fluorogenic responses towards superoxide, and enable multiplexed imaging of various cellular superoxide levels in an organelle-resolved way. Notably, the probes reveal the aberrant superoxide generation in the pathology of myocardial ischemia/reperfusion injury, and facilitate the establishment of a high-content screening pipeline for mediators of superoxide homeostasis. One such identified mediator, coprostanone, is shown to effectively ameliorating oxidative stress-induced injury in mice with myocardial ischemia/reperfusion injury. Collectively, these results showcase the potential of 1,2,4,5-tetrazine-tethered probes as versatile tools to monitor superoxide in a range of pathophysiological settings.

Jiang Xuefeng, Li Min, Wang Yule, Wang Chao, Wang Yingchao, Shen Tianruo, Shen Lili, Liu Xiaogang, Wang Yi, Li Xin

2023-Mar-14

Surgery Surgery

Pathologist Validation of a Machine Learning-Derived Feature for Colon Cancer Risk Stratification.

In JAMA network open

IMPORTANCE : Identifying new prognostic features in colon cancer has the potential to refine histopathologic review and inform patient care. Although prognostic artificial intelligence systems have recently demonstrated significant risk stratification for several cancer types, studies have not yet shown that the machine learning-derived features associated with these prognostic artificial intelligence systems are both interpretable and usable by pathologists.

OBJECTIVE : To evaluate whether pathologist scoring of a histopathologic feature previously identified by machine learning is associated with survival among patients with colon cancer.

DESIGN, SETTING, AND PARTICIPANTS : This prognostic study used deidentified, archived colorectal cancer cases from January 2013 to December 2015 from the University of Milano-Bicocca. All available histologic slides from 258 consecutive colon adenocarcinoma cases were reviewed from December 2021 to February 2022 by 2 pathologists, who conducted semiquantitative scoring for tumor adipose feature (TAF), which was previously identified via a prognostic deep learning model developed with an independent colorectal cancer cohort.

MAIN OUTCOMES AND MEASURES : Prognostic value of TAF for overall survival and disease-specific survival as measured by univariable and multivariable regression analyses. Interpathologist agreement in TAF scoring was also evaluated.

RESULTS : A total of 258 colon adenocarcinoma histopathologic cases from 258 patients (138 men [53%]; median age, 67 years [IQR, 65-81 years]) with stage II (n = 119) or stage III (n = 139) cancer were included. Tumor adipose feature was identified in 120 cases (widespread in 63 cases, multifocal in 31, and unifocal in 26). For overall survival analysis after adjustment for tumor stage, TAF was independently prognostic in 2 ways: TAF as a binary feature (presence vs absence: hazard ratio [HR] for presence of TAF, 1.55 [95% CI, 1.07-2.25]; P = .02) and TAF as a semiquantitative categorical feature (HR for widespread TAF, 1.87 [95% CI, 1.23-2.85]; P = .004). Interpathologist agreement for widespread TAF vs lower categories (absent, unifocal, or multifocal) was 90%, corresponding to a κ metric at this threshold of 0.69 (95% CI, 0.58-0.80).

CONCLUSIONS AND RELEVANCE : In this prognostic study, pathologists were able to learn and reproducibly score for TAF, providing significant risk stratification on this independent data set. Although additional work is warranted to understand the biological significance of this feature and to establish broadly reproducible TAF scoring, this work represents the first validation to date of human expert learning from machine learning in pathology. Specifically, this validation demonstrates that a computationally identified histologic feature can represent a human-identifiable, prognostic feature with the potential for integration into pathology practice.

L’Imperio Vincenzo, Wulczyn Ellery, Plass Markus, Müller Heimo, Tamini Nicolò, Gianotti Luca, Zucchini Nicola, Reihs Robert, Corrado Greg S, Webster Dale R, Peng Lily H, Chen Po-Hsuan Cameron, Lavitrano Marialuisa, Liu Yun, Steiner David F, Zatloukal Kurt, Pagni Fabio

2023-Mar-01

Surgery Surgery

An artificial intelligence approach for predicting death or organ failure after hospitalization for COVID-19: development of a novel risk prediction tool and comparisons with ISARIC-4C, CURB-65, qSOFA, and MEWS scoring systems.

In Respiratory research ; h5-index 45.0

BACKGROUND : We applied machine learning (ML) algorithms to generate a risk prediction tool [Collaboration for Risk Evaluation in COVID-19 (CORE-COVID-19)] for predicting the composite of 30-day endotracheal intubation, intravenous administration of vasopressors, or death after COVID-19 hospitalization and compared it with the existing risk scores.

METHODS : This is a retrospective study of adults hospitalized with COVID-19 from March 2020 to February 2021. Patients, each with 92 variables, and one composite outcome underwent feature selection process to identify the most predictive variables. Selected variables were modeled to build four ML algorithms (artificial neural network, support vector machine, gradient boosting machine, and Logistic regression) and an ensemble model to generate a CORE-COVID-19 model to predict the composite outcome and compared with existing risk prediction scores. The net benefit for clinical use of each model was assessed by decision curve analysis.

RESULTS : Of 1796 patients, 278 (15%) patients reached primary outcome. Six most predictive features were identified. Four ML algorithms achieved comparable discrimination (P > 0.827) with c-statistics ranged 0.849-0.856, calibration slopes 0.911-1.173, and Hosmer-Lemeshow P > 0.141 in validation dataset. These 6-variable fitted CORE-COVID-19 model revealed a c-statistic of 0.880, which was significantly (P < 0.04) higher than ISARIC-4C (0.751), CURB-65 (0.735), qSOFA (0.676), and MEWS (0.674) for outcome prediction. The net benefit of the CORE-COVID-19 model was greater than that of the existing risk scores.

CONCLUSION : The CORE-COVID-19 model accurately assigned 88% of patients who potentially progressed to 30-day composite events and revealed improved performance over existing risk scores, indicating its potential utility in clinical practice.

Kwok Stephen Wai Hang, Wang Guanjin, Sohel Ferdous, Kashani Kianoush B, Zhu Ye, Wang Zhen, Antpack Eduardo, Khandelwal Kanika, Pagali Sandeep R, Nanda Sanjeev, Abdalrhim Ahmed D, Sharma Umesh M, Bhagra Sumit, Dugani Sagar, Takahashi Paul Y, Murad Mohammad H, Yousufuddin Mohammed

2023-Mar-13

COVID-19, Machine learning algorithms, Mortality, Organ failure, Prediction models

Surgery Surgery

An artificial intelligence approach for predicting death or organ failure after hospitalization for COVID-19: development of a novel risk prediction tool and comparisons with ISARIC-4C, CURB-65, qSOFA, and MEWS scoring systems.

In Respiratory research ; h5-index 45.0

BACKGROUND : We applied machine learning (ML) algorithms to generate a risk prediction tool [Collaboration for Risk Evaluation in COVID-19 (CORE-COVID-19)] for predicting the composite of 30-day endotracheal intubation, intravenous administration of vasopressors, or death after COVID-19 hospitalization and compared it with the existing risk scores.

METHODS : This is a retrospective study of adults hospitalized with COVID-19 from March 2020 to February 2021. Patients, each with 92 variables, and one composite outcome underwent feature selection process to identify the most predictive variables. Selected variables were modeled to build four ML algorithms (artificial neural network, support vector machine, gradient boosting machine, and Logistic regression) and an ensemble model to generate a CORE-COVID-19 model to predict the composite outcome and compared with existing risk prediction scores. The net benefit for clinical use of each model was assessed by decision curve analysis.

RESULTS : Of 1796 patients, 278 (15%) patients reached primary outcome. Six most predictive features were identified. Four ML algorithms achieved comparable discrimination (P > 0.827) with c-statistics ranged 0.849-0.856, calibration slopes 0.911-1.173, and Hosmer-Lemeshow P > 0.141 in validation dataset. These 6-variable fitted CORE-COVID-19 model revealed a c-statistic of 0.880, which was significantly (P < 0.04) higher than ISARIC-4C (0.751), CURB-65 (0.735), qSOFA (0.676), and MEWS (0.674) for outcome prediction. The net benefit of the CORE-COVID-19 model was greater than that of the existing risk scores.

CONCLUSION : The CORE-COVID-19 model accurately assigned 88% of patients who potentially progressed to 30-day composite events and revealed improved performance over existing risk scores, indicating its potential utility in clinical practice.

Kwok Stephen Wai Hang, Wang Guanjin, Sohel Ferdous, Kashani Kianoush B, Zhu Ye, Wang Zhen, Antpack Eduardo, Khandelwal Kanika, Pagali Sandeep R, Nanda Sanjeev, Abdalrhim Ahmed D, Sharma Umesh M, Bhagra Sumit, Dugani Sagar, Takahashi Paul Y, Murad Mohammad H, Yousufuddin Mohammed

2023-Mar-13

COVID-19, Machine learning algorithms, Mortality, Organ failure, Prediction models

oncology Oncology

A non-antibiotic-disrupted gut microbiome is associated with clinical responses to CD19-CAR-T cell cancer immunotherapy.

In Nature medicine ; h5-index 170.0

Increasing evidence suggests that the gut microbiome may modulate the efficacy of cancer immunotherapy. In a B cell lymphoma patient cohort from five centers in Germany and the United States (Germany, n = 66; United States, n = 106; total, n = 172), we demonstrate that wide-spectrum antibiotics treatment ('high-risk antibiotics') prior to CD19-targeted chimeric antigen receptor (CAR)-T cell therapy is associated with adverse outcomes, but this effect is likely to be confounded by an increased pretreatment tumor burden and systemic inflammation in patients pretreated with high-risk antibiotics. To resolve this confounding effect and gain insights into antibiotics-masked microbiome signals impacting CAR-T efficacy, we focused on the high-risk antibiotics non-exposed patient population. Indeed, in these patients, significant correlations were noted between pre-CAR-T infusion Bifidobacterium longum and microbiome-encoded peptidoglycan biosynthesis, and CAR-T treatment-associated 6-month survival or lymphoma progression. Furthermore, predictive pre-CAR-T treatment microbiome-based machine learning algorithms trained on the high-risk antibiotics non-exposed German cohort and validated by the respective US cohort robustly segregated long-term responders from non-responders. Bacteroides, Ruminococcus, Eubacterium and Akkermansia were most important in determining CAR-T responsiveness, with Akkermansia also being associated with pre-infusion peripheral T cell levels in these patients. Collectively, we identify conserved microbiome features across clinical and geographical variations, which may enable cross-cohort microbiome-based predictions of outcomes in CAR-T cell immunotherapy.

Stein-Thoeringer Christoph K, Saini Neeraj Y, Zamir Eli, Blumenberg Viktoria, Schubert Maria-Luisa, Mor Uria, Fante Matthias A, Schmidt Sabine, Hayase Eiko, Hayase Tomo, Rohrbach Roman, Chang Chia-Chi, McDaniel Lauren, Flores Ivonne, Gaiser Rogier, Edinger Matthias, Wolff Daniel, Heidenreich Martin, Strati Paolo, Nair Ranjit, Chihara Dai, Fayad Luis E, Ahmed Sairah, Iyer Swaminathan P, Steiner Raphael E, Jain Preetesh, Nastoupil Loretta J, Westin Jason, Arora Reetakshi, Wang Michael L, Turner Joel, Menges Meghan, Hidalgo-Vargas Melanie, Reid Kayla, Dreger Peter, Schmitt Anita, Müller-Tidow Carsten, Locke Frederick L, Davila Marco L, Champlin Richard E, Flowers Christopher R, Shpall Elizabeth J, Poeck Hendrik, Neelapu Sattva S, Schmitt Michael, Subklewe Marion, Jain Michael D, Jenq Robert R, Elinav Eran

2023-Mar-13

Pathology Pathology

Interpretable and context-free deconvolution of multi-scale whole transcriptomic data with UniCell deconvolve.

In Nature communications ; h5-index 260.0

We introduce UniCell: Deconvolve Base (UCDBase), a pre-trained, interpretable, deep learning model to deconvolve cell type fractions and predict cell identity across Spatial, bulk-RNA-Seq, and scRNA-Seq datasets without contextualized reference data. UCD is trained on 10 million pseudo-mixtures from a fully-integrated scRNA-Seq training database comprising over 28 million annotated single cells spanning 840 unique cell types from 898 studies. We show that our UCDBase and transfer-learning models achieve comparable or superior performance on in-silico mixture deconvolution to existing, reference-based, state-of-the-art methods. Feature attribute analysis uncovers gene signatures associated with cell-type specific inflammatory-fibrotic responses in ischemic kidney injury, discerns cancer subtypes, and accurately deconvolves tumor microenvironments. UCD identifies pathologic changes in cell fractions among bulk-RNA-Seq data for several disease states. Applied to lung cancer scRNA-Seq data, UCD annotates and distinguishes normal from cancerous cells. Overall, UCD enhances transcriptomic data analysis, aiding in assessment of cellular and spatial context.

Charytonowicz Daniel, Brody Rachel, Sebra Robert

2023-Mar-11

Internal Medicine Internal Medicine

Effects of gastroesophageal reflux disease treatment with proton pump inhibitors on the risk of acute exacerbation and pneumonia in patients with COPD.

In Respiratory research ; h5-index 45.0

BACKGROUND : Gastroesophageal reflux disease (GERD) has been suggested as a risk factor for acute exacerbation of chronic obstructive pulmonary disease (COPD). However, it remains undetermined whether proton pump inhibitor (PPI) treatment reduces the risk of exacerbation or affects the risk of pneumonia. This study aimed to evaluate the risks of both exacerbation and pneumonia following PPI treatment for GERD in patients with COPD.

METHODS : This study used a reimbursement database of the Republic of Korea. Patients aged ≥ 40 years with COPD as a main diagnosis and who received PPI treatment for GERD at least for 14 consecutive days between January 2013 and December 2018 were included in the study. A self-controlled case series analysis was conducted to calculate the risk of moderate and severe exacerbation and pneumonia.

RESULTS : A total of 104,439 patients with prevalent COPD received PPI treatment for GERD. The risk of moderate exacerbation was significantly lower during the PPI treatment than at baseline. The risk of severe exacerbation increased during the PPI treatment but significantly decreased in the post-treatment period. Pneumonia risk was not significantly increased during the PPI treatment. The results were similar in patients with incident COPD.

CONCLUSIONS : The risk of exacerbation was significantly reduced after PPI treatment compared with the non-treated period. Severe exacerbation may increase due to uncontrolled GERD but subsequently decrease following PPI treatment. There was no evidence of an increased risk of pneumonia.

Kang Jieun, Lee Rugyeom, Lee Sei Won

2023-Mar-11

Acute exacerbation, Chronic obstructive pulmonary disease, Gastroesophageal reflux disease, Pneumonia, Proton pump inhibitor

General General

Smartphone-based platforms implementing microfluidic detection with image-based artificial intelligence.

In Nature communications ; h5-index 260.0

The frequent outbreak of global infectious diseases has prompted the development of rapid and effective diagnostic tools for the early screening of potential patients in point-of-care testing scenarios. With advances in mobile computing power and microfluidic technology, the smartphone-based mobile health platform has drawn significant attention from researchers developing point-of-care testing devices that integrate microfluidic optical detection with artificial intelligence analysis. In this article, we summarize recent progress in these mobile health platforms, including the aspects of microfluidic chips, imaging modalities, supporting components, and the development of software algorithms. We document the application of mobile health platforms in terms of the detection objects, including molecules, viruses, cells, and parasites. Finally, we discuss the prospects for future development of mobile health platforms.

Wang Bangfeng, Li Yiwei, Zhou Mengfan, Han Yulong, Zhang Mingyu, Gao Zhaolong, Liu Zetai, Chen Peng, Du Wei, Zhang Xingcai, Feng Xiaojun, Liu Bi-Feng

2023-Mar-11

General General

A deep intronic TCTN2 variant activating a cryptic exon predicted by SpliceRover in a patient with Joubert syndrome.

In Journal of human genetics

The recent introduction of genome sequencing in genetic analysis has led to the identification of pathogenic variants located in deep introns. Recently, several new tools have emerged to predict the impact of variants on splicing. Here, we present a Japanese boy of Joubert syndrome with biallelic TCTN2 variants. Exome sequencing identified only a heterozygous maternal nonsense TCTN2 variant (NM_024809.5:c.916C >T, p.(Gln306Ter)). Subsequent genome sequencing identified a deep intronic variant (c.1033+423G>A) inherited from his father. The machine learning algorithms SpliceAI, Squirls, and Pangolin were unable to predict alterations in splicing by the c.1033+423G>A variant. SpliceRover, a tool for splice site prediction using FASTA sequence, was able to detect a cryptic exon which was 85-bp away from the variant and within the inverted Alu sequence while SpliceRover scores for these splice sites showed slight increase (donor) or decrease (acceptor) between the reference and mutant sequences. RNA sequencing and RT-PCR using urinary cells confirmed inclusion of the cryptic exon. The patient showed major symptoms of TCTN2-related disorders such as developmental delay, dysmorphic facial features and polydactyly. He also showed uncommon features such as retinal dystrophy, exotropia, abnormal pattern of respiration, and periventricular heterotopia, confirming these as one of features of TCTN2-related disorders. Our study highlights usefulness of genome sequencing and RNA sequencing using urinary cells for molecular diagnosis of genetic disorders and suggests that database of cryptic splice sites predicted in introns by SpliceRover using the reference sequences can be helpful in extracting candidate variants from large numbers of intronic variants in genome sequencing.

Hiraide Takuya, Shimizu Kenji, Okumura Yoshinori, Miyamoto Sachiko, Nakashima Mitsuko, Ogata Tsutomu, Saitsu Hirotomo

2023-Mar-10

Surgery Surgery

Mechanism of Injury and Age Predict Operative Intervention in Pediatric Perineal Injury.

In Pediatric emergency care

OBJECTIVES : Literature characterizing pediatric perineal trauma is sparse and generally limited to females. The purpose of this study was to characterize pediatric perineal injuries with specific focus on patient demographics, mechanisms of injury, and care patterns at a regional level 1 pediatric trauma center.

METHODS : Retrospective review of children aged younger than 18 years evaluated at a level 1 pediatric trauma center from 2006 to 2017. Patients were identified by International Classification of Diseases-9 and 10 codes. Extracted data included demographics, injury mechanism, diagnostic studies, hospital course, and structures injured. The χ2 and t tests were used to examine differences between subgroups. Machine learning was used to predict variable importance in determining the need for operative interventions.

RESULTS : One hundred ninety-seven patients met inclusion criteria. Mean age was 8.5 years. A total of 50.8% were girls. Blunt trauma accounted for 83.8% of injuries. Motor vehicle collisions and foreign bodies were more common in patients aged 12 years and older, whereas falls and bicycle-related injuries were more common in those younger than 12 years (P < 0.01). Patients younger than 12 years were more likely to sustain blunt trauma with isolated external genital injuries (P < 0.01). Patients aged 12 and older had a higher incidence of pelvic fractures, bladder/urethral injuries, and colorectal injuries, suggesting more severe injury patterns (P < 0.01). Half of patients required operative intervention. Children aged 3 years or younger and older than 12 years had longer mean hospital stays compared with children aged 4 to 11 years (P < 0.01). Mechanism of injury and age constituted more than 75% of the variable importance in predicting operative intervention.

CONCLUSIONS : Perineal trauma in children varies by age, sex, and mechanism. Blunt mechanisms are the most common, with patients frequently requiring surgical intervention. Mechanism of injury and age may be important in deciding which patients will require operative intervention. This study describes injury patterns in pediatric perineal trauma that can be used to guide future practice and inform injury prevention efforts.

McLaughlin Christopher J, Martin Kathryn L

2023-Mar-07

General General

Human-machine collaboration for improving semiconductor process development.

In Nature ; h5-index 368.0

One of the bottlenecks to building semiconductor chips is the increasing cost required to develop chemical plasma processes that form the transistors and memory storage cells1,2. These processes are still developed manually using highly trained engineers searching for a combination of tool parameters that produces an acceptable result on the silicon wafer3. The challenge for computer algorithms is the availability of limited experimental data owing to the high cost of acquisition, making it difficult to form a predictive model with accuracy to the atomic scale. Here we study Bayesian optimization algorithms to investigate how artificial intelligence (AI) might decrease the cost of developing complex semiconductor chip processes. In particular, we create a controlled virtual process game to systematically benchmark the performance of humans and computers for the design of a semiconductor fabrication process. We find that human engineers excel in the early stages of development, whereas the algorithms are far more cost-efficient near the tight tolerances of the target. Furthermore, we show that a strategy using both human designers with high expertise and algorithms in a human first-computer last strategy can reduce the cost-to-target by half compared with only human designers. Finally, we highlight cultural challenges in partnering humans with computers that need to be addressed when introducing artificial intelligence in developing semiconductor processes.

Kanarik Keren J, Osowiecki Wojciech T, Lu Yu Joe, Talukder Dipongkar, Roschewsky Niklas, Park Sae Na, Kamon Mattan, Fried David M, Gottscho Richard A

2023-Mar-08

General General

Diagnosis of Alzheimer's Disease and Tauopathies on Whole Slide Histopathology Images Using a Weakly Supervised Deep Learning Algorithm.

In Laboratory investigation; a journal of technical methods and pathology ; h5-index 42.0

Neuropathological assessment at autopsy is the gold standard for diagnosing neurodegenerative disorders. Neurodegenerative conditions, such as Alzheimer's disease (AD), neuropathological changes are a continuous process from normal aging rather than categorical; therefore, diagnosing neurodegenerative disorders is a complicated task. We aimed to develop a pipeline for diagnosing AD and other tauopathies, including corticobasal degeneration (CBD), globular glial tauopathy (GGT), Pick's disease (PiD), and progressive supranuclear palsy (PSP). We used a weakly supervised deep learning-based approach called clustering-constrained-attention multiple instance learning (CLAM) on whole slide images (WSIs) of patients with AD (n=30), CBD (n=20), GGT (n=10), PiD (n=20), and PSP (n=20), as well as non-tauopathy controls (n=21). Three sections (A: motor cortex; B: cingulate gyrus and superior frontal gyrus; C: corpus striatum) that had been immunostained for phosphorylated-tau were scanned and converted to WSIs. We evaluated three models (classical multiple instance learning, single-attention-branch CLAM, and multi-attention-branch CLAM) using 5-fold cross-validation. Attention-based interpretation analysis was performed to identify morphological features contributing to the classification. Within highly attended regions, we also augmented gradient-weighted class activation mapping (Grad-CAM) to the model to visualize cellular-level evidence of the model's decisions. The multi-attention-branch CLAM model using Section B achieved the highest area under the curve (0.970 ± 0.037) and diagnostic accuracy (0.873 ± 0.087). A heatmap showed the highest attention in the gray matter of the superior frontal gyrus in AD and the white matter of the cingulate gyrus in CBD. Grad-CAM showed the highest attention in characteristic tau lesions for each disease (e.g., numerous tau-positive threads in the white matter inclusions for CBD). Our findings supported the feasibility of deep learning-based approaches for the classification of neurodegenerative disorders on WSIs. Further investigation of this method focusing on clinicopathological correlations, is warranted.

Kim Minji, Sekiya Hiroaki, Yao Gary, Martin Nicholas B, Castanedes-Casey Monica, Dickson Dennis W, Hwang Tae Hyun, Koga Shunsuke

2023-Mar-06

Alzheimer’s disease, CLAM, Grad-CAM, Pick’s disease, corticobasal degeneration, globular glial tauopathy, neuropathology, progressive supranuclear palsy, tauopathy, weakly supervised deep learning

Radiology Radiology

Autonomous Chest Radiograph Reporting Using AI: Estimation of Clinical Impact.

In Radiology ; h5-index 91.0

Background Automated interpretation of normal chest radiographs could alleviate the workload of radiologists. However, the performance of such an artificial intelligence (AI) tool compared with clinical radiology reports has not been established. Purpose To perform an external evaluation of a commercially available AI tool for (a) the number of chest radiographs autonomously reported, (b) the sensitivity for AI detection of abnormal chest radiographs, and (c) the performance of AI compared with that of the clinical radiology reports. Materials and Methods In this retrospective study, consecutive posteroanterior chest radiographs from adult patients in four hospitals in the capital region of Denmark were obtained in January 2020, including images from emergency department patients, in-hospital patients, and outpatients. Three thoracic radiologists labeled chest radiographs in a reference standard based on chest radiograph findings into the following categories: critical, other remarkable, unremarkable, or normal (no abnormalities). AI classified chest radiographs as high confidence normal (normal) or not high confidence normal (abnormal). Results A total of 1529 patients were included for analysis (median age, 69 years [IQR, 55-69 years]; 776 women), with 1100 (72%) classified by the reference standard as having abnormal radiographs, 617 (40%) as having critical abnormal radiographs, and 429 (28%) as having normal radiographs. For comparison, clinical radiology reports were classified based on the text and insufficient reports excluded (n = 22). The sensitivity of AI was 99.1% (95% CI: 98.3, 99.6; 1090 of 1100 patients) for abnormal radiographs and 99.8% (95% CI: 99.1, 99.9; 616 of 617 patients) for critical radiographs. Corresponding sensitivities for radiologist reports were 72.3% (95% CI: 69.5, 74.9; 779 of 1078 patients) and 93.5% (95% CI: 91.2, 95.3; 558 of 597 patients), respectively. Specificity of AI, and hence the potential autonomous reporting rate, was 28.0% of all normal posteroanterior chest radiographs (95% CI: 23.8, 32.5; 120 of 429 patients), or 7.8% (120 of 1529 patients) of all posteroanterior chest radiographs. Conclusion Of all normal posteroanterior chest radiographs, 28% were autonomously reported by AI with a sensitivity for any abnormalities higher than 99%. This corresponded to 7.8% of the entire posteroanterior chest radiograph production. © RSNA, 2023 Supplemental material is available for this article. See also the editorial by Park in this issue.

Plesner Louis L, Müller Felix C, Nybing Janus D, Laustrup Lene C, Rasmussen Finn, Nielsen Olav W, Boesen Mikael, Andersen Michael B

2023-Mar-07

Radiology Radiology

Deep Learning for Head and Neck CT Angiography: Stenosis and Plaque Classification.

In Radiology ; h5-index 91.0

Background Studies have rarely investigated stenosis detection from head and neck CT angiography scans because accurate interpretation is time consuming and labor intensive. Purpose To develop an automated convolutional neural network-based method for accurate stenosis detection and plaque classification in head and neck CT angiography images and compare its performance with that of radiologists. Materials and Methods A deep learning (DL) algorithm was constructed and trained with use of head and neck CT angiography images that were collected retrospectively from four tertiary hospitals between March 2020 and July 2021. CT scans were partitioned into training, validation, and independent test sets at a ratio of 7:2:1. An independent test set of CT angiography scans was collected prospectively between October 2021 and December 2021 in one of the four tertiary centers. Stenosis grade categories were as follows: mild stenosis (<50%), moderate stenosis (50%-69%), severe stenosis (70%-99%), and occlusion (100%). The stenosis diagnosis and plaque classification of the algorithm were compared with the ground truth of consensus by two radiologists (with more than 10 years of experience). The performance of the models was analyzed in terms of accuracy, sensitivity, specificity, and areas under the receiver operating characteristic curve. Results There were 3266 patients (mean age ± SD, 62 years ± 12; 2096 men) evaluated. The consistency between radiologists and the DL-assisted algorithm on plaque classification was 85.6% (320 of 374 cases [95% CI: 83.2, 88.6]) on a per-vessel basis. Moreover, the artificial intelligence model assisted in visual assessment, such as increasing confidence in the degree of stenosis. This reduced the time needed for diagnosis and report writing of radiologists from 28.8 minutes ± 5.6 to 12.4 minutes ± 2.0 (P < .001). Conclusion A deep learning algorithm for head and neck CT angiography interpretation accurately determined vessel stenosis and plaque classification and had equivalent diagnostic performance when compared with experienced radiologists. © RSNA, 2023 Supplemental material is available for this article.

Fu Fan, Shan Yi, Yang Guang, Zheng Chao, Zhang Miao, Rong Dongdong, Wang Ximing, Lu Jie

2023-Mar-07

General General

Predicting functional effects of ion channel variants using new phenotypic machine learning methods.

In PLoS computational biology

Missense variants in genes encoding ion channels are associated with a spectrum of severe diseases. Variant effects on biophysical function correlate with clinical features and can be categorized as gain- or loss-of-function. This information enables a timely diagnosis, facilitates precision therapy, and guides prognosis. Functional characterization presents a bottleneck in translational medicine. Machine learning models may be able to rapidly generate supporting evidence by predicting variant functional effects. Here, we describe a multi-task multi-kernel learning framework capable of harmonizing functional results and structural information with clinical phenotypes. This novel approach extends the human phenotype ontology towards kernel-based supervised machine learning. Our gain- or loss-of-function classifier achieves high performance (mean accuracy 0.853 SD 0.016, mean AU-ROC 0.912 SD 0.025), outperforming both conventional baseline and state-of-the-art methods. Performance is robust across different phenotypic similarity measures and largely insensitive to phenotypic noise or sparsity. Localized multi-kernel learning offered biological insight and interpretability by highlighting channels with implicit genotype-phenotype correlations or latent task similarity for downstream analysis.

Boßelmann Christian Malte, Hedrich Ulrike B S, Lerche Holger, Pfeifer Nico

2023-Mar-06

General General

CustOmics: A versatile deep-learning based strategy for multi-omics integration.

In PLoS computational biology

The availability of patient cohorts with several types of omics data opens new perspectives for exploring the disease's underlying biological processes and developing predictive models. It also comes with new challenges in computational biology in terms of integrating high-dimensional and heterogeneous data in a fashion that captures the interrelationships between multiple genes and their functions. Deep learning methods offer promising perspectives for integrating multi-omics data. In this paper, we review the existing integration strategies based on autoencoders and propose a new customizable one whose principle relies on a two-phase approach. In the first phase, we adapt the training to each data source independently before learning cross-modality interactions in the second phase. By taking into account each source's singularity, we show that this approach succeeds at taking advantage of all the sources more efficiently than other strategies. Moreover, by adapting our architecture to the computation of Shapley Additive explanations, our model can provide interpretable results in a multi-source setting. Using multiple omics sources from different TCGA cohorts, we demonstrate the performance of the proposed method for cancer on test cases for several tasks, such as the classification of tumor types and breast cancer subtypes, as well as survival outcome prediction. We show through our experiments the great performances of our architecture on seven different datasets with various sizes and provide some interpretations of the results obtained. Our code is available on (https://github.com/HakimBenkirane/CustOmics).

Benkirane Hakim, Pradat Yoann, Michiels Stefan, Cournède Paul-Henry

2023-Mar-06

General General

Evaluation of Risk of Bias in Neuroimaging-Based Artificial Intelligence Models for Psychiatric Diagnosis: A Systematic Review.

In JAMA network open

IMPORTANCE : Neuroimaging-based artificial intelligence (AI) diagnostic models have proliferated in psychiatry. However, their clinical applicability and reporting quality (ie, feasibility) for clinical practice have not been systematically evaluated.

OBJECTIVE : To systematically assess the risk of bias (ROB) and reporting quality of neuroimaging-based AI models for psychiatric diagnosis.

EVIDENCE REVIEW : PubMed was searched for peer-reviewed, full-length articles published between January 1, 1990, and March 16, 2022. Studies aimed at developing or validating neuroimaging-based AI models for clinical diagnosis of psychiatric disorders were included. Reference lists were further searched for suitable original studies. Data extraction followed the CHARMS (Checklist for Critical Appraisal and Data Extraction for Systematic Reviews of Prediction Modeling Studies) and PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-analyses) guidelines. A closed-loop cross-sequential design was used for quality control. The PROBAST (Prediction Model Risk of Bias Assessment Tool) and modified CLEAR (Checklist for Evaluation of Image-Based Artificial Intelligence Reports) benchmarks were used to systematically evaluate ROB and reporting quality.

FINDINGS : A total of 517 studies presenting 555 AI models were included and evaluated. Of these models, 461 (83.1%; 95% CI, 80.0%-86.2%) were rated as having a high overall ROB based on the PROBAST. The ROB was particular high in the analysis domain, including inadequate sample size (398 of 555 models [71.7%; 95% CI, 68.0%-75.6%]), poor model performance examination (with 100% of models lacking calibration examination), and lack of handling data complexity (550 of 555 models [99.1%; 95% CI, 98.3%-99.9%]). None of the AI models was perceived to be applicable to clinical practices. Overall reporting completeness (ie, number of reported items/number of total items) for the AI models was 61.2% (95% CI, 60.6%-61.8%), and the completeness was poorest for the technical assessment domain with 39.9% (95% CI, 38.8%-41.1%).

CONCLUSIONS AND RELEVANCE : This systematic review found that the clinical applicability and feasibility of neuroimaging-based AI models for psychiatric diagnosis were challenged by a high ROB and poor reporting quality. Particularly in the analysis domain, ROB in AI diagnostic models should be addressed before clinical application.

Chen Zhiyi, Liu Xuerong, Yang Qingwu, Wang Yan-Jiang, Miao Kuan, Gong Zheng, Yu Yang, Leonov Artemiy, Liu Chunlei, Feng Zhengzhi, Chuan-Peng Hu

2023-Mar-01

General General

An AI-Aided Diagnostic Framework for Hematologic Neoplasms Based on Morphologic Features and Medical Expertise.

In Laboratory investigation; a journal of technical methods and pathology ; h5-index 42.0

A morphologic examination is essential for the diagnosis of hematological diseases. However, its conventional manual operation is time-consuming and laborious. Herein, we attempt to establish an artificial intelligence (AI)-aided diagnostic framework integrating medical expertise. This framework acts as a virtual hematological morphologist (VHM) for diagnosing hematological neoplasms. Two datasets were established as follows: An image dataset was used to train the Faster Region-based Convolutional Neural Network to develop an image-based morphologic feature extraction model. A case dataset containing retrospective morphologic diagnostic data was used to train a support vector machine algorithm to develop a feature-based case identification model based on diagnostic criteria. Integrating these 2 models established a whole-process AI-aided diagnostic framework, namely, VHM, and a 2-stage strategy was applied to practice case diagnosis. The recall and precision of VHM in bone marrow cell classification were 94.65% and 93.95%, respectively. The balanced accuracy, sensitivity, and specificity of VHM were 97.16%, 99.09%, and 92%, respectively, in the differential diagnosis of normal and abnormal cases, and 99.23%, 97.96%, and 100%, respectively, in the precise diagnosis of chronic myelogenous leukemia in chronic phase. This work represents the first attempt, to our knowledge, to extract multimodal morphologic features and to integrate a feature-based case diagnosis model for designing a comprehensive AI-aided morphologic diagnostic framework. The performance of our knowledge-based framework was superior to that of the widely used end-to-end AI-based diagnostic framework in terms of testing accuracy (96.88% vs 68.75%) or generalization ability (97.11% vs 68.75%) in differentiating normal and abnormal cases. The remarkable advantage of VHM is that it follows the logic of clinical diagnostic procedures, making it a reliable and interpretable hematological diagnostic tool.

Li Nan, Fan Liquan, Xu Hang, Zhang Xiwen, Bai Zanzhou, Li Miaohui, Xiong Shumin, Jiang Lu, Yang Jie, Chen Saijuan, Qiao Yu, Chen Bing

2023-Jan-10

artificial intelligence, bone marrow morphology, diagnosis, hematology, multimodal features

General General

Revealing the importance of prenatal gut microbiome in offspring neurodevelopment in humans.

In EBioMedicine

BACKGROUND : It has been widely recognized that a critical time window for neurodevelopment occurs in early life and the host's gut microbiome plays an important role in neurodevelopment. Following recent demonstrations that the maternal prenatal gut microbiome influences offspring brain development in murine models, we aim to explore whether the critical time window for the association between the gut microbiome and neurodevelopment is prenatal or postnatal for human.

METHODS : Here we leverage a large-scale human study and compare the associations between the gut microbiota and metabolites from mothers during pregnancy and their children with the children's neurodevelopment. Specifically, using multinomial regression integrated in Songbird, we assessed the discriminating power of the maternal prenatal and child gut microbiome for children's neurodevelopment at early life as measured by the Ages & Stages Questionnaires (ASQ).

FINDINGS : We show that the maternal prenatal gut microbiome is more relevant than the children's gut microbiome to the children's neurodevelopment in the first year of life (maximum Q2 = 0.212 and 0.096 separately using the taxa at the class level). Moreover, we found that Fusobacteriia is more associated with high fine motor skills in ASQ in the maternal prenatal gut microbiota but become more associated with low fine motor skills in the infant gut microbiota (rank = 0.084 and -0.047 separately), suggesting the roles of the same taxa with respect to neurodevelopment can be opposite at the two stages of fetal neurodevelopment.

INTERPRETATION : These findings shed light, especially in terms of timing, on potential therapeutic interventions to prevent neurodevelopmental disorders.

FUNDING : This work was supported by the National Institutes of Health (grant numbers: R01AI141529, R01HD093761, RF1AG067744, UH3OD023268, U19AI095219, U01HL089856, R01HL141826, K08HL148178, K01HL146980), and the Charles A. King Trust Postdoctoral Fellowship.

Sun Zheng, Lee-Sarwar Kathleen, Kelly Rachel S, Lasky-Su Jessica A, Litonjua Augusto A, Weiss Scott T, Liu Yang-Yu

2023-Mar-01

Ages and stages questionnaire, Childhood neurodevelopment, Early-life gut microbiome, Maternal gut microbiome

Pathology Pathology

Integrated Cytometry With Machine Learning Applied to High-Content Imaging of Human Kidney Tissue for In Situ Cell Classification and Neighborhood Analysis.

In Laboratory investigation; a journal of technical methods and pathology ; h5-index 42.0

The human kidney is a complex organ with various cell types that are intricately organized to perform key physiological functions and maintain homeostasis. New imaging modalities, such as mesoscale and highly multiplexed fluorescence microscopy, are increasingly being applied to human kidney tissue to create single-cell resolution data sets that are both spatially large and multidimensional. These single-cell resolution high-content imaging data sets have great potential to uncover the complex spatial organization and cellular makeup of the human kidney. Tissue cytometry is a novel approach used for the quantitative analysis of imaging data; however, the scale and complexity of such data sets pose unique challenges for processing and analysis. We have developed the Volumetric Tissue Exploration and Analysis (VTEA) software, a unique tool that integrates image processing, segmentation, and interactive cytometry analysis into a single framework on desktop computers. Supported by an extensible and open-source framework, VTEA's integrated pipeline now includes enhanced analytical tools, such as machine learning, data visualization, and neighborhood analyses, for hyperdimensional large-scale imaging data sets. These novel capabilities enable the analysis of mesoscale 2- and 3-dimensional multiplexed human kidney imaging data sets (such as co-detection by indexing and 3-dimensional confocal multiplexed fluorescence imaging). We demonstrate the utility of this approach in identifying cell subtypes in the kidney on the basis of labels, spatial association, and their microenvironment or neighborhood membership. VTEA provides an integrated and intuitive approach to decipher the cellular and spatial complexity of the human kidney and complements other transcriptomics and epigenetic efforts to define the landscape of kidney cell types.

Winfree Seth, McNutt Andrew T, Khochare Suraj, Borgard Tyler J, Barwinska Daria, Sabo Angela R, Ferkowicz Michael J, Williams James C, Lingeman James E, Gulbronson Connor J, Kelly Katherine J, Sutton Timothy A, Dagher Pierre C, Eadon Michael T, Dunn Kenneth W, El-Achkar Tarek M

2023-Feb-04

3D, CODEX, FIJI is just ImageJ, confocal microscopy, cytometry, machine learning

General General

Cryo-EM structure of human heptameric pannexin 2 channel.

In Nature communications ; h5-index 260.0

Pannexin 2 (Panx2) is a large-pore ATP-permeable channel with critical roles in various physiological processes, such as the inflammatory response, energy production and apoptosis. Its dysfunction is related to numerous pathological conditions including ischemic brain injury, glioma and glioblastoma multiforme. However, the working mechanism of Panx2 remains unclear. Here, we present the cryo-electron microscopy structure of human Panx2 at a resolution of 3.4 Å. Panx2 structure assembles as a heptamer, forming an exceptionally wide channel pore across the transmembrane and intracellular domains, which is compatible with ATP permeation. Comparing Panx2 with Panx1 structures in different states reveals that the Panx2 structure corresponds to an open channel state. A ring of seven arginine residues located at the extracellular entrance forms the narrowest site of the channel, which serves as the critical molecular filter controlling the permeation of substrate molecules. This is further verified by molecular dynamics simulations and ATP release assays. Our studies reveal the architecture of the Panx2 channel and provide insights into the molecular mechanism of its channel gating.

Zhang Hang, Wang Shiyu, Zhang Zhenzhen, Hou Mengzhuo, Du Chunyu, Zhao Zhenye, Vogel Horst, Li Zhifang, Yan Kaige, Zhang Xiaokang, Lu Jianping, Liang Yujie, Yuan Shuguang, Wang Daping, Zhang Huawei

2023-Mar-03

Public Health Public Health

Artificial intelligence (AI) for breast cancer screening: BreastScreen population-based cohort study of cancer detection.

In EBioMedicine

BACKGROUND : Artificial intelligence (AI) has been proposed to reduce false-positive screens, increase cancer detection rates (CDRs), and address resourcing challenges faced by breast screening programs. We compared the accuracy of AI versus radiologists in real-world population breast cancer screening, and estimated potential impacts on CDR, recall and workload for simulated AI-radiologist reading.

METHODS : External validation of a commercially-available AI algorithm in a retrospective cohort of 108,970 consecutive mammograms from a population-based screening program, with ascertained outcomes (including interval cancers by registry linkage). Area under the ROC curve (AUC), sensitivity and specificity for AI were compared with radiologists who interpreted the screens in practice. CDR and recall were estimated for simulated AI-radiologist reading (with arbitration) and compared with program metrics.

FINDINGS : The AUC for AI was 0.83 compared with 0.93 for radiologists. At a prospective threshold, sensitivity for AI (0.67; 95% CI: 0.64-0.70) was comparable to radiologists (0.68; 95% CI: 0.66-0.71) with lower specificity (0.81 [95% CI: 0.81-0.81] versus 0.97 [95% CI: 0.97-0.97]). Recall rate for AI-radiologist reading (3.14%) was significantly lower than for the BSWA program (3.38%) (-0.25%; 95% CI: -0.31 to -0.18; P < 0.001). CDR was also lower (6.37 versus 6.97 per 1000) (-0.61; 95% CI: -0.77 to -0.44; P < 0.001); however, AI detected interval cancers that were not found by radiologists (0.72 per 1000; 95% CI: 0.57-0.90). AI-radiologist reading increased arbitration but decreased overall screen-reading volume by 41.4% (95% CI: 41.2-41.6).

INTERPRETATION : Replacement of one radiologist by AI (with arbitration) resulted in lower recall and overall screen-reading volume. There was a small reduction in CDR for AI-radiologist reading. AI detected interval cases that were not identified by radiologists, suggesting potentially higher CDR if radiologists were unblinded to AI findings. These results indicate AI's potential role as a screen-reader of mammograms, but prospective trials are required to determine whether CDR could improve if AI detection was actioned in double-reading with arbitration.

FUNDING : National Breast Cancer Foundation (NBCF), National Health and Medical Research Council (NHMRC).

Marinovich M Luke, Wylie Elizabeth, Lotter William, Lund Helen, Waddell Andrew, Madeley Carolyn, Pereira Gavin, Houssami Nehmat

2023-Feb-28

Artificial intelligence, Breast neoplasms, Diagnostic screening programs, Sensitivity and specificity

General General

Assessment of Natural Language Processing of Electronic Health Records to Measure Goals-of-Care Discussions as a Clinical Trial Outcome.

In JAMA network open

IMPORTANCE : Many clinical trial outcomes are documented in free-text electronic health records (EHRs), making manual data collection costly and infeasible at scale. Natural language processing (NLP) is a promising approach for measuring such outcomes efficiently, but ignoring NLP-related misclassification may lead to underpowered studies.

OBJECTIVE : To evaluate the performance, feasibility, and power implications of using NLP to measure the primary outcome of EHR-documented goals-of-care discussions in a pragmatic randomized clinical trial of a communication intervention.

DESIGN, SETTING, AND PARTICIPANTS : This diagnostic study compared the performance, feasibility, and power implications of measuring EHR-documented goals-of-care discussions using 3 approaches: (1) deep-learning NLP, (2) NLP-screened human abstraction (manual verification of NLP-positive records), and (3) conventional manual abstraction. The study included hospitalized patients aged 55 years or older with serious illness enrolled between April 23, 2020, and March 26, 2021, in a pragmatic randomized clinical trial of a communication intervention in a multihospital US academic health system.

MAIN OUTCOMES AND MEASURES : Main outcomes were natural language processing performance characteristics, human abstractor-hours, and misclassification-adjusted statistical power of methods of measuring clinician-documented goals-of-care discussions. Performance of NLP was evaluated with receiver operating characteristic (ROC) curves and precision-recall (PR) analyses and examined the effects of misclassification on power using mathematical substitution and Monte Carlo simulation.

RESULTS : A total of 2512 trial participants (mean [SD] age, 71.7 [10.8] years; 1456 [58%] female) amassed 44 324 clinical notes during 30-day follow-up. In a validation sample of 159 participants, deep-learning NLP trained on a separate training data set from identified patients with documented goals-of-care discussions with moderate accuracy (maximal F1 score, 0.82; area under the ROC curve, 0.924; area under the PR curve, 0.879). Manual abstraction of the outcome from the trial data set would require an estimated 2000 abstractor-hours and would power the trial to detect a risk difference of 5.4% (assuming 33.5% control-arm prevalence, 80% power, and 2-sided α = .05). Measuring the outcome by NLP alone would power the trial to detect a risk difference of 7.6%. Measuring the outcome by NLP-screened human abstraction would require 34.3 abstractor-hours to achieve estimated sensitivity of 92.6% and would power the trial to detect a risk difference of 5.7%. Monte Carlo simulations corroborated misclassification-adjusted power calculations.

CONCLUSIONS AND RELEVANCE : In this diagnostic study, deep-learning NLP and NLP-screened human abstraction had favorable characteristics for measuring an EHR outcome at scale. Adjusted power calculations accurately quantified power loss from NLP-related misclassification, suggesting that incorporation of this approach into the design of studies using NLP would be beneficial.

Lee Robert Y, Kross Erin K, Torrence Janaki, Li Kevin S, Sibley James, Cohen Trevor, Lober William B, Engelberg Ruth A, Curtis J Randall

2023-Mar-01

Dermatology Dermatology

A deep-learning algorithm to classify skin lesions from mpox virus infection.

In Nature medicine ; h5-index 170.0

Undetected infection and delayed isolation of infected individuals are key factors driving the monkeypox virus (now termed mpox virus or MPXV) outbreak. To enable earlier detection of MPXV infection, we developed an image-based deep convolutional neural network (named MPXV-CNN) for the identification of the characteristic skin lesions caused by MPXV. We assembled a dataset of 139,198 skin lesion images, split into training/validation and testing cohorts, comprising non-MPXV images (n = 138,522) from eight dermatological repositories and MPXV images (n = 676) from the scientific literature, news articles, social media and a prospective cohort of the Stanford University Medical Center (n = 63 images from 12 patients, all male). In the validation and testing cohorts, the sensitivity of the MPXV-CNN was 0.83 and 0.91, the specificity was 0.965 and 0.898 and the area under the curve was 0.967 and 0.966, respectively. In the prospective cohort, the sensitivity was 0.89. The classification performance of the MPXV-CNN was robust across various skin tones and body regions. To facilitate the usage of the algorithm, we developed a web-based app by which the MPXV-CNN can be accessed for patient guidance. The capability of the MPXV-CNN for identifying MPXV lesions has the potential to aid in MPXV outbreak mitigation.

Thieme Alexander H, Zheng Yuanning, Machiraju Gautam, Sadee Chris, Mittermaier Mirja, Gertler Maximilian, Salinas Jorge L, Srinivasan Krithika, Gyawali Prashnna, Carrillo-Perez Francisco, Capodici Angelo, Uhlig Maximilian, Habenicht Daniel, Löser Anastassia, Kohler Maja, Schuessler Maximilian, Kaul David, Gollrad Johannes, Ma Jackie, Lippert Christoph, Billick Kendall, Bogoch Isaac, Hernandez-Boussard Tina, Geldsetzer Pascal, Gevaert Olivier

2023-Mar-02

General General

Sub-continental-scale carbon stocks of individual trees in African drylands.

In Nature ; h5-index 368.0

The distribution of dryland trees and their density, cover, size, mass and carbon content are not well known at sub-continental to continental scales1-14. This information is important for ecological protection, carbon accounting, climate mitigation and restoration efforts of dryland ecosystems15-18. We assessed more than 9.9 billion trees derived from more than 300,000 satellite images, covering semi-arid sub-Saharan Africa north of the Equator. We attributed wood, foliage and root carbon to every tree in the 0-1,000 mm year-1 rainfall zone by coupling field data19, machine learning20-22, satellite data and high-performance computing. Average carbon stocks of individual trees ranged from 0.54 Mg C ha-1 and 63 kg C tree-1 in the arid zone to 3.7 Mg C ha-1 and 98 kg tree-1 in the sub-humid zone. Overall, we estimated the total carbon for our study area to be 0.84 (±19.8%) Pg C. Comparisons with 14 previous TRENDY numerical simulation studies23 for our area found that the density and carbon stocks of scattered trees have been underestimated by three models and overestimated by 11 models, respectively. This benchmarking can help understand the carbon cycle and address concerns about land degradation24-29. We make available a linked database of wood mass, foliage mass, root mass and carbon stock of each tree for scientists, policymakers, dryland-restoration practitioners and farmers, who can use it to estimate farmland tree carbon stocks from tablets or laptops.

Tucker Compton, Brandt Martin, Hiernaux Pierre, Kariryaa Ankit, Rasmussen Kjeld, Small Jennifer, Igel Christian, Reiner Florian, Melocik Katherine, Meyer Jesse, Sinno Scott, Romero Eric, Glennie Erin, Fitts Yasmin, Morin August, Pinzon Jorge, McClain Devin, Morin Paul, Porter Claire, Loeffler Shane, Kergoat Laurent, Issoufou Bil-Assanou, Savadogo Patrice, Wigneron Jean-Pierre, Poulter Benjamin, Ciais Philippe, Kaufmann Robert, Myneni Ranga, Saatchi Sassan, Fensholt Rasmus

2023-Mar

General General

Prediction of transition state structures of gas-phase chemical reactions via machine learning.

In Nature communications ; h5-index 260.0

The elucidation of transition state (TS) structures is essential for understanding the mechanisms of chemical reactions and exploring reaction networks. Despite significant advances in computational approaches, TS searching remains a challenging problem owing to the difficulty of constructing an initial structure and heavy computational costs. In this paper, a machine learning (ML) model for predicting the TS structures of general organic reactions is proposed. The proposed model derives the interatomic distances of a TS structure from atomic pair features reflecting reactant, product, and linearly interpolated structures. The model exhibits excellent accuracy, particularly for atomic pairs in which bond formation or breakage occurs. The predicted TS structures yield a high success ratio (93.8%) for quantum chemical saddle point optimizations, and 88.8% of the optimization results have energy errors of less than 0.1 kcal mol-1. Additionally, as a proof of concept, the exploration of multiple reaction paths of an organic reaction is demonstrated based on ML inferences. I envision that the proposed approach will aid in the construction of initial geometries for TS optimization and reaction path exploration.

Choi Sunghwan

2023-Mar-01

General General

Automatic and accurate ligand structure determination guided by cryo-electron microscopy maps.

In Nature communications ; h5-index 260.0

Advances in cryo-electron microscopy (cryoEM) and deep-learning guided protein structure prediction have expedited structural studies of protein complexes. However, methods for accurately determining ligand conformations are lacking. In this manuscript, we develop EMERALD, a tool for automatically determining ligand structures guided by medium-resolution cryoEM density. We show this method is robust at predicting ligands along with surrounding side chains in maps as low as 4.5 Å local resolution. Combining this with a measure of placement confidence and running on all protein/ligand structures in the EMDB, we show that 57% of ligands replicate the deposited model, 16% confidently find alternate conformations, 22% have ambiguous density where multiple conformations might be present, and 5% are incorrectly placed. For five cases where our approach finds an alternate conformation with high confidence, high-resolution crystal structures validate our placement. EMERALD and the resulting analysis should prove critical in using cryoEM to solve protein-ligand complexes.

Muenks Andrew, Zepeda Samantha, Zhou Guangfeng, Veesler David, DiMaio Frank

2023-Mar-01

oncology Oncology

Circulating tumor DNA reveals complex biological features with clinical relevance in metastatic breast cancer.

In Nature communications ; h5-index 260.0

Liquid biopsy has proven valuable in identifying individual genetic alterations; however, the ability of plasma ctDNA to capture complex tumor phenotypes with clinical value is unknown. To address this question, we have performed 0.5X shallow whole-genome sequencing in plasma from 459 patients with metastatic breast cancer, including 245 patients treated with endocrine therapy and a CDK4/6 inhibitor (ET + CDK4/6i) from 2 independent cohorts. We demonstrate that machine learning multi-gene signatures, obtained from ctDNA, identify complex biological features, including measures of tumor proliferation and estrogen receptor signaling, similar to what is accomplished using direct tumor tissue DNA or RNA profiling. More importantly, 4 DNA-based subtypes, and a ctDNA-based genomic signature tracking retinoblastoma loss-of-heterozygosity, are significantly associated with poor response and survival outcome following ET + CDK4/6i, independently of plasma tumor fraction. Our approach opens opportunities for the discovery of additional multi-feature genomic predictors coming from ctDNA in breast cancer and other cancer-types.

Prat Aleix, Brasó-Maristany Fara, Martínez-Sáez Olga, Sanfeliu Esther, Xia Youli, Bellet Meritxell, Galván Patricia, Martínez Débora, Pascual Tomás, Marín-Aguilera Mercedes, Rodríguez Anna, Chic Nuria, Adamo Barbara, Paré Laia, Vidal Maria, Margelí Mireia, Ballana Ester, Gómez-Rey Marina, Oliveira Mafalda, Felip Eudald, Matito Judit, Sánchez-Bayona Rodrigo, Suñol Anna, Saura Cristina, Ciruelos Eva, Tolosa Pablo, Muñoz Montserrat, González-Farré Blanca, Villagrasa Patricia, Parker Joel S, Perou Charles M, Vivancos Ana

2023-Mar-01

Public Health Public Health

Development and validation of echocardiography-based machine-learning models to predict mortality.

In EBioMedicine

BACKGROUND : Echocardiography (echo) based machine learning (ML) models may be useful in identifying patients at high-risk of all-cause mortality.

METHODS : We developed ML models (ResNet deep learning using echo videos and CatBoost gradient boosting using echo measurements) to predict 1-year, 3-year, and 5-year mortality. Models were trained on the Mackay dataset, Taiwan (6083 echos, 3626 patients) and validated in the Alberta HEART dataset, Canada (997 echos, 595 patients). We examined the performance of the models overall, and in subgroups (healthy controls, at risk of heart failure (HF), HF with reduced ejection fraction (HFrEF) and HF with preserved ejection fraction (HFpEF)). We compared the models' performance to the MAGGIC risk score, and examined the correlation between the models' predicted probability of death and baseline quality of life as measured by the Kansas City Cardiomyopathy Questionnaire (KCCQ).

FINDINGS : Mortality rates at 1-, 3- and 5-years were 14.9%, 28.6%, and 42.5% in the Mackay cohort, and 3.0%, 10.3%, and 18.7%, in the Alberta HEART cohort. The ResNet and CatBoost models achieved area under the receiver-operating curve (AUROC) between 85% and 92% in internal validation. In external validation, the AUROCs for the ResNet (82%, 82%, and 78%) were significantly better than CatBoost (78%, 73%, and 75%), for 1-, 3- and 5-year mortality prediction respectively, with better or comparable performance to the MAGGIC score. ResNet models predicted higher probability of death in the HFpEF and HFrEF (30%-50%) subgroups than in controls and at risk patients (5%-20%). The predicted probabilities of death correlated with KCCQ scores (all p < 0.05).

INTERPRETATION : Echo-based ML models to predict mortality had good internal and external validity, were generalizable, correlated with patients' quality of life, and are comparable to an established HF risk score. These models can be leveraged for automated risk stratification at point-of-care.

FUNDING : Funding for Alberta HEART was provided by an Alberta Innovates - Health Solutions Interdisciplinary Team Grant no. AHFMRITG 200801018. P.K. holds a Canadian Institutes of Health Research (CIHR) Sex and Gender Science Chair and a Heart & Stroke Foundation Chair in Cardiovascular Research. A.V. and V.S. received funding from the Mitacs Globalink Research Internship.

Valsaraj Akshay, Kalmady Sunil Vasu, Sharma Vaibhav, Frost Matthew, Sun Weijie, Sepehrvand Nariman, Ong Marcus, Equibec Cyril, Dyck Jason R B, Anderson Todd, Becher Harald, Weeks Sarah, Tromp Jasper, Hung Chung-Lieh, Ezekowitz Justin A, Kaul Padma

2023-Feb-27

Deep learning, Echocardiography, Functional status, Heart failure, Machine learning, Mortality, Prognostic models

General General

Identifying subtypes of chronic kidney disease with machine learning: development, internal validation and prognostic validation using linked electronic health records in 350,067 individuals.

In EBioMedicine

BACKGROUND : Although chronic kidney disease (CKD) is associated with high multimorbidity, polypharmacy, morbidity and mortality, existing classification systems (mild to severe, usually based on estimated glomerular filtration rate, proteinuria or urine albumin-creatinine ratio) and risk prediction models largely ignore the complexity of CKD, its risk factors and its outcomes. Improved subtype definition could improve prediction of outcomes and inform effective interventions.

METHODS : We analysed individuals ≥18 years with incident and prevalent CKD (n = 350,067 and 195,422 respectively) from a population-based electronic health record resource (2006-2020; Clinical Practice Research Datalink, CPRD). We included factors (n = 264 with 2670 derived variables), e.g. demography, history, examination, blood laboratory values and medications. Using a published framework, we identified subtypes through seven unsupervised machine learning (ML) methods (K-means, Diana, HC, Fanny, PAM, Clara, Model-based) with 66 (of 2670) variables in each dataset. We evaluated subtypes for: (i) internal validity (within dataset, across methods); (ii) prognostic validity (predictive accuracy for 5-year all-cause mortality and admissions); and (iii) medications (new and existing by British National Formulary chapter).

FINDINGS : After identifying five clusters across seven approaches, we labelled CKD subtypes: 1. Early-onset, 2. Late-onset, 3. Cancer, 4. Metabolic, and 5. Cardiometabolic. Internal validity: We trained a high performing model (using XGBoost) that could predict disease subtypes with 95% accuracy for incident and prevalent CKD (Sensitivity: 0.81-0.98, F1 score:0.84-0.97). Prognostic validity: 5-year all-cause mortality, hospital admissions, and incidence of new chronic diseases differed across CKD subtypes. The 5-year risk of mortality and admissions in the overall incident CKD population were highest in cardiometabolic subtype: 43.3% (42.3-42.8%) and 29.5% (29.1-30.0%), respectively, and lowest in the early-onset subtype: 5.7% (5.5-5.9%) and 18.7% (18.4-19.1%).

MEDICATIONS : Across CKD subtypes, the distribution of prescription medication classes at baseline varied, with highest medication burden in cardiometabolic and metabolic subtypes, and higher burden in prevalent than incident CKD.

INTERPRETATION : In the largest CKD study using ML, to-date, we identified five distinct subtypes in individuals with incident and prevalent CKD. These subtypes have relevance to study of aetiology, therapeutics and risk prediction.

FUNDING : AstraZeneca UK Ltd, Health Data Research UK.

Dashtban Ashkan, Mizani Mehrdad A, Pasea Laura, Denaxas Spiros, Corbett Richard, Mamza Jil B, Gao He, Morris Tamsin, Hemingway Harry, Banerjee Amitava

2023-Feb-27

CKD subtype, Cluster analysis, Machine learning, Survival analysis, Unsupervised clustering

General General

Inferring protein fitness landscapes from laboratory evolution experiments.

In PLoS computational biology

Directed laboratory evolution applies iterative rounds of mutation and selection to explore the protein fitness landscape and provides rich information regarding the underlying relationships between protein sequence, structure, and function. Laboratory evolution data consist of protein sequences sampled from evolving populations over multiple generations and this data type does not fit into established supervised and unsupervised machine learning approaches. We develop a statistical learning framework that models the evolutionary process and can infer the protein fitness landscape from multiple snapshots along an evolutionary trajectory. We apply our modeling approach to dihydrofolate reductase (DHFR) laboratory evolution data and the resulting landscape parameters capture important aspects of DHFR structure and function. We use the resulting model to understand the structure of the fitness landscape and find numerous examples of epistasis but an overall global peak that is evolutionarily accessible from most starting sequences. Finally, we use the model to perform an in silico extrapolation of the DHFR laboratory evolution trajectory and computationally design proteins from future evolutionary rounds.

D’Costa Sameer, Hinds Emily C, Freschlin Chase R, Song Hyebin, Romero Philip A

2023-Mar-01

oncology Oncology

Benchmarking omics-based prediction of asthma development in children.

In Respiratory research ; h5-index 45.0

BACKGROUND : Asthma is a heterogeneous disease with high morbidity. Advancement in high-throughput multi-omics approaches has enabled the collection of molecular assessments at different layers, providing a complementary perspective of complex diseases. Numerous computational methods have been developed for the omics-based patient classification or disease outcome prediction. Yet, a systematic benchmarking of those methods using various combinations of omics data for the prediction of asthma development is still lacking.

OBJECTIVE : We aimed to investigate the computational methods in disease status prediction using multi-omics data.

METHOD : We systematically benchmarked 18 computational methods using all the 63 combinations of six omics data (GWAS, miRNA, mRNA, microbiome, metabolome, DNA methylation) collected in The Vitamin D Antenatal Asthma Reduction Trial (VDAART) cohort. We evaluated each method using standard performance metrics for each of the 63 omics combinations.

RESULTS : Our results indicate that overall Logistic Regression, Multi-Layer Perceptron, and MOGONET display superior performance, and the combination of transcriptional, genomic and microbiome data achieves the best prediction. Moreover, we find that including the clinical data can further improve the prediction performance for some but not all the omics combinations.

CONCLUSIONS : Specific omics combinations can reach the optimal prediction of asthma development in children. And certain computational methods showed superior performance than other methods.

Wang Xu-Wen, Wang Tong, Schaub Darius P, Chen Can, Sun Zheng, Ke Shanlin, Hecker Julian, Maaser-Hecker Anna, Zeleznik Oana A, Zeleznik Roman, Litonjua Augusto A, DeMeo Dawn L, Lasky-Su Jessica, Silverman Edwin K, Liu Yang-Yu, Weiss Scott T

2023-Feb-26

Asthma, Disease status, Multi-omics, Prediction

General General

A framework to identify ethical concerns with ML-guided care workflows: a case study of mortality prediction to guide advance care planning.

In Journal of the American Medical Informatics Association : JAMIA

OBJECTIVE : Identifying ethical concerns with ML applications to healthcare (ML-HCA) before problems arise is now a stated goal of ML design oversight groups and regulatory agencies. Lack of accepted standard methodology for ethical analysis, however, presents challenges. In this case study, we evaluate use of a stakeholder "values-collision" approach to identify consequential ethical challenges associated with an ML-HCA for advanced care planning (ACP). Identification of ethical challenges could guide revision and improvement of the ML-HCA.

MATERIALS AND METHODS : We conducted semistructured interviews of the designers, clinician-users, affiliated administrators, and patients, and inductive qualitative analysis of transcribed interviews using modified grounded theory.

RESULTS : Seventeen stakeholders were interviewed. Five "values-collisions"-where stakeholders disagreed about decisions with ethical implications-were identified: (1) end-of-life workflow and how model output is introduced; (2) which stakeholders receive predictions; (3) benefit-harm trade-offs; (4) whether the ML design team has a fiduciary relationship to patients and clinicians; and, (5) how and if to protect early deployment research from external pressures, like news scrutiny, before research is completed.

DISCUSSION : From these findings, the ML design team prioritized: (1) alternative workflow implementation strategies; (2) clarification that prediction was only evaluated for ACP need, not other mortality-related ends; and (3) shielding research from scrutiny until endpoint driven studies were completed.

CONCLUSION : In this case study, our ethical analysis of this ML-HCA for ACP was able to identify multiple sites of intrastakeholder disagreement that mark areas of ethical and value tension. These findings provided a useful initial ethical screening.

Cagliero Diana, Deuitch Natalie, Shah Nigam, Feudtner Chris, Char Danton

2023-Feb-24

artificial intelligence, clinical, end-of-life care, ethics, machine learning, palliative care

General General

Field-dependent deep learning enables high-throughput whole-cell 3D super-resolution imaging.

In Nature methods ; h5-index 152.0

Single-molecule localization microscopy in a typical wide-field setup has been widely used for investigating subcellular structures with super resolution; however, field-dependent aberrations restrict the field of view (FOV) to only tens of micrometers. Here, we present a deep-learning method for precise localization of spatially variant point emitters (FD-DeepLoc) over a large FOV covering the full chip of a modern sCMOS camera. Using a graphic processing unit-based vectorial point spread function (PSF) fitter, we can fast and accurately model the spatially variant PSF of a high numerical aperture objective in the entire FOV. Combined with deformable mirror-based optimal PSF engineering, we demonstrate high-accuracy three-dimensional single-molecule localization microscopy over a volume of ~180 × 180 × 5 μm3, allowing us to image mitochondria and nuclear pore complexes in entire cells in a single imaging cycle without hardware scanning; a 100-fold increase in throughput compared to the state of the art.

Fu Shuang, Shi Wei, Luo Tingdan, He Yingchuan, Zhou Lulu, Yang Jie, Yang Zhichao, Liu Jiadong, Liu Xiaotian, Guo Zhiyong, Yang Chengyu, Liu Chao, Huang Zhen-Li, Ries Jonas, Zhang Mingjie, Xi Peng, Jin Dayong, Li Yiming

2023-Feb-23

oncology Oncology

Targeted plasma proteomics reveals signatures discriminating COVID-19 from sepsis with pneumonia.

In Respiratory research ; h5-index 45.0

BACKGROUND : COVID-19 remains a major public health challenge, requiring the development of tools to improve diagnosis and inform therapeutic decisions. As dysregulated inflammation and coagulation responses have been implicated in the pathophysiology of COVID-19 and sepsis, we studied their plasma proteome profiles to delineate similarities from specific features.

METHODS : We measured 276 plasma proteins involved in Inflammation, organ damage, immune response and coagulation in healthy controls, COVID-19 patients during acute and convalescence phase, and sepsis patients; the latter included (i) community-acquired pneumonia (CAP) caused by Influenza, (ii) bacterial CAP, (iii) non-pneumonia sepsis, and (iv) septic shock patients.

RESULTS : We identified a core response to infection consisting of 42 proteins altered in both COVID-19 and sepsis, although higher levels of cytokine storm-associated proteins were evident in sepsis. Furthermore, microbiologic etiology and clinical endotypes were linked to unique signatures. Finally, through machine learning, we identified biomarkers, such as TRIM21, PTN and CASP8, that accurately differentiated COVID-19 from CAP-sepsis with higher accuracy than standard clinical markers.

CONCLUSIONS : This study extends the understanding of host responses underlying sepsis and COVID-19, indicating varying disease mechanisms with unique signatures. These diagnostic and severity signatures are candidates for the development of personalized management of COVID-19 and sepsis.

Palma Medina Laura M, Babačić Haris, Dzidic Majda, Parke Åsa, Garcia Marina, Maleki Kimia T, Unge Christian, Lourda Magda, Kvedaraite Egle, Chen Puran, Muvva Jagadeeswara Rao, Cornillet Martin, Emgård Johanna, Moll Kirsten, Michaëlsson Jakob, Flodström-Tullberg Malin, Brighenti Susanna, Buggert Marcus, Mjösberg Jenny, Malmberg Karl-Johan, Sandberg Johan K, Gredmark-Russ Sara, Rooyackers Olav, Svensson Mattias, Chambers Benedict J, Eriksson Lars I, Pernemalm Maria, Björkström Niklas K, Aleman Soo, Ljunggren Hans-Gustaf, Klingström Jonas, Strålin Kristoffer, Norrby-Teglund Anna

2023-Feb-24

COVID-19, Community acquired pneumonia, Olink proximity extension assays, Sepsis, Septic shock

General General

Social complexity, life-history and lineage influence the molecular basis of castes in vespid wasps.

In Nature communications ; h5-index 260.0

A key mechanistic hypothesis for the evolution of division of labour in social insects is that a shared set of genes co-opted from a common solitary ancestral ground plan (a genetic toolkit for sociality) regulates caste differentiation across levels of social complexity. Using brain transcriptome data from nine species of vespid wasps, we test for overlap in differentially expressed caste genes and use machine learning models to predict castes using different gene sets. We find evidence of a shared genetic toolkit across species representing different levels of social complexity. We also find evidence of additional fine-scale differences in predictive gene sets, functional enrichment and rates of gene evolution that are related to level of social complexity, lineage and of colony founding. These results suggest that the concept of a shared genetic toolkit for sociality may be too simplistic to fully describe the process of the major transition to sociality.

Wyatt Christopher Douglas Robert, Bentley Michael Andrew, Taylor Daisy, Favreau Emeline, Brock Ryan Edward, Taylor Benjamin Aaron, Bell Emily, Leadbeater Ellouise, Sumner Seirian

2023-Feb-24

oncology Oncology

Targeted plasma proteomics reveals signatures discriminating COVID-19 from sepsis with pneumonia.

In Respiratory research ; h5-index 45.0

BACKGROUND : COVID-19 remains a major public health challenge, requiring the development of tools to improve diagnosis and inform therapeutic decisions. As dysregulated inflammation and coagulation responses have been implicated in the pathophysiology of COVID-19 and sepsis, we studied their plasma proteome profiles to delineate similarities from specific features.

METHODS : We measured 276 plasma proteins involved in Inflammation, organ damage, immune response and coagulation in healthy controls, COVID-19 patients during acute and convalescence phase, and sepsis patients; the latter included (i) community-acquired pneumonia (CAP) caused by Influenza, (ii) bacterial CAP, (iii) non-pneumonia sepsis, and (iv) septic shock patients.

RESULTS : We identified a core response to infection consisting of 42 proteins altered in both COVID-19 and sepsis, although higher levels of cytokine storm-associated proteins were evident in sepsis. Furthermore, microbiologic etiology and clinical endotypes were linked to unique signatures. Finally, through machine learning, we identified biomarkers, such as TRIM21, PTN and CASP8, that accurately differentiated COVID-19 from CAP-sepsis with higher accuracy than standard clinical markers.

CONCLUSIONS : This study extends the understanding of host responses underlying sepsis and COVID-19, indicating varying disease mechanisms with unique signatures. These diagnostic and severity signatures are candidates for the development of personalized management of COVID-19 and sepsis.

Palma Medina Laura M, Babačić Haris, Dzidic Majda, Parke Åsa, Garcia Marina, Maleki Kimia T, Unge Christian, Lourda Magda, Kvedaraite Egle, Chen Puran, Muvva Jagadeeswara Rao, Cornillet Martin, Emgård Johanna, Moll Kirsten, Michaëlsson Jakob, Flodström-Tullberg Malin, Brighenti Susanna, Buggert Marcus, Mjösberg Jenny, Malmberg Karl-Johan, Sandberg Johan K, Gredmark-Russ Sara, Rooyackers Olav, Svensson Mattias, Chambers Benedict J, Eriksson Lars I, Pernemalm Maria, Björkström Niklas K, Aleman Soo, Ljunggren Hans-Gustaf, Klingström Jonas, Strålin Kristoffer, Norrby-Teglund Anna

2023-Feb-24

COVID-19, Community acquired pneumonia, Olink proximity extension assays, Sepsis, Septic shock

Pathology Pathology

Unstained Tissue Imaging and Virtual Hematoxylin and Eosin Staining of Histologic Whole Slide Images.

In Laboratory investigation; a journal of technical methods and pathology ; h5-index 42.0

Tissue structures, phenotypes, and pathology are routinely investigated based on histology. This includes chemically staining the transparent tissue sections to make them visible to the human eye. Although chemical staining is fast and routine, it permanently alters the tissue and often consumes hazardous reagents. On the other hand, on using adjacent tissue sections for combined measurements, the cell-wise resolution is lost owing to sections representing different parts of the tissue. Hence, techniques providing visual information of the basic tissue structure enabling additional measurements from the exact same tissue section are required. Here we tested unstained tissue imaging for the development of computational hematoxylin and eosin (HE) staining. We used unsupervised deep learning (CycleGAN) and whole slide images of prostate tissue sections to compare the performance of imaging tissue in paraffin, as deparaffinized in air, and as deparaffinized in mounting medium with section thicknesses varying between 3 and 20 μm. We showed that although thicker sections increase the information content of tissue structures in the images, thinner sections generally perform better in providing information that can be reproduced in virtual staining. According to our results, tissue imaged in paraffin and as deparaffinized provides a good overall representation of the tissue for virtually HE-stained images. Further, using a pix2pix model, we showed that the reproduction of overall tissue histology can be clearly improved with image-to-image translation using supervised learning and pixel-wise ground truth. We also showed that virtual HE staining can be used for various tissues and used with both 20× and 40× imaging magnifications. Although the performance and methods of virtual staining need further development, our study provides evidence of the feasibility of whole slide unstained microscopy as a fast, cheap, and feasible approach to producing virtual staining of tissue histology while sparing the exact same tissue section ready for subsequent utilization with follow-up methods at single-cell resolution.

Koivukoski Sonja, Khan Umair, Ruusuvuori Pekka, Latonen Leena

2023-Jan-25

HE staining, computational histology, digital pathology, histology, virtual staining, whole slide imaging (WSI)

Surgery Surgery

NIR-II fluorescence imaging-guided colorectal cancer surgery targeting CEACAM5 by a nanobody.

In EBioMedicine

BACKGROUND : Surgery is the cornerstone of colorectal cancer (CRC) treatment, yet complete removal of the tumour remains a challenge. The second near-infrared window (NIR-II, 1000-1700 nm) fluorescent molecular imaging is a novel technique, which has broad application prospects in tumour surgical navigation. We aimed to evaluate the ability of CEACAM5-targeted probe for CRC recognition and the value of NIR-II imaging-guided CRC resection.

METHODS : We constructed the probe 2D5-IRDye800CW by conjugated anti-CEACAM5 nanobody (2D5) with near-infrared fluorescent dye IRDye800CW. The performance and benefits of 2D5-IRDye800CW at NIR-II were confirmed by imaging experiments in mouse vascular and capillary phantom. Then mouse colorectal cancer subcutaneous tumour model (n = 15), orthotopic model (n = 15), and peritoneal metastasis model (n = 10) were constructed to investigate biodistribution of probe and imaging differences between NIR-I and NIR-II in vivo, and then tumour resection was guided by NIR-II fluorescence. Fresh human colorectal cancer specimens were incubated with 2D5-IRDye800CW to verify its specific targeting ability.

FINDINGS : 2D5-IRDye800CW had an NIR-II fluorescence signal extending to 1600 nm and bound specifically to CEACAM5 with an affinity of 2.29 nM. In vivo imaging, 2D5-IRDye800CW accumulated rapidly in tumour (15 min) and could specifically identify orthotopic colorectal cancer and peritoneal metastases. All tumours were resected under NIR-II fluorescence guidance, even smaller than 2 mm tumours were detected, and NIR-II had a higher tumour-to-background ratio than NIR-I (2.55 ± 0.38, 1.94 ± 0.20, respectively). 2D5-IRDye800CW could precisely identify CEACAM5-positive human colorectal cancer tissue.

INTERPRETATION : 2D5-IRDye800CW combined with NIR-II fluorescence has translational potential as an aid to improve R0 surgery of colorectal cancer.

FUNDINGS : This study was supported by Beijing Natural Science Foundation (JQ19027), the National Key Research and Development Program of China (2017YFA0205200), National Natural Science Foundation of China (NSFC) (61971442, 62027901, 81930053, 92059207, 81227901, 82102236), Beijing Natural Science Foundation (L222054), CAS Youth Interdisciplinary Team (JCTD-2021-08), the Strategic Priority Research Program of the Chinese Academy of Sciences (XDA16021200), the Zhuhai High-level Health Personnel Team Project (Zhuhai HLHPTP201703), the Fundamental Research Funds for the Central Universities (JKF-YG-22-B005) and Capital Clinical Characteristic Application Research (Z181100001718178). The authors would like to acknowledge the instrumental and technical support of the multi-modal biomedical imaging experimental platform, Institute of Automation, Chinese Academy of Sciences.

Guo Xiaoyong, Li Changjian, Jia Xiaohua, Qu Yawei, Li Miaomiao, Cao Caiguang, Zhang Zeyu, Qu Qiaojun, Luo Shuangling, Tang Jianqiang, Liu Haifeng, Hu Zhenhua, Tian Jie

2023-Feb-16

CEACAM5, Colorectal cancer, Molecular imaging, Nanobody, Second near-infrared window

General General

Network of hotspot interactions cluster tau amyloid folds.

In Nature communications ; h5-index 260.0

Cryogenic electron microscopy has revealed unprecedented molecular insight into the conformations of β-sheet-rich protein amyloids linked to neurodegenerative diseases. It remains unknown how a protein can adopt a diversity of folds and form multiple distinct fibrillar structures. Here we develop an in silico alanine scan method to estimate the relative energetic contribution of each amino acid in an amyloid assembly. We apply our method to twenty-seven ex vivo and in vitro fibril structural polymorphs of the microtubule-associated protein tau. We uncover networks of energetically important interactions involving amyloid-forming motifs that stabilize the different fibril folds. We evaluate our predictions in cellular and in vitro aggregation assays. Using a machine learning approach, we classify the structures based on residue energetics to identify distinguishing and unifying features. Our energetic profiling suggests that minimal sequence elements control the stability of tau fibrils, allowing future design of protein sequences that fold into unique structures.

Mullapudi Vishruth, Vaquer-Alicea Jaime, Bommareddy Vaibhav, Vega Anthony R, Ryder Bryan D, White Charles L, Diamond Marc I, Joachimiak Lukasz A

2023-Feb-16

Surgery Surgery

How the UK public views the use of diagnostic decision aids by physicians: a vignette-based experiment.

In Journal of the American Medical Informatics Association : JAMIA

OBJECTIVE : Physicians' low adoption of diagnostic decision aids (DDAs) may be partially due to concerns about patient/public perceptions. We investigated how the UK public views DDA use and factors affecting perceptions.

MATERIALS AND METHODS : In this online experiment, 730 UK adults were asked to imagine attending a medical appointment where the doctor used a computerized DDA. The DDA recommended a test to rule out serious disease. We varied the test's invasiveness, the doctor's adherence to DDA advice, and the severity of the patient's disease. Before disease severity was revealed, respondents indicated how worried they felt. Both before [t1] and after [t2] severity was revealed, we measured satisfaction with the consultation, likelihood of recommending the doctor, and suggested frequency of DDA use.

RESULTS : At both timepoints, satisfaction and likelihood of recommending the doctor increased when the doctor adhered to DDA advice (P ≤ .01), and when the DDA suggested an invasive versus noninvasive test (P ≤ .05). The effect of adherence to DDA advice was stronger when participants were worried (P ≤ .05), and the disease turned out to be serious (P ≤ .01). Most respondents felt that DDAs should be used by doctors "sparingly" (34%[t1]/29%[t2]), "frequently," (43%[t1]/43%[t2]) or "always" (17%[t1]/21%[t2]).

DISCUSSION : People are more satisfied when doctors adhere to DDA advice, especially when worried, and when it helps to spot serious disease. Having to undergo an invasive test does not appear to dampen satisfaction.

CONCLUSION : Positive attitudes regarding DDA use and satisfaction with doctors adhering to DDA advice could encourage greater use of DDAs in consultations.

Nurek Martine, Kostopoulou Olga

2023-Feb-16

artificial intelligence, decision aids, diagnosis, health regulatory focus, individual differences, trust in physicians, worry

General General

The 2022 n2c2/UW shared task on extracting social determinants of health.

In Journal of the American Medical Informatics Association : JAMIA

OBJECTIVE : The n2c2/UW SDOH Challenge explores the extraction of social determinant of health (SDOH) information from clinical notes. The objectives include the advancement of natural language processing (NLP) information extraction techniques for SDOH and clinical information more broadly. This article presents the shared task, data, participating teams, performance results, and considerations for future work.

MATERIALS AND METHODS : The task used the Social History Annotated Corpus (SHAC), which consists of clinical text with detailed event-based annotations for SDOH events, such as alcohol, drug, tobacco, employment, and living situation. Each SDOH event is characterized through attributes related to status, extent, and temporality. The task includes 3 subtasks related to information extraction (Subtask A), generalizability (Subtask B), and learning transfer (Subtask C). In addressing this task, participants utilized a range of techniques, including rules, knowledge bases, n-grams, word embeddings, and pretrained language models (LM).

RESULTS : A total of 15 teams participated, and the top teams utilized pretrained deep learning LM. The top team across all subtasks used a sequence-to-sequence approach achieving 0.901 F1 for Subtask A, 0.774 F1 Subtask B, and 0.889 F1 for Subtask C.

CONCLUSIONS : Similar to many NLP tasks and domains, pretrained LM yielded the best performance, including generalizability and learning transfer. An error analysis indicates extraction performance varies by SDOH, with lower performance achieved for conditions, like substance use and homelessness, which increase health risks (risk factors) and higher performance achieved for conditions, like substance abstinence and living with family, which reduce health risks (protective factors).

Lybarger Kevin, Yetisgen Meliha, Uzuner Özlem

2023-Feb-16

data mining, electronic health records, machine learning, natural language processing, social determinants of health

Surgery Surgery

A hierarchical multilabel graph attention network method to predict the deterioration paths of chronic hepatitis B patients.

In Journal of the American Medical Informatics Association : JAMIA

OBJECTIVE : Estimating the deterioration paths of chronic hepatitis B (CHB) patients is critical for physicians' decisions and patient management. A novel, hierarchical multilabel graph attention-based method aims to predict patient deterioration paths more effectively. Applied to a CHB patient data set, it offers strong predictive utilities and clinical value.

MATERIALS AND METHODS : The proposed method incorporates patients' responses to medications, diagnosis event sequences, and outcome dependencies to estimate deterioration paths. From the electronic health records maintained by a major healthcare organization in Taiwan, we collect clinical data about 177 959 patients diagnosed with hepatitis B virus infection. We use this sample to evaluate the proposed method's predictive efficacy relative to 9 existing methods, as measured by precision, recall, F-measure, and area under the curve (AUC).

RESULTS : We use 20% of the sample as holdouts to test each method's prediction performance. The results indicate that our method consistently and significantly outperforms all benchmark methods. It attains the highest AUC, with a 4.8% improvement over the best-performing benchmark, as well as 20.9% and 11.4% improvements in precision and F-measures, respectively. The comparative results demonstrate that our method is more effective for predicting CHB patients' deterioration paths than existing predictive methods.

DISCUSSION AND CONCLUSION : The proposed method underscores the value of patient-medication interactions, temporal sequential patterns of distinct diagnosis, and patient outcome dependencies for capturing dynamics that underpin patient deterioration over time. Its efficacious estimates grant physicians a more holistic view of patient progressions and can enhance their clinical decision-making and patient management.

Wu Zejian Eric, Xu Da, Hu Paul Jen-Hwa, Huang Ting-Shuo

2023-Feb-15

chronic hepatitis B patients, deep learning, deterioration path predictions, graph attention network, predictive analytics

General General

Algorithmic encoding of protected characteristics in chest X-ray disease detection models.

In EBioMedicine

BACKGROUND : It has been rightfully emphasized that the use of AI for clinical decision making could amplify health disparities. An algorithm may encode protected characteristics, and then use this information for making predictions due to undesirable correlations in the (historical) training data. It remains unclear how we can establish whether such information is actually used. Besides the scarcity of data from underserved populations, very little is known about how dataset biases manifest in predictive models and how this may result in disparate performance. This article aims to shed some light on these issues by exploring methodology for subgroup analysis in image-based disease detection models.

METHODS : We utilize two publicly available chest X-ray datasets, CheXpert and MIMIC-CXR, to study performance disparities across race and biological sex in deep learning models. We explore test set resampling, transfer learning, multitask learning, and model inspection to assess the relationship between the encoding of protected characteristics and disease detection performance across subgroups.

FINDINGS : We confirm subgroup disparities in terms of shifted true and false positive rates which are partially removed after correcting for population and prevalence shifts in the test sets. We find that transfer learning alone is insufficient for establishing whether specific patient information is used for making predictions. The proposed combination of test-set resampling, multitask learning, and model inspection reveals valuable insights about the way protected characteristics are encoded in the feature representations of deep neural networks.

INTERPRETATION : Subgroup analysis is key for identifying performance disparities of AI models, but statistical differences across subgroups need to be taken into account when analyzing potential biases in disease detection. The proposed methodology provides a comprehensive framework for subgroup analysis enabling further research into the underlying causes of disparities.

FUNDING : European Research Council Horizon 2020, UK Research and Innovation.

Glocker Ben, Jones Charles, Bernhardt Mélanie, Winzeck Stefan

2023-Feb-13

Algorithmic bias, Artificial intelligence, Image-based disease detection, Subgroup disparities

Surgery Surgery

Data-driven longitudinal characterization of neonatal health and morbidity.

In Science translational medicine ; h5-index 138.0

Although prematurity is the single largest cause of death in children under 5 years of age, the current definition of prematurity, based on gestational age, lacks the precision needed for guiding care decisions. Here, we propose a longitudinal risk assessment for adverse neonatal outcomes in newborns based on a deep learning model that uses electronic health records (EHRs) to predict a wide range of outcomes over a period starting shortly before conception and ending months after birth. By linking the EHRs of the Lucile Packard Children's Hospital and the Stanford Healthcare Adult Hospital, we developed a cohort of 22,104 mother-newborn dyads delivered between 2014 and 2018. Maternal and newborn EHRs were extracted and used to train a multi-input multitask deep learning model, featuring a long short-term memory neural network, to predict 24 different neonatal outcomes. An additional cohort of 10,250 mother-newborn dyads delivered at the same Stanford Hospitals from 2019 to September 2020 was used to validate the model. Areas under the receiver operating characteristic curve at delivery exceeded 0.9 for 10 of the 24 neonatal outcomes considered and were between 0.8 and 0.9 for 7 additional outcomes. Moreover, comprehensive association analysis identified multiple known associations between various maternal and neonatal features and specific neonatal outcomes. This study used linked EHRs from more than 30,000 mother-newborn dyads and would serve as a resource for the investigation and prediction of neonatal outcomes. An interactive website is available for independent investigators to leverage this unique dataset: https://maternal-child-health-associations.shinyapps.io/shiny_app/.

De Francesco Davide, Reiss Jonathan D, Roger Jacquelyn, Tang Alice S, Chang Alan L, Becker Martin, Phongpreecha Thanaphong, Espinosa Camilo, Morin Susanna, Berson Eloïse, Thuraiappah Melan, Le Brian L, Ravindra Neal G, Payrovnaziri Seyedeh Neelufar, Mataraso Samson, Kim Yeasul, Xue Lei, Rosenstein Melissa G, Oskotsky Tomiko, Marić Ivana, Gaudilliere Brice, Carvalho Brendan, Bateman Brian T, Angst Martin S, Prince Lawrence S, Blumenfeld Yair J, Benitz William E, Fuerch Janene H, Shaw Gary M, Sylvester Karl G, Stevenson David K, Sirota Marina, Aghaeepour Nima

2023-Feb-15

Public Health Public Health

The gut microbiome and early-life growth in a population with high prevalence of stunting.

In Nature communications ; h5-index 260.0

Stunting affects one-in-five children globally and is associated with greater infectious morbidity, mortality and neurodevelopmental deficits. Recent evidence suggests that the early-life gut microbiome affects child growth through immune, metabolic and endocrine pathways. Using whole metagenomic sequencing, we map the assembly of the gut microbiome in 335 children from rural Zimbabwe from 1-18 months of age who were enrolled in the Sanitation, Hygiene, Infant Nutrition Efficacy Trial (SHINE; NCT01824940), a randomized trial of improved water, sanitation and hygiene (WASH) and infant and young child feeding (IYCF). Here, we show that the early-life gut microbiome undergoes programmed assembly that is unresponsive to the randomized interventions intended to improve linear growth. However, maternal HIV infection is associated with over-diversification and over-maturity of the early-life gut microbiome in their uninfected children, in addition to reduced abundance of Bifidobacterium species. Using machine learning models (XGBoost), we show that taxonomic microbiome features are poorly predictive of child growth, however functional metagenomic features, particularly B-vitamin and nucleotide biosynthesis pathways, moderately predict both attained linear and ponderal growth and growth velocity. New approaches targeting the gut microbiome in early childhood may complement efforts to combat child undernutrition.

Robertson Ruairi C, Edens Thaddeus J, Carr Lynnea, Mutasa Kuda, Gough Ethan K, Evans Ceri, Geum Hyun Min, Baharmand Iman, Gill Sandeep K, Ntozini Robert, Smith Laura E, Chasekwa Bernard, Majo Florence D, Tavengwa Naume V, Mutasa Batsirai, Francis Freddy, Tome Joice, Stoltzfus Rebecca J, Humphrey Jean H, Prendergast Andrew J, Manges Amee R

2023-Feb-14

Radiology Radiology

Deep Learning-based Approach to Predict Pulmonary Function at Chest CT.

In Radiology ; h5-index 91.0

Background Low-dose chest CT screening is recommended for smokers with the potential for lung function abnormality, but its role in predicting lung function remains unclear. Purpose To develop a deep learning algorithm to predict pulmonary function with low-dose CT images in participants using health screening services. Materials and Methods In this retrospective study, participants underwent health screening with same-day low-dose CT and pulmonary function testing with spirometry at a university affiliated tertiary referral general hospital between January 2015 and December 2018. The data set was split into a development set (model training, validation, and internal test sets) and temporally independent test set according to first visit year. A convolutional neural network was trained to predict the forced expiratory volume in the first second of expiration (FEV1) and forced vital capacity (FVC) from low-dose CT. The mean absolute error and concordance correlation coefficient (CCC) were used to evaluate agreement between spirometry as the reference standard and deep-learning prediction as the index test. FVC and FEV1 percent predicted (hereafter, FVC% and FEV1%) values less than 80% and percent of FVC exhaled in first second (hereafter, FEV1/FVC) less than 70% were used to classify participants at high risk. Results A total of 16 148 participants were included (mean age, 55 years ± 10 [SD]; 10 981 men) and divided into a development set (n = 13 428) and temporally independent test set (n = 2720). In the temporally independent test set, the mean absolute error and CCC were 0.22 L and 0.94, respectively, for FVC and 0.22 L and 0.91 for FEV1. For the prediction of the respiratory high-risk group, FVC%, FEV1%, and FEV1/FVC had respective accuracies of 89.6% (2436 of 2720 participants; 95% CI: 88.4, 90.7), 85.9% (2337 of 2720 participants; 95% CI: 84.6, 87.2), and 90.2% (2453 of 2720 participants; 95% CI: 89.1, 91.3) in the same testing data set. The sensitivities were 61.6% (242 of 393 participants; 95% CI: 59.7, 63.4), 46.9% (226 of 482 participants; 95% CI: 45.0, 48.8), and 36.1% (91 of 252 participants; 95% CI: 34.3, 37.9), respectively. Conclusion A deep learning model applied to volumetric chest CT predicted pulmonary function with relatively good performance. © RSNA, 2023 Supplemental material is available for this article.

Park Hyunjung, Yun Jihye, Lee Sang Min, Hwang Hye Jeon, Seo Joon Beom, Jung Young Ju, Hwang Jeongeun, Lee Se Hee, Lee Sei Won, Kim Namkug

2023-Feb-14

General General

AI reveals insights into link between CD33 and cognitive impairment in Alzheimer's Disease.

In PLoS computational biology

Modeling biological mechanisms is a key for disease understanding and drug-target identification. However, formulating quantitative models in the field of Alzheimer's Disease is challenged by a lack of detailed knowledge of relevant biochemical processes. Additionally, fitting differential equation systems usually requires time resolved data and the possibility to perform intervention experiments, which is difficult in neurological disorders. This work addresses these challenges by employing the recently published Variational Autoencoder Modular Bayesian Networks (VAMBN) method, which we here trained on combined clinical and patient level gene expression data while incorporating a disease focused knowledge graph. Our approach, called iVAMBN, resulted in a quantitative model that allowed us to simulate a down-expression of the putative drug target CD33, including potential impact on cognitive impairment and brain pathophysiology. Experimental validation demonstrated a high overlap of molecular mechanism predicted to be altered by CD33 perturbation with cell line data. Altogether, our modeling approach may help to select promising drug targets.

Raschka Tamara, Sood Meemansa, Schultz Bruce, Altay Aybuge, Ebeling Christian, Fröhlich Holger

2023-Feb-13

Pathology Pathology

Predicting 1p/19q co-deletion status from magnetic resonance imaging using deep learning in adult-type diffuse lower-grade gliomas: a discovery and validation study.

In Laboratory investigation; a journal of technical methods and pathology ; h5-index 42.0

Determination of 1p/19q co-deletion status is important for the classification, prognostication, and personalized therapy in diffuse lower-grade gliomas (LGG). We developed and validated a deep learning imaging signature (DLIS) from preoperative magnetic resonance imaging (MRI) for predicting the 1p/19q status in patients with LGG. The DLIS was constructed on a training dataset (n = 330) and validated on both an internal validation dataset (n = 123) and a public TCIA dataset (n = 102). The receiver operating characteristic (ROC) analysis and precision recall curves (PRC) were used to measure the classification performance. The area under ROC curves (AUC) of the DLIS was 0.999 for training dataset, 0.986 for validation dataset, and 0.983 for testing dataset. The F1-score of the prediction model was 0.992 for training dataset, 0.940 for validation dataset, and 0.925 for testing dataset. Our data suggests that DLIS could be used to predict the 1p/19q status from preoperative imaging in patients with LGG. The imaging-based deep learning has the potential to be a noninvasive tool predictive of molecular markers in adult diffuse gliomas. The authors developed a deep learning model predictive of 1p/19q status from preoperative imaging in 555 lower-grade gliomas (LGG), and achieved an area under the curve (AUC) of 0.983 in the testing dataset. They reveal that developing deep learning imaging signatures could be a noninvasive tool for predicting molecular markers in LGG.

Yan Jing, Zhang Shenghai, Sun Qiuchang, Wang Weiwei, Duan Wenchao, Wang Li, Ding Tianqing, Pei Dongling, Sun Chen, Wang Wenqing, Liu Zhen, Hong Xuanke, Wang Xiangxiang, Guo Yu, Li Wencai, Cheng Jingliang, Liu Xianzhi, Li Zhi-Cheng, Zhang Zhenyu

2022-Feb

General General

A convolutional neural network model for survival prediction based on prognosis-related cascaded Wx feature selection.

In Laboratory investigation; a journal of technical methods and pathology ; h5-index 42.0

Great advances in deep learning have provided effective solutions for prediction tasks in the biomedical field. However, accurate prognosis prediction using cancer genomics data remains challenging due to the severe overfitting problem caused by curse of dimensionality inherent to high-throughput sequencing data. Moreover, there are unique challenges to perform survival analysis, arising from the difficulty in utilizing censored samples whose events of interest are not observed. Convolutional neural network (CNN) models provide us the opportunity to extract meaningful hierarchical features to characterize cancer subtype and prognosis outcomes. On the other hand, feature selection can mitigate overfitting and reduce subsequent model training computation burden by screening out significant genes from redundant genes. To accomplish model simplification, we developed a concise and efficient survival analysis model, named CNN-Cox model, which combines a special CNN framework with prognosis-related feature selection cascaded Wx, with the advantage of less computation demand utilizing light training parameters. Experiment results show that CNN-Cox model achieved consistent higher C-index values and better survival prediction performance across seven cancer type datasets in The Cancer Genome Atlas cohort, including bladder carcinoma, head and neck squamous cell carcinoma, kidney renal cell carcinoma, brain low-grade glioma, lung adenocarcinoma (LUAD), lung squamous cell carcinoma, and skin cutaneous melanoma, compared with the existing state-of-the-art survival analysis methods. As an illustration of model interpretation, we examined potential prognostic gene signatures of LUAD dataset using the proposed CNN-Cox model. We conducted protein-protein interaction network analysis to identify potential prognostic genes and further analyzed the biological function of 13 hub genes, including ANLN, RACGAP1, KIF4A, KIF20A, KIF14, ASPM, CDK1, SPC25, NCAPG, MKI67, HJURP, EXO1, HMMR, whose high expression is significantly associated with poor survival of LUAD patients. These findings confirmed that CNN-Cox model is effective in extracting not only prognosis factors but also biologically meaningful gene features. The codes are available at the GitHub website: https://github.com/wangwangCCChen/CNN-Cox.

Yin Qingyan, Chen Wangwang, Zhang Chunxia, Wei Zhi

2022-Oct

General General

Modeling CRISPR-Cas13d on-target and off-target effects using machine learning approaches.

In Nature communications ; h5-index 260.0

A major challenge in the application of the CRISPR-Cas13d system is to accurately predict its guide-dependent on-target and off-target effect. Here, we perform CRISPR-Cas13d proliferation screens and design a deep learning model, named DeepCas13, to predict the on-target activity from guide sequences and secondary structures. DeepCas13 outperforms existing methods to predict the efficiency of guides targeting both protein-coding and non-coding RNAs. Guides targeting non-essential genes display off-target viability effects, which are closely related to their on-target efficiencies. Choosing proper negative control guides during normalization mitigates the associated false positives in proliferation screens. We apply DeepCas13 to the guides targeting lncRNAs, and identify lncRNAs that affect cell viability and proliferation in multiple cell lines. The higher prediction accuracy of DeepCas13 over existing methods is extensively confirmed via a secondary CRISPR-Cas13d screen and quantitative RT-PCR experiments. DeepCas13 is freely accessible via http://deepcas13.weililab.org .

Cheng Xiaolong, Li Zexu, Shan Ruocheng, Li Zihan, Wang Shengnan, Zhao Wenchang, Zhang Han, Chao Lumen, Peng Jian, Fei Teng, Li Wei

2023-Feb-10

General General

Multi-omics and machine learning reveal context-specific gene regulatory activities of PML::RARA in acute promyelocytic leukemia.

In Nature communications ; h5-index 260.0

The PML::RARA fusion protein is the hallmark driver of Acute Promyelocytic Leukemia (APL) and disrupts retinoic acid signaling, leading to wide-scale gene expression changes and uncontrolled proliferation of myeloid precursor cells. While known to be recruited to binding sites across the genome, its impact on gene regulation and expression is under-explored. Using integrated multi-omics datasets, we characterize the influence of PML::RARA binding on gene expression and regulation in an inducible PML::RARA cell line model and APL patient ex vivo samples. We find that genes whose regulatory elements recruit PML::RARA are not uniformly transcriptionally repressed, as commonly suggested, but also may be upregulated or remain unchanged. We develop a computational machine learning implementation called Regulatory Element Behavior Extraction Learning to deconvolute the complex, local transcription factor binding site environment at PML::RARA bound positions to reveal distinct signatures that modulate how PML::RARA directs the transcriptional response.

Villiers William, Kelly Audrey, He Xiaohan, Kaufman-Cook James, Elbasir Abdurrahman, Bensmail Halima, Lavender Paul, Dillon Richard, Mifsud Borbála, Osborne Cameron S

2023-Feb-09

Radiology Radiology

Altered global signal topography in Alzheimer's disease.

In EBioMedicine

BACKGROUND : Alzheimer's disease (AD) is a neurodegenerative disease associated with widespread disruptions in intrinsic local specialization and global integration in the functional system of the brain. These changes in integration may further disrupt the global signal (GS) distribution, which might represent the local relative contribution to global activity in functional magnetic resonance imaging (fMRI).

METHODS : fMRI scans from a discovery dataset (n = 809) and a validated dataset (n = 542) were used in the analysis. We investigated the alteration of GS topography using the GS correlation (GSCORR) in patients with mild cognitive impairment (MCI) and AD. The association between GS alterations and functional network properties was also investigated based on network theory. The underlying mechanism of GSCORR alterations was elucidated using imaging-transcriptomics.

FINDINGS : Significantly increased GS topography in the frontal lobe and decreased GS topography in the hippocampus, cingulate gyrus, caudate, and middle temporal gyrus were observed in patients with AD (Padj < 0.05). Notably, topographical GS changes in these regions correlated with cognitive ability (P < 0.05). The changes in GS topography also correlated with the changes in functional network segregation (ρ = 0.5). Moreover, the genes identified based on GS topographical changes were enriched in pathways associated with AD and neurodegenerative diseases.

INTERPRETATION : Our findings revealed significant changes in GS topography and its molecular basis, confirming the informative role of GS in AD and further contributing to the understanding of the relationship between global and local neuronal activities in patients with AD.

FUNDING : Beijing Natural Science Funds for Distinguished Young Scholars, China; Fundamental Research Funds for the Central Universities, China; National Natural Science Foundation, China.

Chen Pindong, Zhao Kun, Zhang Han, Wei Yongbin, Wang Pan, Wang Dawei, Song Chengyuan, Yang Hongwei, Zhang Zengqiang, Yao Hongxiang, Qu Yida, Kang Xiaopeng, Du Kai, Fan Lingzhong, Han Tong, Yu Chunshui, Zhou Bo, Jiang Tianzi, Zhou Yuying, Lu Jie, Han Ying, Zhang Xi, Liu Bing, Liu Yong

2023-Feb-07

“Alzheimers disease”, Functional network, Global signal, Transcriptomics

Surgery Surgery

UCseek: ultrasensitive early detection and recurrence monitoring of urothelial carcinoma by shallow-depth genome-wide bisulfite sequencing of urinary sediment DNA.

In EBioMedicine

BACKGROUND : Current methods for the detection and surveillance of urothelial carcinomas (UCs) are often invasive, costly, and not effective for low-grade, early-stage, and minimal residual disease (MRD) tumors. We aimed to develop and validate a model from urine sediments to predict different grade and stage UCs with low cost and high accuracy.

METHODS : We collected 167 samples, including 90 tumors and 77 individuals without tumors, as a discovery cohort. We assessed copy number variations and methylation values for them and constructed a diagnostic classifier to detect UC, UCseek, by using an individual read-based method and support vector machine. The performance of UCseek was validated in an independent cohort derived from three hospitals (n = 206) and a relapse cohort (n = 42) for monitoring recurrence.

FINDINGS : We constructed UCseek, which could predict UCs with high sensitivity (92.7%), high specificity (90.7%), and high accuracy (91.7%) in the independent validation set. The accuracy of UCseek in low-grade and early-stage patients reached 91.8% and 94.3%, respectively. Notably, UCseek retained great performance at ultralow sequencing depths (0.3X-0.5X). It also demonstrated a powerful ability to monitor recurrence in a surveillance cohort compared with cystoscopy (90.91% vs. 59.09%).

INTERPRETATION : We optimized an improved approach named UCseek for the noninvasive diagnosis and monitoring of UCs in both low- and high-grade tumors and in early- and advanced-stage tumors, even at ultralow sequencing depths, which may reduce the burden of cystoscopy and blind second surgery.

FUNDING : A full list of funding bodies that contributed to this study can be found in the Acknowledgments section.

Wang Ping, Shi Yue, Zhang Jianye, Shou Jianzhong, Zhang Mingxin, Zou Daojia, Liang Yuan, Li Juan, Tan Yezhen, Zhang Mei, Bi Xingang, Zhou Liqun, Ci Weimin, Li Xuesong

2023-Feb-07

Diagnosis, Machine learning, Molecular diagnostics, Relapse monitoring, Tumor markers, Urothelial carcinoma

Ophthalmology Ophthalmology

Neurologic Dysfunction Assessment in Parkinson Disease Based on Fundus Photographs Using Deep Learning.

In JAMA ophthalmology ; h5-index 58.0

IMPORTANCE : Until now, other than complex neurologic tests, there have been no readily accessible and reliable indicators of neurologic dysfunction among patients with Parkinson disease (PD). This study was conducted to determine the role of fundus photography as a noninvasive and readily available tool for assessing neurologic dysfunction among patients with PD using deep learning methods.

OBJECTIVE : To develop an algorithm that can predict Hoehn and Yahr (H-Y) scale and Unified Parkinson's Disease Rating Scale part III (UPDRS-III) score using fundus photography among patients with PD.

DESIGN, SETTINGS, AND PARTICIPANTS : This was a prospective decision analytical model conducted at a single tertiary-care hospital. The fundus photographs of participants with PD and participants with non-PD atypical motor abnormalities who visited the neurology department of Kangbuk Samsung Hospital from October 7, 2020, to April 30, 2021, were analyzed in this study. A convolutional neural network was developed to predict both the H-Y scale and UPDRS-III score based on fundus photography findings and participants' demographic characteristics.

MAIN OUTCOMES AND MEASURES : The area under the receiver operating characteristic curve (AUROC) was calculated for sensitivity and specificity analyses for both the internal and external validation data sets.

RESULTS : A total of 615 participants were included in the study: 266 had PD (43.3%; mean [SD] age, 70.8 [8.3] years; 134 male individuals [50.4%]), and 349 had non-PD atypical motor abnormalities (56.7%; mean [SD] age, 70.7 [7.9] years; 236 female individuals [67.6%]). For the internal validation data set, the sensitivity was 83.23% (95% CI, 82.07%-84.38%) and 82.61% (95% CI, 81.38%-83.83%) for the H-Y scale and UPDRS-III score, respectively. The specificity was 66.81% (95% CI, 64.97%-68.65%) and 65.75% (95% CI, 62.56%-68.94%) for the H-Y scale and UPDRS-III score, respectively. For the external validation data set, the sensitivity and specificity were 70.73% (95% CI, 66.30%-75.16%) and 66.66% (95% CI, 50.76%-82.25%), respectively. Lastly, the calculated AUROC and accuracy were 0.67 (95% CI, 0.55-0.79) and 70.45% (95% CI, 66.85%-74.04%), respectively.

CONCLUSIONS AND RELEVANCE : This decision analytical model reveals amalgamative insights into the neurologic dysfunction among PD patients by providing information on how to apply a deep learning method to evaluate the association between the retina and brain. Study data may help clarify recent research findings regarding dopamine pathologic cascades between the retina and brain among patients with PD; however, further research is needed to expand the clinical implication of this algorithm.

Ahn Sangil, Shin Jitae, Song Su Jeong, Yoon Won Tae, Sagong Min, Jeong Areum, Kim Joon Hyung, Yu Hyeong Gon

2023-Feb-09

General General

Association of Recent SARS-CoV-2 Infection With New-Onset Alcohol Use Disorder, January 2020 Through January 2022.

In JAMA network open

IMPORTANCE : The COVID-19 pandemic affects many diseases, including alcohol use disorders (AUDs). As the pandemic evolves, understanding the association of a new diagnosis of AUD with COVID-19 over time is required to mitigate negative consequences.

OBJECTIVE : To examine the association of COVID-19 infection with new diagnosis of AUD over time from January 2020 through January 2022.

DESIGN, SETTING, AND PARTICIPANTS : In this retrospective cohort study of electronic health records of US patients 12 years of age or older, new diagnoses of AUD were compared between patients with COVID-19 and patients with other respiratory infections who had never had COVID-19 by 3-month intervals from January 20, 2020, through January 27, 2022.

EXPOSURES : SARS-CoV-2 infection or non-SARS-CoV-2 respiratory infection.

MAIN OUTCOMES AND MEASURES : New diagnoses of AUD were compared in COVID-19 and propensity score-matched control cohorts by hazard ratios (HRs) and 95% CIs from either 14 days to 3 months or 3 to 6 months after the index event.

RESULTS : This study comprised 1 201 082 patients with COVID-19 (56.9% female patients; 65.7% White; mean [SD] age at index, 46.2 [18.9] years) and 1 620 100 patients with other respiratory infections who had never had COVID-19 (60.4% female patients; 71.1% White; mean [SD] age at index, 44.5 [20.6] years). There was a significantly increased risk of a new diagnosis of AUD in the 3 months after COVID-19 was contracted during the first 3 months of the pandemic (block 1) compared with control cohorts (HR, 2.53 [95% CI, 1.82-3.51]), but the risk decreased to nonsignificance in the next 3 time blocks (April 2020 to January 2021). The risk for AUD diagnosis increased after infection in January to April 2021 (HR, 1.30 [95% CI, 1.08-1.56]) and April to July 2021 (HR, 1.80 [95% CI, 1.47-2.21]). The result became nonsignificant again in blocks 7 and 8 (COVID-19 diagnosis between July 2021 and January 2022). A similar temporal pattern was seen for new diagnosis of AUD 3 to 6 months after infection with COVID-19 vs control index events.

CONCLUSIONS AND RELEVANCE : Elevated risk for AUD after COVID-19 infection compared with non-COVID-19 respiratory infections during some time frames may suggest an association of SARS-CoV-2 infection with the pandemic-associated increase in AUD. However, the lack of excess hazard in most time blocks makes it likely that the circumstances surrounding the pandemic and the fear and anxiety they created also were important factors associated with new diagnoses of AUD.

Olaker Veronica R, Kendall Ellen K, Wang Christina X, Parran Theodore V, Terebuh Pauline, Kaelber David C, Xu Rong, Davis Pamela B

2023-Feb-01

General General

Association of Recent SARS-CoV-2 Infection With New-Onset Alcohol Use Disorder, January 2020 Through January 2022.

In JAMA network open

IMPORTANCE : The COVID-19 pandemic affects many diseases, including alcohol use disorders (AUDs). As the pandemic evolves, understanding the association of a new diagnosis of AUD with COVID-19 over time is required to mitigate negative consequences.

OBJECTIVE : To examine the association of COVID-19 infection with new diagnosis of AUD over time from January 2020 through January 2022.

DESIGN, SETTING, AND PARTICIPANTS : In this retrospective cohort study of electronic health records of US patients 12 years of age or older, new diagnoses of AUD were compared between patients with COVID-19 and patients with other respiratory infections who had never had COVID-19 by 3-month intervals from January 20, 2020, through January 27, 2022.

EXPOSURES : SARS-CoV-2 infection or non-SARS-CoV-2 respiratory infection.

MAIN OUTCOMES AND MEASURES : New diagnoses of AUD were compared in COVID-19 and propensity score-matched control cohorts by hazard ratios (HRs) and 95% CIs from either 14 days to 3 months or 3 to 6 months after the index event.

RESULTS : This study comprised 1 201 082 patients with COVID-19 (56.9% female patients; 65.7% White; mean [SD] age at index, 46.2 [18.9] years) and 1 620 100 patients with other respiratory infections who had never had COVID-19 (60.4% female patients; 71.1% White; mean [SD] age at index, 44.5 [20.6] years). There was a significantly increased risk of a new diagnosis of AUD in the 3 months after COVID-19 was contracted during the first 3 months of the pandemic (block 1) compared with control cohorts (HR, 2.53 [95% CI, 1.82-3.51]), but the risk decreased to nonsignificance in the next 3 time blocks (April 2020 to January 2021). The risk for AUD diagnosis increased after infection in January to April 2021 (HR, 1.30 [95% CI, 1.08-1.56]) and April to July 2021 (HR, 1.80 [95% CI, 1.47-2.21]). The result became nonsignificant again in blocks 7 and 8 (COVID-19 diagnosis between July 2021 and January 2022). A similar temporal pattern was seen for new diagnosis of AUD 3 to 6 months after infection with COVID-19 vs control index events.

CONCLUSIONS AND RELEVANCE : Elevated risk for AUD after COVID-19 infection compared with non-COVID-19 respiratory infections during some time frames may suggest an association of SARS-CoV-2 infection with the pandemic-associated increase in AUD. However, the lack of excess hazard in most time blocks makes it likely that the circumstances surrounding the pandemic and the fear and anxiety they created also were important factors associated with new diagnoses of AUD.

Olaker Veronica R, Kendall Ellen K, Wang Christina X, Parran Theodore V, Terebuh Pauline, Kaelber David C, Xu Rong, Davis Pamela B

2023-Feb-01

General General

Dissecting cell identity via network inference and in silico gene perturbation.

In Nature ; h5-index 368.0

Cell identity is governed by the complex regulation of gene expression, represented as gene-regulatory networks1. Here we use gene-regulatory networks inferred from single-cell multi-omics data to perform in silico transcription factor perturbations, simulating the consequent changes in cell identity using only unperturbed wild-type data. We apply this machine-learning-based approach, CellOracle, to well-established paradigms-mouse and human haematopoiesis, and zebrafish embryogenesis-and we correctly model reported changes in phenotype that occur as a result of transcription factor perturbation. Through systematic in silico transcription factor perturbation in the developing zebrafish, we simulate and experimentally validate a previously unreported phenotype that results from the loss of noto, an established notochord regulator. Furthermore, we identify an axial mesoderm regulator, lhx1a. Together, these results show that CellOracle can be used to analyse the regulation of cell identity by transcription factors, and can provide mechanistic insights into development and differentiation.

Kamimoto Kenji, Stringa Blerta, Hoffmann Christy M, Jindal Kunal, Solnica-Krezel Lilianna, Morris Samantha A

2023-Feb-08

oncology Oncology

Cartography of Genomic Interactions Enables Deep Analysis of Single-Cell Expression Data.

In Nature communications ; h5-index 260.0

Remarkable advances in single cell genomics have presented unique challenges and opportunities for interrogating a wealth of biomedical inquiries. High dimensional genomic data are inherently complex because of intertwined relationships among the genes. Existing methods, including emerging deep learning-based approaches, do not consider the underlying biological characteristics during data processing, which greatly compromises the performance of data analysis and hinders the maximal utilization of state-of-the-art genomic techniques. In this work, we develop an entropy-based cartography strategy to contrive the high dimensional gene expression data into a configured image format, referred to as genomap, with explicit integration of the genomic interactions. This unique cartography casts the gene-gene interactions into the spatial configuration of genomaps and enables us to extract the deep genomic interaction features and discover underlying discriminative patterns of the data. We show that, for a wide variety of applications (cell clustering and recognition, gene signature extraction, single cell data integration, cellular trajectory analysis, dimensionality reduction, and visualization), the proposed approach drastically improves the accuracies of data analyses as compared to the state-of-the-art techniques.

Islam Md Tauhidul, Xing Lei

2023-Feb-08

Radiology Radiology

Comparison of Chest Radiograph Captions Based on Natural Language Processing vs Completed by Radiologists.

In JAMA network open

IMPORTANCE : Artificial intelligence (AI) can interpret abnormal signs in chest radiography (CXR) and generate captions, but a prospective study is needed to examine its practical value.

OBJECTIVE : To prospectively compare natural language processing (NLP)-generated CXR captions and the diagnostic findings of radiologists.

DESIGN, SETTING, AND PARTICIPANTS : A multicenter diagnostic study was conducted. The training data set included CXR images and reports retrospectively collected from February 1, 2014, to February 28, 2018. The retrospective test data set included consecutive images and reports from April 1 to July 31, 2019. The prospective test data set included consecutive images and reports from May 1 to September 30, 2021.

EXPOSURES : A bidirectional encoder representation from a transformers model was used to extract language entities and relationships from unstructured CXR reports to establish 23 labels of abnormal signs to train convolutional neural networks. The participants in the prospective test group were randomly assigned to 1 of 3 different caption generation models: a normal template, NLP-generated captions, and rule-based captions based on convolutional neural networks. For each case, a resident drafted the report based on the randomly assigned captions and an experienced radiologist finalized the report blinded to the original captions. A total of 21 residents and 19 radiologists were involved.

MAIN OUTCOMES AND MEASURES : Time to write reports based on different caption generation models.

RESULTS : The training data set consisted of 74 082 cases (39 254 [53.0%] women; mean [SD] age, 50.0 [17.1] years). In the retrospective (n = 8126; 4345 [53.5%] women; mean [SD] age, 47.9 [15.9] years) and prospective (n = 5091; 2416 [47.5%] women; mean [SD] age, 45.1 [15.6] years) test data sets, the mean (SD) area under the curve of abnormal signs was 0.87 (0.11) in the retrospective data set and 0.84 (0.09) in the prospective data set. The residents' mean (SD) reporting time using the NLP-generated model was 283 (37) seconds-significantly shorter than the normal template (347 [58] seconds; P < .001) and the rule-based model (296 [46] seconds; P < .001). The NLP-generated captions showed the highest similarity to the final reports with a mean (SD) bilingual evaluation understudy score of 0.69 (0.24)-significantly higher than the normal template (0.37 [0.09]; P < .001) and the rule-based model (0.57 [0.19]; P < .001).

CONCLUSIONS AND RELEVANCE : In this diagnostic study of NLP-generated CXR captions, prior information provided by NLP was associated with greater efficiency in the reporting process, while maintaining good consistency with the findings of radiologists.

Zhang Yaping, Liu Mingqian, Zhang Lu, Wang Lingyun, Zhao Keke, Hu Shundong, Chen Xu, Xie Xueqian

2023-Feb-01

Radiology Radiology

AI Improves Nodule Detection on Chest Radiographs in a Health Screening Population: A Randomized Controlled Trial.

In Radiology ; h5-index 91.0

Background The impact of artificial intelligence (AI)-based computer-aided detection (CAD) software has not been prospectively explored in real-world populations. Purpose To investigate whether commercial AI-based CAD software could improve the detection rate of actionable lung nodules on chest radiographs in participants undergoing health checkups. Materials and Methods In this single-center, pragmatic, open-label randomized controlled trial, participants who underwent chest radiography between July 2020 and December 2021 in a health screening center were enrolled and randomized into intervention (AI group) and control (non-AI group) arms. One of three designated radiologists with 13-36 years of experience interpreted each radiograph, referring to the AI-based CAD results for the AI group. The primary outcome was the detection rate, that is, the number of true-positive radiographs divided by the total number of radiographs, of actionable lung nodules confirmed on CT scans obtained within 3 months. Actionable nodules were defined as solid nodules larger than 8 mm or subsolid nodules with a solid portion larger than 6 mm (Lung Imaging Reporting and Data System, or Lung-RADS, category 4). Secondary outcomes included the positive-report rate, sensitivity, false-referral rate, and malignant lung nodule detection rate. Clinical outcomes were compared between the two groups using univariable logistic regression analyses. Results A total of 10 476 participants (median age, 59 years [IQR, 50-66 years]; 5121 men) were randomized to an AI group (n = 5238) or non-AI group (n = 5238). The trial met the predefined primary outcome, demonstrating an improved detection rate of actionable nodules in the AI group compared with the non-AI group (0.59% [31 of 5238 participants] vs 0.25% [13 of 5238 participants], respectively; odds ratio, 2.4; 95% CI: 1.3, 4.7; P = .008). The detection rate for malignant lung nodules was higher in the AI group compared with the non-AI group (0.15% [eight of 5238 participants] vs 0.0% [0 of 5238 participants], respectively; P = .008). The AI and non-AI groups showed similar false-referral rates (45.9% [56 of 122 participants] vs 56.0% [56 of 100 participants], respectively; P = .14) and positive-report rates (2.3% [122 of 5238 participants] vs 1.9% [100 of 5238 participants]; P = .14). Conclusion In health checkup participants, artificial intelligence-based software improved the detection of actionable lung nodules on chest radiographs. © RSNA, 2023 Supplemental material is available for this article. See also the editorial by Auffermann in this isssue.

Nam Ju Gang, Hwang Eui Jin, Kim Jayoun, Park Nanhee, Lee Eun Hee, Kim Hyun Jin, Nam Miyeon, Lee Jong Hyuk, Park Chang Min, Goo Jin Mo

2023-Feb-07

Radiology Radiology

Beyond Breast Density: Risk Measures for Breast Cancer in Multiple Imaging Modalities.

In Radiology ; h5-index 91.0

Breast density is an independent risk factor for breast cancer. In digital mammography and digital breast tomosynthesis, breast density is assessed visually using the four-category scale developed by the American College of Radiology Breast Imaging Reporting and Data System (5th edition as of November 2022). Epidemiologically based risk models, such as the Tyrer-Cuzick model (version 8), demonstrate superior modeling performance when mammographic density is incorporated. Beyond just density, a separate mammographic measure of breast cancer risk is parenchymal textural complexity. With advancements in radiomics and deep learning, mammographic textural patterns can be assessed quantitatively and incorporated into risk models. Other supplemental screening modalities, such as breast US and MRI, offer independent risk measures complementary to those derived from mammography. Breast US allows the two components of fibroglandular tissue (stromal and glandular) to be visualized separately in a manner that is not possible with mammography. A higher glandular component at screening breast US is associated with higher risk. With MRI, a higher background parenchymal enhancement of the fibroglandular tissue has also emerged as an imaging marker for risk assessment. Imaging markers observed at mammography, US, and MRI are powerful tools in refining breast cancer risk prediction, beyond mammographic density alone.

Acciavatti Raymond J, Lee Su Hyun, Reig Beatriu, Moy Linda, Conant Emily F, Kontos Despina, Moon Woo Kyung

2023-Feb-07

Ophthalmology Ophthalmology

Multicenter Evaluation of AI-generated DIR and PSIR for Cortical and Juxtacortical Multiple Sclerosis Lesion Detection.

In Radiology ; h5-index 91.0

Background Cortical multiple sclerosis lesions are clinically relevant but inconspicuous at conventional clinical MRI. Double inversion recovery (DIR) and phase-sensitive inversion recovery (PSIR) are more sensitive but often unavailable. In the past 2 years, artificial intelligence (AI) was used to generate DIR and PSIR from standard clinical sequences (eg, T1-weighted, T2-weighted, and fluid-attenuated inversion-recovery sequences), but multicenter validation is crucial for further implementation. Purpose To evaluate cortical and juxtacortical multiple sclerosis lesion detection for diagnostic and disease monitoring purposes on AI-generated DIR and PSIR images compared with MRI-acquired DIR and PSIR images in a multicenter setting. Materials and Methods Generative adversarial networks were used to generate AI-based DIR (n = 50) and PSIR (n = 43) images. The number of detected lesions between AI-generated images and MRI-acquired (reference) images was compared by randomized blinded scoring by seven readers (all with >10 years of experience in lesion assessment). Reliability was expressed as the intraclass correlation coefficient (ICC). Differences in lesion subtype were determined using Wilcoxon signed-rank tests. Results MRI scans of 202 patients with multiple sclerosis (mean age, 46 years ± 11 [SD]; 127 women) were retrospectively collected from seven centers (February 2020 to January 2021). In total, 1154 lesions were detected on AI-generated DIR images versus 855 on MRI-acquired DIR images (mean difference per reader, 35.0% ± 22.8; P < .001). On AI-generated PSIR images, 803 lesions were detected versus 814 on MRI-acquired PSIR images (98.9% ± 19.4; P = .87). Reliability was good for both DIR (ICC, 0.81) and PSIR (ICC, 0.75) across centers. Regionally, more juxtacortical lesions were detected on AI-generated DIR images than on MRI-acquired DIR images (495 [42.9%] vs 338 [39.5%]; P < .001). On AI-generated PSIR images, fewer juxtacortical lesions were detected than on MRI-acquired PSIR images (232 [28.9%] vs 282 [34.6%]; P = .02). Conclusion Artificial intelligence-generated double inversion-recovery and phase-sensitive inversion-recovery images performed well compared with their MRI-acquired counterparts and can be considered reliable in a multicenter setting, with good between-reader and between-center interpretative agreement. © RSNA, 2023 Supplemental material is available for this article. See also the editorial by Zivadinov and Dwyer in this issue.

Bouman Piet M, Noteboom Samantha, Nobrega Santos Fernando A, Beck Erin S, Bliault Gregory, Castellaro Marco, Calabrese Massimiliano, Chard Declan T, Eichinger Paul, Filippi Massimo, Inglese Matilde, Lapucci Caterina, Marciniak Andrzej, Moraal Bastiaan, Morales Pinzon Alfredo, Mühlau Mark, Preziosa Paolo, Reich Daniel S, Rocca Maria A, Schoonheim Menno M, Twisk Jos W R, Wiestler Benedict, Jonkman Laura E, Guttmann Charles R G, Geurts Jeroen J G, Steenwijk Martijn D

2023-Feb-07

Pathology Pathology

Deepfake Histologic Images for Enhancing Digital Pathology.

In Laboratory investigation; a journal of technical methods and pathology ; h5-index 42.0

A pathologist's optical microscopic examination of thinly cut, stained tissue on glass slides prepared from a formalin-fixed paraffin-embedded tissue blocks is the gold standard for tissue diagnostics. In addition, the diagnostic abilities and expertise of pathologists is dependent on their direct experience with common and rarer variant morphologies. Recently, deep learning approaches have been used to successfully show a high level of accuracy for such tasks. However, obtaining expert-level annotated images is an expensive and time-consuming task, and artificially synthesized histologic images can prove greatly beneficial. In this study, we present an approach to not only generate histologic images that reproduce the diagnostic morphologic features of common disease but also provide a user ability to generate new and rare morphologies. Our approach involves developing a generative adversarial network model that synthesizes pathology images constrained by class labels. We investigated the ability of this framework in synthesizing realistic prostate and colon tissue images and assessed the utility of these images in augmenting the diagnostic ability of machine learning methods and their usability by a panel of experienced anatomic pathologists. Synthetic data generated by our framework performed similar to real data when training a deep learning model for diagnosis. Pathologists were not able to distinguish between real and synthetic images, and their analyses showed a similar level of interobserver agreement for prostate cancer grading. We extended the approach to significantly more complex images from colon biopsies and showed that the morphology of the complex microenvironment in such tissues can be reproduced. Finally, we present the ability for a user to generate deepfake histologic images using a simple markup of sematic labels.

Falahkheirkhah Kianoush, Tiwari Saumya, Yeh Kevin, Gupta Sounak, Herrera-Hernandez Loren, McCarthy Michael R, Jimenez Rafael E, Cheville John C, Bhargava Rohit

2023-Jan

computer vision, deep learning, deepfake pathology, digital pathology synthetic data

General General

Funneling modulatory peptide design with generative models: Discovery and characterization of disruptors of calcineurin protein-protein interactions.

In PLoS computational biology

Design of peptide binders is an attractive strategy for targeting "undruggable" protein-protein interfaces. Current design protocols rely on the extraction of an initial sequence from one known protein interactor of the target protein, followed by in-silico or in-vitro mutagenesis-based optimization of its binding affinity. Wet lab protocols can explore only a minor portion of the vast sequence space and cannot efficiently screen for other desirable properties such as high specificity and low toxicity, while in-silico design requires intensive computational resources and often relies on simplified binding models. Yet, for a multivalent protein target, dozens to hundreds of natural protein partners already exist in the cellular environment. Here, we describe a peptide design protocol that harnesses this diversity via a machine learning generative model. After identifying putative natural binding fragments by literature and homology search, a compositional Restricted Boltzmann Machine is trained and sampled to yield hundreds of diverse candidate peptides. The latter are further filtered via flexible molecular docking and an in-vitro microchip-based binding assay. We validate and test our protocol on calcineurin, a calcium-dependent protein phosphatase involved in various cellular pathways in health and disease. In a single screening round, we identified multiple 16-length peptides with up to six mutations from their closest natural sequence that successfully interfere with the binding of calcineurin to its substrates. In summary, integrating protein interaction and sequence databases, generative modeling, molecular docking and interaction assays enables the discovery of novel protein-protein interaction modulators.

Tubiana Jérôme, Adriana-Lifshits Lucia, Nissan Michael, Gabay Matan, Sher Inbal, Sova Marina, Wolfson Haim J, Gal Maayan

2023-Feb-02

General General

Assessment of Pediatric Telemedicine Using Remote Physical Examinations With a Mobile Medical Device: A Nonrandomized Controlled Trial.

In JAMA network open

IMPORTANCE : The number of innovations in health care based on the use of platforms, digital devices, apps, and artificial intelligence has grown exponentially in recent years. When used correctly, these technologies allow inequities in access to health care to be addressed by optimizing care and reducing social and geographic barriers. However, most of the technological health care solutions proposed have not undergone rigorous clinical studies.

OBJECTIVE : To assess the concordance between measurements from a remote physical examination using a mobile medical device and measurements from a conventional in-person physical examination.

DESIGN, SETTING, AND PARTICIPANTS : This nonrandomized controlled trial was conducted from January 1 to December 31, 2020. The clinical parameters compared were heart rate; body temperature; heart, lung, and abdominal auscultation; otoscopy; throat and oral examination; and skin examination. A total of 690 patients with clinical stability and various symptoms who were seen in the emergency department of 2 Brazilian pediatric hospitals were eligible to enter this study.

MAIN OUTCOMES AND MEASURES : The primary outcome was concordance between measurements from a telemedicine physical examination using a mobile medical device and measurements from a conventional in-person physical examination. The secondary outcome was the specificity and sensitivity of the digital device, considering the conventional in-person consultation as the gold standard.

RESULTS : Among 690 patients, the median (IQR) age at study entry was 5 (1-9) years; 348 (50.4%) were female, and 331 (48.0%) presented with a chronic disease. Regarding the primary outcome, the concordance values were 90% or greater for skin examination (94% for rash, 100% for hemorrhagic suffusion, and 95% for signs of secondary infection), characteristics of the mucosa (98% for hydration and 97% for coloring), and heart (95% for murmur, 97% for rhythms, and 98% for sounds), lung (91% for adventitious sounds, 97% for vesicular sounds, and 90% for fever), and abdominal (92% for abdominal sounds) auscultations. Concordance values were lower for otoscopy (72% for the ear canal and 86% for the tympanic membrane), throat and oral examination (72%), and rhinoscopy (79% for mucosa and 81% for secretion). The specificity was greater than 70% (ranging from 74.5% for the ear canal to 99.7% for hemorrhagic suffusion) for all variables. The sensitivity was greater than 52% for skin examination (58.0% for rash and 54.8% for signs of secondary infection), throat and oral examination (52.7%), and otoscopy (66.1% for the ear canal and 64.4% for the tympanic membrane).

CONCLUSIONS AND RELEVANCE : In this study, measurements from remote physical examination with a mobile medical device had satisfactory concordance with measurements from in-person physical examination for otoscopy, throat and oral examination, skin examination, and heart and lung auscultation, with limitations regarding heart and lung auscultation in infants and abdominal auscultation in children of all ages. Measurements from remote physical examination via a mobile medical device were comparable with those from in-person physical examination in children older than 2 years. These findings suggest that telemedicine may be an alternative to in-person examination in certain contexts and may help to optimize access to health care services and reduce social and geographic barriers.

TRIAL REGISTRATION : Brazilian Registry of Clinical Trials Identifier: RBR-346ymn.

Wagner Rafaela, Lima Thalita Cecília, Silva Marielen Ribeiro Tavares da, Rabha Anna Clara Pereira, Ricieri Marinei Campos, Fachi Mariana Millan, Afonso Rogério Carballo, Motta Fábio Araújo

2023-Feb-01

Surgery Surgery

Single-cell spatial landscapes of the lung tumour immune microenvironment.

In Nature ; h5-index 368.0

Single-cell technologies have revealed the complexity of the tumour immune microenvironment with unparalleled resolution1-9. Most clinical strategies rely on histopathological stratification of tumour subtypes, yet the spatial context of single-cell phenotypes within these stratified subgroups is poorly understood. Here we apply imaging mass cytometry to characterize the tumour and immunological landscape of samples from 416 patients with lung adenocarcinoma across five histological patterns. We resolve more than 1.6 million cells, enabling spatial analysis of immune lineages and activation states with distinct clinical correlates, including survival. Using deep learning, we can predict with high accuracy those patients who will progress after surgery using a single 1-mm2 tumour core, which could be informative for clinical management following surgical resection. Our dataset represents a valuable resource for the non-small cell lung cancer research community and exemplifies the utility of spatial resolution within single-cell analyses. This study also highlights how artificial intelligence can improve our understanding of microenvironmental features that underlie cancer progression and may influence future clinical practice.

Sorin Mark, Rezanejad Morteza, Karimi Elham, Fiset Benoit, Desharnais Lysanne, Perus Lucas J M, Milette Simon, Yu Miranda W, Maritan Sarah M, Doré Samuel, Pichette Émilie, Enlow William, Gagné Andréanne, Wei Yuhong, Orain Michele, Manem Venkata S K, Rayes Roni, Siegel Peter M, Camilleri-Broët Sophie, Fiset Pierre Olivier, Desmeules Patrice, Spicer Jonathan D, Quail Daniela F, Joubert Philippe, Walsh Logan A

2023-Feb-01

General General

Automated extraction of clinical measures from videos of oculofacial disorders using machine learning: feasibility, validity and reliability.

In Eye (London, England) ; h5-index 41.0

OBJECTIVES : To determine the feasibility, validity and reliability of automatically extracting clinically meaningful eyelid measurements from consumer-grade videos of individuals with oculofacial disorders.

METHODS : A custom computer program was designed to automatically extract clinical measures from consumer-grade videos. This program was applied to publicly available videos of individuals with oculofacial disorders, and age-matched controls. The primary outcomes were margin reflex distance 1 (MRD1) and 2 (MRD2), blink lagophthalmos, and ocular surface area exposure. Test-retest reliability was evaluated using Bland-Altman analysis to compare the agreement in obtained measures between separate videos of the same individual taken within 48 h of each other.

RESULTS : MRD1 was reduced in individuals with ptosis versus controls (2.2 mm versus 3.4 mm, p < 0.001), and increased in individuals with facial nerve palsy (FNP) (3.9 mm, p = 0.049) and thyroid eye disease (TED) (4.1 mm; p = 0.038). Blink lagophthalmos was increased in individuals with FNP (3.7 mm); p < 0.001) and those with TED (0.1 mm, p = 0.003) versus controls (0.0 mm). Ocular surface exposure was reduced in individuals with ptosis compared with controls (12.2 mm2 versus 13.1 mm2; p < 0.001) and increased in TED (13.7 mm2; p 0.002). Bland-Altmann analysis demonstrated 95% limits of agreement for video-derived measures: median MRD1: -1.1 to 1.1 mm; median MRD2: -0.9 to 1.0 mm; blink lagophthalmos: -3.5 to 3.7 mm; and average ocular surface area exposure: -1.6 to 1.6 mm2.

CONCLUSIONS : The presented program is capable of taking consumer grade videos of patients with oculofacial disease and providing clinically meaningful and reliable eyelid measurements that show promising validity.

Schulz Christopher B, Clarke Holly, Makuloluwe Sarith, Thomas Peter B, Kang Swan

2023-Feb-01

General General

Highly-scaled and fully-integrated 3-dimensional ferroelectric transistor array for hardware implementation of neural networks.

In Nature communications ; h5-index 260.0

Hardware-based neural networks (NNs) can provide a significant breakthrough in artificial intelligence applications due to their ability to extract features from unstructured data and learn from them. However, realizing complex NN models remains challenging because different tasks, such as feature extraction and classification, should be performed at different memory elements and arrays. This further increases the required number of memory arrays and chip size. Here, we propose a three-dimensional ferroelectric NAND (3D FeNAND) array for the area-efficient hardware implementation of NNs. Vector-matrix multiplication is successfully demonstrated using the integrated 3D FeNAND arrays, and excellent pattern classification is achieved. By allocating each array of vertical layers in 3D FeNAND as the hidden layer of NN, each layer can be used to perform different tasks, and the classification of color-mixed patterns is achieved. This work provides a practical strategy to realize high-performance and highly efficient NN systems by stacking computation components vertically.

Kim Ik-Jyae, Kim Min-Kyu, Lee Jang-Sik

2023-Jan-31

Radiology Radiology

Development and Validation of a Deep Learning-Based Synthetic Bone-Suppressed Model for Pulmonary Nodule Detection in Chest Radiographs.

In JAMA network open

IMPORTANCE : Dual-energy chest radiography exhibits better sensitivity than single-energy chest radiography, partly due to its ability to remove overlying anatomical structures.

OBJECTIVES : To develop and validate a deep learning-based synthetic bone-suppressed (DLBS) nodule-detection algorithm for pulmonary nodule detection on chest radiographs.

DESIGN, SETTING, AND PARTICIPANTS : This decision analytical modeling study used data from 3 centers between November 2015 and July 2019 from 1449 patients. The DLBS nodule-detection algorithm was trained using single-center data (institute 1) of 998 chest radiographs. The DLBS algorithm was validated using 2 external data sets (institute 2, 246 patients; and institute 3, 205 patients). Statistical analysis was performed from March to December 2021.

EXPOSURES : DLBS nodule-detection algorithm.

MAIN OUTCOMES AND MEASURES : The nodule-detection performance of DLBS model was compared with the convolution neural network nodule-detection algorithm (original model). Reader performance testing was conducted by 3 thoracic radiologists assisted by the DLBS algorithm or not. Sensitivity and false-positive markings per image (FPPI) were compared.

RESULTS : Training data consisted of 998 patients (539 men [54.0%]; mean [SD] age, 54.2 [9.82] years), and 2 external validation data sets consisted of 246 patients (133 men [54.1%]; mean [SD] age, 55.3 [8.7] years) and 205 patients (105 men [51.2%]; mean [SD] age, 51.8 [9.1] years). Using the external validation data set of institute 2, the bone-suppressed model showed higher sensitivity compared with that of the original model for nodule detection (91.5% [109 of 119] vs 79.8% [95 of 119]; P < .001). The overall mean of FPPI with the bone-suppressed model was reduced compared with the original model (0.07 [17 of 246] vs 0.09 [23 of 246]; P < .001). For the observer performance testing with the data of institute 3, the mean sensitivity of 3 radiologists was 77.5% (95% [CI], 69.9%-85.2%), whereas that of radiologists assisted by DLBS modeling was 92.1% (95% CI, 86.3%-97.3%; P < .001). The 3 radiologists had a reduced number of FPPI when assisted by the DLBS model (0.071 [95% CI, 0.041-0.111] vs 0.151 [95% CI, 0.111-0.210]; P < .001).

CONCLUSIONS AND RELEVANCE : This decision analytical modeling study found that the DLBS model was more sensitive to detecting pulmonary nodules on chest radiographs compared with the original model. These findings suggest that the DLBS model could be beneficial to radiologists in the detection of lung nodules in chest radiographs without need of the specialized equipment or increase of radiation dose.

Kim Hwiyoung, Lee Kye Ho, Han Kyunghwa, Lee Ji Won, Kim Jin Young, Im Dong Jin, Hong Yoo Jin, Choi Byoung Wook, Hur Jin

2023-Jan-03

General General

Assessment of the Role of Artificial Intelligence in the Association Between Time of Day and Colonoscopy Quality.

In JAMA network open

IMPORTANCE : Time of day was associated with a decline in adenoma detection during colonoscopy. Artificial intelligence (AI) systems are effective in improving the adenoma detection rate (ADR), but the performance of AI during different times of the day remains unknown.

OBJECTIVE : To validate whether the assistance of an AI system could overcome the time-related decline in ADR during colonoscopy.

DESIGN, SETTING, AND PARTICIPANTS : This cohort study is a secondary analysis of 2 prospective randomized controlled trials (RCT) from Renmin Hospital of Wuhan University. Consecutive patients undergoing colonoscopy were randomly assigned to either the AI-assisted group or unassisted group from June 18, 2019, to September 6, 2019, and July 1, 2020, to October 15, 2020. The ADR of early and late colonoscopy sessions per half day were compared before and after the intervention of the AI system. Data were analyzed from March to June 2022.

EXPOSURE : Conventional colonoscopy or AI-assisted colonoscopy.

MAIN OUTCOMES AND MEASURES : Adenoma detection rate.

RESULTS : A total of 1780 patients (mean [SD] age, 48.61 [13.35] years, 837 [47.02%] women) were enrolled. A total of 1041 procedures (58.48%) were performed in early sessions, with 357 randomized into the unassisted group (34.29%) and 684 into the AI group (65.71%). A total of 739 procedures (41.52%) were performed in late sessions, with 263 randomized into the unassisted group (35.59%) and 476 into the AI group (64.41%). In the unassisted group, the ADR in early sessions was significantly higher compared with that of late sessions (13.73% vs 5.70%; P = .005; OR, 2.42; 95% CI, 1.31-4.47). After the intervention of the AI system, as expected, no statistically significant difference was found (22.95% vs 22.06%, P = .78; OR, 0.96; 95% CI; 0.71-1.29). Furthermore, the AI systems showed better assistance ability on ADR in late sessions compared with early sessions (odds ratio, 3.81; 95% CI, 2.10-6.91 vs 1.60; 95% CI, 1.10-2.34).

CONCLUSIONS AND RELEVANCE : In this cohort study, AI systems showed higher assistance ability in late sessions per half day, which suggests the potential to maintain high quality and homogeneity of colonoscopies and further improve endoscopist performance in large screening programs and centers with high workloads.

Lu Zihua, Zhang Lihui, Yao Liwen, Gong Dexin, Wu Lianlian, Xia Meiqing, Zhang Jun, Zhou Wei, Huang Xu, He Chunping, Wu Huiling, Zhang Chenxia, Li Xun, Yu Honggang

2023-Jan-03

General General

Machine Learning-Based Prediction of Hospital Admission Among Children in an Emergency Care Center.

In Pediatric emergency care

OBJECTIVES : Machine learning-based prediction of hospital admissions may have the potential to optimize patient disposition and improve clinical outcomes by minimizing both undertriage and overtriage in crowded emergency care. We developed and validated the predictive abilities of machine learning-based predictions of hospital admissions in a pediatric emergency care center.

METHODS : A prognostic study was performed using retrospectively collected data of children younger than 16 years who visited a single pediatric emergency care center in Osaka, Japan, between August 1, 2016, and October 15, 2019. Generally, the center treated walk-in children and did not treat trauma injuries. The main outcome was hospital admission as determined by the physician. The 83 potential predictors available at presentation were selected from the following categories: demographic characteristics, triage level, physiological parameters, and symptoms. To identify predictive abilities for hospital admission, maximize the area under the precision-recall curve, and address imbalanced outcome classes, we developed the following models for the preperiod training cohort (67% of the samples) and also used them in the 1-year postperiod validation cohort (33% of the samples): (1) logistic regression, (2) support vector machine, (3) random forest, and (4) extreme gradient boosting.

RESULTS : Among 88,283 children who were enrolled, the median age was 3.9 years, with 47,931 (54.3%) boys and 1985 (2.2%) requiring hospital admission. Among the models, extreme gradient boosting achieved the highest predictive abilities (eg, area under the precision-recall curve, 0.26; 95% confidence interval, 0.25-0.27; area under the receiver operating characteristic curve, 0.86; 95% confidence interval, 0.84-0.88; sensitivity, 0.77; and specificity, 0.82). With an optimal threshold, the positive and negative likelihood ratios were 4.22, and 0.28, respectively.

CONCLUSIONS : Machine learning-based prediction of hospital admissions may support physicians' decision-making for hospital admissions. However, further improvements are required before implementing these models in real clinical settings.

Hatachi Takeshi, Hashizume Takao, Taniguchi Masashi, Inata Yu, Aoki Yoshihiro, Kawamura Atsushi, Takeuchi Muneyuki

2023-Feb-01

Radiology Radiology

Deep Learning Image Reconstruction for CT: Technical Principles and Clinical Prospects.

In Radiology ; h5-index 91.0

Filtered back projection (FBP) has been the standard CT image reconstruction method for 4 decades. A simple, fast, and reliable technique, FBP has delivered high-quality images in several clinical applications. However, with faster and more advanced CT scanners, FBP has become increasingly obsolete. Higher image noise and more artifacts are especially noticeable in lower-dose CT imaging using FBP. This performance gap was partly addressed by model-based iterative reconstruction (MBIR). Yet, its "plastic" image appearance and long reconstruction times have limited widespread application. Hybrid iterative reconstruction partially addressed these limitations by blending FBP with MBIR and is currently the state-of-the-art reconstruction technique. In the past 5 years, deep learning reconstruction (DLR) techniques have become increasingly popular. DLR uses artificial intelligence to reconstruct high-quality images from lower-dose CT faster than MBIR. However, the performance of DLR algorithms relies on the quality of data used for model training. Higher-quality training data will become available with photon-counting CT scanners. At the same time, spectral data would greatly benefit from the computational abilities of DLR. This review presents an overview of the principles, technical approaches, and clinical applications of DLR, including metal artifact reduction algorithms. In addition, emerging applications and prospects are discussed.

Koetzier Lennart R, Mastrodicasa Domenico, Szczykutowicz Timothy P, van der Werf Niels R, Wang Adam S, Sandfort Veit, van der Molen Aart J, Fleischmann Dominik, Willemink Martin J

2023-Jan-31

Pathology Pathology

Next-Generation Morphometry for pathomics-data mining in histopathology.

In Nature communications ; h5-index 260.0

Pathology diagnostics relies on the assessment of morphology by trained experts, which remains subjective and qualitative. Here we developed a framework for large-scale histomorphometry (FLASH) performing deep learning-based semantic segmentation and subsequent large-scale extraction of interpretable, quantitative, morphometric features in non-tumour kidney histology. We use two internal and three external, multi-centre cohorts to analyse over 1000 kidney biopsies and nephrectomies. By associating morphometric features with clinical parameters, we confirm previous concepts and reveal unexpected relations. We show that the extracted features are independent predictors of long-term clinical outcomes in IgA-nephropathy. We introduce single-structure morphometric analysis by applying techniques from single-cell transcriptomics, identifying distinct glomerular populations and morphometric phenotypes along a trajectory of disease progression. Our study provides a concept for Next-generation Morphometry (NGM), enabling comprehensive quantitative pathology data mining, i.e., pathomics.

Hölscher David L, Bouteldja Nassim, Joodaki Mehdi, Russo Maria L, Lan Yu-Chia, Sadr Alireza Vafaei, Cheng Mingbo, Tesar Vladimir, Stillfried Saskia V, Klinkhammer Barbara M, Barratt Jonathan, Floege Jürgen, Roberts Ian S D, Coppo Rosanna, Costa Ivan G, Bülow Roman D, Boor Peter

2023-Jan-28

Radiology Radiology

Development and Validation of a Deep Learning Algorithm to Differentiate Colon Carcinoma From Acute Diverticulitis in Computed Tomography Images.

In JAMA network open

IMPORTANCE : Differentiating between malignant and benign etiology in large-bowel wall thickening on computed tomography (CT) images can be a challenging task. Artificial intelligence (AI) support systems can improve the diagnostic accuracy of radiologists, as shown for a variety of imaging tasks. Improvements in diagnostic performance, in particular the reduction of false-negative findings, may be useful in patient care.

OBJECTIVE : To develop and evaluate a deep learning algorithm able to differentiate colon carcinoma (CC) and acute diverticulitis (AD) on CT images and analyze the impact of the AI-support system in a reader study.

DESIGN, SETTING, AND PARTICIPANTS : In this diagnostic study, patients who underwent surgery between July 1, 2005, and October 1, 2020, for CC or AD were included. Three-dimensional (3-D) bounding boxes including the diseased bowel segment and surrounding mesentery were manually delineated and used to develop a 3-D convolutional neural network (CNN). A reader study with 10 observers of different experience levels was conducted. Readers were asked to classify the testing cohort under reading room conditions, first without and then with algorithmic support.

MAIN OUTCOMES AND MEASURES : To evaluate the diagnostic performance, sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were calculated for all readers and reader groups with and without AI support. Metrics were compared using the McNemar test and relative and absolute predictive value comparisons.

RESULTS : A total of 585 patients (AD: n = 267, CC: n = 318; mean [SD] age, 63.2 [13.4] years; 341 men [58.3%]) were included. The 3-D CNN reached a sensitivity of 83.3% (95% CI, 70.0%-96.6%) and specificity of 86.6% (95% CI, 74.5%-98.8%) for the test set, compared with the mean reader sensitivity of 77.6% (95% CI, 72.9%-82.3%) and specificity of 81.6% (95% CI, 77.2%-86.1%). The combined group of readers improved significantly with AI support from a sensitivity of 77.6% to 85.6% (95% CI, 81.3%-89.3%; P < .001) and a specificity of 81.6% to 91.3% (95% CI, 88.1%-94.5%; P < .001). Artificial intelligence support significantly reduced the number of false-negative and false-positive findings (NPV from 78.5% to 86.4% and PPV from 80.9% to 90.8%; P < .001).

CONCLUSIONS AND RELEVANCE : The findings of this study suggest that a deep learning model able to distinguish CC and AD in CT images as a support system may significantly improve the diagnostic performance of radiologists, which may improve patient care.

Ziegelmayer Sebastian, Reischl Stefan, Havrda Hannah, Gawlitza Joshua, Graf Markus, Lenhart Nicolas, Nehls Nadja, Lemke Tristan, Wilhelm Dirk, Lohöfer Fabian, Burian Egon, Neumann Philipp-Alexander, Makowski Marcus, Braren Rickmer

2023-Jan-03

Ophthalmology Ophthalmology

Early detection of visual impairment in young children using a smartphone-based deep learning system.

In Nature medicine ; h5-index 170.0

Early detection of visual impairment is crucial but is frequently missed in young children, who are capable of only limited cooperation with standard vision tests. Although certain features of visually impaired children, such as facial appearance and ocular movements, can assist ophthalmic practice, applying these features to real-world screening remains challenging. Here, we present a mobile health (mHealth) system, the smartphone-based Apollo Infant Sight (AIS), which identifies visually impaired children with any of 16 ophthalmic disorders by recording and analyzing their gazing behaviors and facial features under visual stimuli. Videos from 3,652 children (≤48 months in age; 54.5% boys) were prospectively collected to develop and validate this system. For detecting visual impairment, AIS achieved an area under the receiver operating curve (AUC) of 0.940 in an internal validation set and an AUC of 0.843 in an external validation set collected in multiple ophthalmology clinics across China. In a further test of AIS for at-home implementation by untrained parents or caregivers using their smartphones, the system was able to adapt to different testing conditions and achieved an AUC of 0.859. This mHealth system has the potential to be used by healthcare professionals, parents and caregivers for identifying young children with visual impairment across a wide range of ophthalmic disorders.

Chen Wenben, Li Ruiyang, Yu Qinji, Xu Andi, Feng Yile, Wang Ruixin, Zhao Lanqin, Lin Zhenzhe, Yang Yahan, Lin Duoru, Wu Xiaohang, Chen Jingjing, Liu Zhenzhen, Wu Yuxuan, Dang Kang, Qiu Kexin, Wang Zilong, Zhou Ziheng, Liu Dong, Wu Qianni, Li Mingyuan, Xiang Yifan, Li Xiaoyan, Lin Zhuoling, Zeng Danqi, Huang Yunjian, Mo Silang, Huang Xiucheng, Sun Shulin, Hu Jianmin, Zhao Jun, Wei Meirong, Hu Shoulong, Chen Liang, Dai Bingfa, Yang Huasheng, Huang Danping, Lin Xiaoming, Liang Lingyi, Ding Xiaoyan, Yang Yangfan, Wu Pengsen, Zheng Feihui, Stanojcic Nick, Li Ji-Peng Olivia, Cheung Carol Y, Long Erping, Chen Chuan, Zhu Yi, Yu-Wai-Man Patrick, Wang Ruixuan, Zheng Wei-Shi, Ding Xiaowei, Lin Haotian

2023-Jan-26

Public Health Public Health

OWL: an optimized and independently validated machine learning prediction model for lung cancer screening based on the UK Biobank, PLCO, and NLST populations.

In EBioMedicine

BACKGROUND : A reliable risk prediction model is critically important for identifying individuals with high risk of developing lung cancer as candidates for low-dose chest computed tomography (LDCT) screening. Leveraging a cutting-edge machine learning technique that accommodates a wide list of questionnaire-based predictors, we sought to optimize and validate a lung cancer prediction model.

METHODS : We developed an Optimized early Warning model for Lung cancer risk (OWL) using the XGBoost algorithm with 323,344 participants from the England area in UK Biobank (training set), and independently validated it with 93,227 participants from UKB Scotland and Wales area (validation set 1), as well as 70,605 and 66,231 participants in the Prostate, Lung, Colorectal, and Ovarian cancer screening trial (PLCO) control and intervention subpopulations, respectively (validation sets 2 & 3) and 23,138 and 18,669 participants in the United States National Lung Screening Trial (NLST) control and intervention subpopulations, respectively (validation sets 4 & 5). By comparing with three competitive prediction models, i.e., PLCO modified 2012 (PLCOm2012), PLCO modified 2014 (PLCOall2014), and the Liverpool Lung cancer Project risk model version 3 (LLPv3), we assessed the discrimination of OWL by the area under receiver operating characteristic curve (AUC) at the designed time point. We further evaluated the calibration using relative improvement in the ratio of expected to observed lung cancer cases (RIEO), and illustrated the clinical utility by the decision curve analysis.

FINDINGS : For general population, with validation set 1, OWL (AUC = 0.855, 95% CI: 0.829-0.880) presented a better discriminative capability than PLCOall2014 (AUC = 0.821, 95% CI: 0.794-0.848) (p < 0.001); with validation sets 2 & 3, AUC of OWL was comparable to PLCOall2014 (AUCPLCOall2014-AUCOWL < 1%). For ever-smokers, OWL outperformed PLCOm2012 and PLCOall2014 among ever-smokers in validation set 1 (AUCOWL = 0.842, 95% CI: 0.814-0.871; AUCPLCOm2012 = 0.792, 95% CI: 0.760-0.823; AUCPLCOall2014 = 0.791, 95% CI: 0.760-0.822, all p < 0.001). OWL remained comparable to PLCOm2012 and PLCOall2014 in discrimination (AUC difference from -0.014 to 0.008) among the ever-smokers in validation sets 2 to 5. In all the validation sets, OWL outperformed LLPv3 among the general population and the ever-smokers. Of note, OWL showed significantly better calibration than PLCOm2012, PLCOall2014 (RIEO from 43.1% to 92.3%, all p < 0.001), and LLPv3 (RIEO from 41.4% to 98.7%, all p < 0.001) in most cases. For clinical utility, OWL exhibited significant improvement in average net benefits (NB) over PLCOall2014 in validation set 1 (NB improvement: 32, p < 0.001); among ever smokers of validation set 1, OWL (average NB = 289) retained significant improvement over PLCOm2012 (average NB = 213) (p < 0.001). OWL had equivalent NBs with PLCOm2012 and PLCOall2014 in PLCO and NLST populations, while outperforming LLPv3 in the three populations.

INTERPRETATION : OWL, with a high degree of predictive accuracy and robustness, is a general framework with scientific justifications and clinical utility that can aid in screening individuals with high risks of lung cancer.

FUNDING : National Natural Science Foundation of China, the US NIH.

Pan Zoucheng, Zhang Ruyang, Shen Sipeng, Lin Yunzhi, Zhang Longyao, Wang Xiang, Ye Qian, Wang Xuan, Chen Jiajin, Zhao Yang, Christiani David C, Li Yi, Chen Feng, Wei Yongyue

2023-Jan-24

External validation, Lung cancer, Machine learning, Risk prediction, UK Biobank

Radiology Radiology

A wearable cardiac ultrasound imager.

In Nature ; h5-index 368.0

Continuous imaging of cardiac functions is highly desirable for the assessment of long-term cardiovascular health, detection of acute cardiac dysfunction and clinical management of critically ill or surgical patients1-4. However, conventional non-invasive approaches to image the cardiac function cannot provide continuous measurements owing to device bulkiness5-11, and existing wearable cardiac devices can only capture signals on the skin12-16. Here we report a wearable ultrasonic device for continuous, real-time and direct cardiac function assessment. We introduce innovations in device design and material fabrication that improve the mechanical coupling between the device and human skin, allowing the left ventricle to be examined from different views during motion. We also develop a deep learning model that automatically extracts the left ventricular volume from the continuous image recording, yielding waveforms of key cardiac performance indices such as stroke volume, cardiac output and ejection fraction. This technology enables dynamic wearable monitoring of cardiac performance with substantially improved accuracy in various environments.

Hu Hongjie, Huang Hao, Li Mohan, Gao Xiaoxiang, Yin Lu, Qi Ruixiang, Wu Ray S, Chen Xiangjun, Ma Yuxiang, Shi Keren, Li Chenghai, Maus Timothy M, Huang Brady, Lu Chengchangfeng, Lin Muyang, Zhou Sai, Lou Zhiyuan, Gu Yue, Chen Yimu, Lei Yusheng, Wang Xinyu, Wang Ruotao, Yue Wentong, Yang Xinyi, Bian Yizhou, Mu Jing, Park Geonho, Xiang Shu, Cai Shengqiang, Corey Paul W, Wang Joseph, Xu Sheng

2023-Jan

General General

Topological identification and interpretation for single-cell gene regulation elucidation across multiple platforms using scMGCA.

In Nature communications ; h5-index 260.0

Single-cell RNA sequencing provides high-throughput gene expression information to explore cellular heterogeneity at the individual cell level. A major challenge in characterizing high-throughput gene expression data arises from challenges related to dimensionality, and the prevalence of dropout events. To address these concerns, we develop a deep graph learning method, scMGCA, for single-cell data analysis. scMGCA is based on a graph-embedding autoencoder that simultaneously learns cell-cell topology representation and cluster assignments. We show that scMGCA is accurate and effective for cell segregation and batch effect correction, outperforming other state-of-the-art models across multiple platforms. In addition, we perform genomic interpretation on the key compressed transcriptomic space of the graph-embedding autoencoder to demonstrate the underlying gene regulation mechanism. We demonstrate that in a pancreatic ductal adenocarcinoma dataset, scMGCA successfully provides annotations on the specific cell types and reveals differential gene expression levels across multiple tumor-associated and cell signalling pathways.

Yu Zhuohan, Su Yanchi, Lu Yifu, Yang Yuning, Wang Fuzhou, Zhang Shixiong, Chang Yi, Wong Ka-Chun, Li Xiangtao

2023-Jan-25

Surgery Surgery

Delineating epileptogenic networks using brain imaging data and personalized modeling in drug-resistant epilepsy.

In Science translational medicine ; h5-index 138.0

Precise estimates of epileptogenic zone networks (EZNs) are crucial for planning intervention strategies to treat drug-resistant focal epilepsy. Here, we present the virtual epileptic patient (VEP), a workflow that uses personalized brain models and machine learning methods to estimate EZNs and to aid surgical strategies. The structural scaffold of the patient-specific whole-brain network model is constructed from anatomical T1 and diffusion-weighted magnetic resonance imaging. Each network node is equipped with a mathematical dynamical model to simulate seizure activity. Bayesian inference methods sample and optimize key parameters of the personalized model using functional stereoelectroencephalography recordings of patients' seizures. These key parameters together with their personalized model determine a given patient's EZN. Personalized models were further used to predict the outcome of surgical intervention using virtual surgeries. We evaluated the VEP workflow retrospectively using 53 patients with drug-resistant focal epilepsy. VEPs reproduced the clinically defined EZNs with a precision of 0.6, where the physical distance between epileptogenic regions identified by VEP and the clinically defined EZNs was small. Compared with the resected brain regions of 25 patients who underwent surgery, VEP showed lower false discovery rates in seizure-free patients (mean, 0.028) than in non-seizure-free patients (mean, 0.407). VEP is now being evaluated in an ongoing clinical trial (EPINOV) with an expected 356 prospective patients with epilepsy.

Wang Huifang E, Woodman Marmaduke, Triebkorn Paul, Lemarechal Jean-Didier, Jha Jayant, Dollomaja Borana, Vattikonda Anirudh Nihalani, Sip Viktor, Medina Villalon Samuel, Hashemi Meysam, Guye Maxime, Makhalova Julia, Bartolomei Fabrice, Jirsa Viktor

2023-Jan-25

oncology Oncology

Development and Validation of a Machine Learning Model for Detection and Classification of Tertiary Lymphoid Structures in Gastrointestinal Cancers.

In JAMA network open

IMPORTANCE : Tertiary lymphoid structures (TLSs) are associated with a favorable prognosis and improved response to cancer immunotherapy. The current approach for evaluation of TLSs is limited by interobserver variability and high complexity and cost of specialized imaging techniques.

OBJECTIVE : To develop a machine learning model for automated and quantitative evaluation of TLSs based on routine histopathology images.

DESIGN, SETTING, AND PARTICIPANTS : In this multicenter, international diagnostic/prognostic study, an interpretable machine learning model was developed and validated for automated detection, enumeration, and classification of TLSs in hematoxylin-eosin-stained images. A quantitative scoring system for TLSs was proposed, and its association with survival was investigated in patients with 1 of 6 types of gastrointestinal cancers. Data analysis was performed between June 2021 and March 2022.

MAIN OUTCOMES AND MEASURES : The diagnostic accuracy for classification of TLSs into 3 maturation states and the association of TLS score with survival were investigated.

RESULTS : A total of 1924 patients with gastrointestinal cancer from 7 independent cohorts (median [IQR] age ranging from 57 [49-64] years to 68 [58-77] years; proportion by sex ranging from 214 of 409 patients who were male [52.3%] to 134 of 155 patients who were male [86.5%]). The machine learning model achieved high accuracies for detecting and classifying TLSs into 3 states (TLS1: 97.7%; 95% CI, 96.4%-99.0%; TLS2: 96.3%; 95% CI, 94.6%-98.0%; TLS3: 95.7%; 95% CI, 93.9%-97.5%). TLSs were detected in 62 of 155 esophageal cancers (40.0%) and up to 267 of 353 gastric cancers (75.6%). Across 6 cancer types, patients were stratified into 3 risk groups (higher and lower TLS score and no TLS) and survival outcomes compared between groups: higher vs lower TLS score (hazard ratio [HR]; 0.27; 95% CI, 0.18-0.41; P < .001) and lower TLS score vs no TLSs (HR, 0.65; 95% CI, 0.56-0.76; P < .001). TLS score remained an independent prognostic factor associated with survival after adjusting for clinicopathologic variables and tumor-infiltrating lymphocytes (eg, for colon cancer: HR, 0.11; 95% CI, 0.02-0.47; P = .003).

CONCLUSIONS AND RELEVANCE : In this study, an interpretable machine learning model was developed that may allow automated and accurate detection of TLSs on routine tissue slide. This model is complementary to the cancer staging system for risk stratification in gastrointestinal cancers.

Li Zhe, Jiang Yuming, Li Bailiang, Han Zhen, Shen Jeanne, Xia Yong, Li Ruijiang

2023-Jan-03

Public Health Public Health

Predictive Accuracy of Stroke Risk Prediction Models Across Black and White Race, Sex, and Age Groups.

In JAMA ; h5-index 211.0

IMPORTANCE : Stroke is the fifth-highest cause of death in the US and a leading cause of serious long-term disability with particularly high risk in Black individuals. Quality risk prediction algorithms, free of bias, are key for comprehensive prevention strategies.

OBJECTIVE : To compare the performance of stroke-specific algorithms with pooled cohort equations developed for atherosclerotic cardiovascular disease for the prediction of new-onset stroke across different subgroups (race, sex, and age) and to determine the added value of novel machine learning techniques.

DESIGN, SETTING, AND PARTICIPANTS : Retrospective cohort study on combined and harmonized data from Black and White participants of the Framingham Offspring, Atherosclerosis Risk in Communities (ARIC), Multi-Ethnic Study for Atherosclerosis (MESA), and Reasons for Geographical and Racial Differences in Stroke (REGARDS) studies (1983-2019) conducted in the US. The 62 482 participants included at baseline were at least 45 years of age and free of stroke or transient ischemic attack.

EXPOSURES : Published stroke-specific algorithms from Framingham and REGARDS (based on self-reported risk factors) as well as pooled cohort equations for atherosclerotic cardiovascular disease plus 2 newly developed machine learning algorithms.

MAIN OUTCOMES AND MEASURES : Models were designed to estimate the 10-year risk of new-onset stroke (ischemic or hemorrhagic). Discrimination concordance index (C index) and calibration ratios of expected vs observed event rates were assessed at 10 years. Analyses were conducted by race, sex, and age groups.

RESULTS : The combined study sample included 62 482 participants (median age, 61 years, 54% women, and 29% Black individuals). Discrimination C indexes were not significantly different for the 2 stroke-specific models (Framingham stroke, 0.72; 95% CI, 0.72-073; REGARDS self-report, 0.73; 95% CI, 0.72-0.74) vs the pooled cohort equations (0.72; 95% CI, 0.71-0.73): differences 0.01 or less (P values >.05) in the combined sample. Significant differences in discrimination were observed by race: the C indexes were 0.76 for all 3 models in White vs 0.69 in Black women (all P values <.001) and between 0.71 and 0.72 in White men and between 0.64 and 0.66 in Black men (all P values ≤.001). When stratified by age, model discrimination was better for younger (<60 years) vs older (≥60 years) adults for both Black and White individuals. The ratios of observed to expected 10-year stroke rates were closest to 1 for the REGARDS self-report model (1.05; 95% CI, 1.00-1.09) and indicated risk overestimation for Framingham stroke (0.86; 95% CI, 0.82-0.89) and pooled cohort equations (0.74; 95% CI, 0.71-0.77). Performance did not significantly improve when novel machine learning algorithms were applied.

CONCLUSIONS AND RELEVANCE : In this analysis of Black and White individuals without stroke or transient ischemic attack among 4 US cohorts, existing stroke-specific risk prediction models and novel machine learning techniques did not significantly improve discriminative accuracy for new-onset stroke compared with the pooled cohort equations, and the REGARDS self-report model had the best calibration. All algorithms exhibited worse discrimination in Black individuals than in White individuals, indicating the need to expand the pool of risk factors and improve modeling techniques to address observed racial disparities and improve model performance.

Hong Chuan, Pencina Michael J, Wojdyla Daniel M, Hall Jennifer L, Judd Suzanne E, Cary Michael, Engelhard Matthew M, Berchuck Samuel, Xian Ying, D’Agostino Ralph, Howard George, Kissela Brett, Henao Ricardo

2023-Jan-24

General General

Explainable deep learning for insights in El Niño and river flows.

In Nature communications ; h5-index 260.0

The El Niño Southern Oscillation (ENSO) is a semi-periodic fluctuation in sea surface temperature (SST) over the tropical central and eastern Pacific Ocean that influences interannual variability in regional hydrology across the world through long-range dependence or teleconnections. Recent research has demonstrated the value of Deep Learning (DL) methods for improving ENSO prediction as well as Complex Networks (CN) for understanding teleconnections. However, gaps in predictive understanding of ENSO-driven river flows include the black box nature of DL, the use of simple ENSO indices to describe a complex phenomenon and translating DL-based ENSO predictions to river flow predictions. Here we show that eXplainable DL (XDL) methods, based on saliency maps, can extract interpretable predictive information contained in global SST and discover SST information regions and dependence structures relevant for river flows which, in tandem with climate network constructions, enable improved predictive understanding. Our results reveal additional information content in global SST beyond ENSO indices, develop understanding of how SSTs influence river flows, and generate improved river flow prediction, including uncertainty estimation. Observations, reanalysis data, and earth system model simulations are used to demonstrate the value of the XDL-CN based methods for future interannual and decadal scale climate projections.

Liu Yumin, Duffy Kate, Dy Jennifer G, Ganguly Auroop R

2023-Jan-20

Pathology Pathology

Multiplexed 3D atlas of state transitions and immune interaction in colorectal cancer.

In Cell ; h5-index 250.0

Advanced solid cancers are complex assemblies of tumor, immune, and stromal cells characterized by high intratumoral variation. We use highly multiplexed tissue imaging, 3D reconstruction, spatial statistics, and machine learning to identify cell types and states underlying morphological features of known diagnostic and prognostic significance in colorectal cancer. Quantitation of these features in high-plex marker space reveals recurrent transitions from one tumor morphology to the next, some of which are coincident with long-range gradients in the expression of oncogenes and epigenetic regulators. At the tumor invasive margin, where tumor, normal, and immune cells compete, T cell suppression involves multiple cell types and 3D imaging shows that seemingly localized 2D features such as tertiary lymphoid structures are commonly interconnected and have graded molecular properties. Thus, while cancer genetics emphasizes the importance of discrete changes in tumor state, whole-specimen imaging reveals large-scale morphological and molecular gradients analogous to those in developing tissues.

Lin Jia-Ren, Wang Shu, Coy Shannon, Chen Yu-An, Yapp Clarence, Tyler Madison, Nariya Maulik K, Heiser Cody N, Lau Ken S, Santagata Sandro, Sorger Peter K

2023-Jan-19

3D microscopy, PD1-PDL1 interaction, cellular, colorectal cancer, intermixed molecular, large-scale, morphological features, multiplexed imaging, spatial gradients, spatial proteomics, spatial transcriptomics, tertiary lymphoid structures, tumor atlas, tumor budding

General General

Multienzyme deep learning models improve peptide de novo sequencing by mass spectrometry proteomics.

In PLoS computational biology

Generating and analyzing overlapping peptides through multienzymatic digestion is an efficient procedure for de novo protein using from bottom-up mass spectrometry (MS). Despite improved instrumentation and software, de novo MS data analysis remains challenging. In recent years, deep learning models have represented a performance breakthrough. Incorporating that technology into de novo protein sequencing workflows require machine-learning models capable of handling highly diverse MS data. In this study, we analyzed the requirements for assembling such generalizable deep learning models by systemcally varying the composition and size of the training set. We assessed the generated models' performances using two test sets composed of peptides originating from the multienzyme digestion of samples from various species. The peptide recall values on the test sets showed that the deep learning models generated from a collection of highly N- and C-termini diverse peptides generalized 76% more over the termini-restricted ones. Moreover, expanding the training set's size by adding peptides from the multienzymatic digestion with five proteases of several species samples led to a 2-3 fold generalizability gain. Furthermore, we tested the applicability of these multienzyme deep learning (MEM) models by fully de novo sequencing the heavy and light monomeric chains of five commercial antibodies (mAbs). MEMs extracted over 10000 matching and overlapped peptides across six different proteases mAb samples, achieving a 100% sequence coverage for 8 of the ten polypeptide chains. We foretell that the MEMs' proven improvements to de novo analysis will positively impact several applications, such as analyzing samples of high complexity, unknown nature, or the peptidomics field.

Gueto-Tettay Carlos, Tang Di, Happonen Lotta, Heusel Moritz, Khakzad Hamed, Malmström Johan, Malmström Lars

2023-Jan-20

oncology Oncology

Association of Neighborhood Deprivation With Prostate Cancer and Immune Markers in African American and European American Men.

In JAMA network open

IMPORTANCE : Neighborhood variables may be factors in the excessive burden of prostate cancer among African American men.

OBJECTIVE : To examine associations between neighborhood deprivation, circulating immune-oncology markers, and prostate cancer among African American and European American men.

DESIGN, SETTING, AND PARTICIPANTS : A case-control study was conducted between January 1, 2005, and January 1, 2016. Participants included men with prostate cancer and age- and race-frequency-matched population controls. Participants were recruited at the Baltimore Veterans Affairs Medical Center and University of Maryland Medical Center; controls were obtained through the Maryland Motor Vehicle Administration database. National Death Index follow-up was performed through December 31, 2020, and data analysis was conducted from February 1, 2022, through October 31, 2022.

EXPOSURES : 2000 Census-tract Neighborhood Deprivation Index as a standardized score.

MAIN OUTCOMES AND MEASURES : Primary outcomes included prostate cancer, all-cause mortality, and disease-specific mortality. Secondary outcomes included the National Comprehensive Cancer Network risk score and serum proteomes for 82 immune-oncology markers with pathway annotation.

RESULTS : Participants included men with prostate cancer (n = 769: 405 African American, 364 European American men) and age- and race-frequency-matched population controls (n = 1023: 479 African American, 544 European American men). The median survival follow-up was 9.70 years (IQR, 5.77 years), with 219 deaths. Among 884 African American men, mean (SD) age at recruitment was 63.8 (7.6) years; mean (SD) age at recruitment among 908 European American men was 66.4 (8.1) years. In the multivariable logistic regression analysis with individual socioeconomic status adjustment, neighborhood deprivation was associated with 55% increased odds of prostate cancer among African American men (odds ratio [OR], 1.55; 95% CI, 1.33-1.81), but was not associated with the disease among European American men. Residing in the most-deprived vs least-deprived neighborhoods corresponded to 88% higher disease odds (OR, 1.88; 95% CI, 1.30-2.75) among all men and an approximate 3-fold increase among African American men (OR, 3.58; 95% CI, 1.72-7.45), but no association was noted among European American men. In Cox proportional hazard regression analyses, socioeconomic status-adjusted neighborhood deprivation was associated with an increased all-cause mortality only among African American men (hazard ratio [HR], 1.28; 95% CI, 1.08-1.53), whereas it was associated with metastatic disease and a 50% increased hazard of a prostate cancer-specific death among all men (HR, 1.50; 95% CI, 1.07-2.09). In analyses restricted to controls, neighborhood deprivation was associated with increased activity scores of serum proteome-defined chemotaxis, inflammation, and tumor immunity suppression.

CONCLUSIONS AND RELEVANCE : The findings of this study suggest that deprived neighborhood residency may increase the risk of African American men for prostate cancer and a related mortality, potentially through its association with systemic immune function and inflammation.

Pichardo Margaret S, Minas Tsion Zewdu, Pichardo Catherine M, Bailey-Whyte Maeve, Tang Wei, Dorsey Tiffany H, Wooten William, Ryan Brid M, Loffredo Christopher A, Ambs Stefan

2023-Jan-03

General General

A wearable motion capture suit and machine learning predict disease progression in Friedreich's ataxia.

In Nature medicine ; h5-index 170.0

Friedreich's ataxia (FA) is caused by a variant of the Frataxin (FXN) gene, leading to its downregulation and progressively impaired cardiac and neurological function. Current gold-standard clinical scales use simplistic behavioral assessments, which require 18- to 24-month-long trials to determine if therapies are beneficial. Here we captured full-body movement kinematics from patients with wearable sensors, enabling us to define digital behavioral features based on the data from nine FA patients (six females and three males) and nine age- and sex-matched controls, who performed the 8-m walk (8-MW) test and 9-hole peg test (9 HPT). We used machine learning to combine these features to longitudinally predict the clinical scores of the FA patients, and compared these with two standard clinical assessments, Spinocerebellar Ataxia Functional Index (SCAFI) and Scale for the Assessment and Rating of Ataxia (SARA). The digital behavioral features enabled longitudinal predictions of personal SARA and SCAFI scores 9 months into the future and were 1.7 and 4 times more precise than longitudinal predictions using only SARA and SCAFI scores, respectively. Unlike the two clinical scales, the digital behavioral features accurately predicted FXN gene expression levels for each FA patient in a cross-sectional manner. Our work demonstrates how data-derived wearable biomarkers can track personal disease trajectories and indicates the potential of such biomarkers for substantially reducing the duration or size of clinical trials testing disease-modifying therapies and for enabling behavioral transcriptomics.

Kadirvelu Balasundaram, Gavriel Constantinos, Nageshwaran Sathiji, Chan Jackson Ping Kei, Nethisinghe Suran, Athanasopoulos Stavros, Ricotti Valeria, Voit Thomas, Giunti Paola, Festenstein Richard, Faisal A Aldo

2023-Jan-19

General General

Federated learning for predicting histological response to neoadjuvant chemotherapy in triple-negative breast cancer.

In Nature medicine ; h5-index 170.0

Triple-negative breast cancer (TNBC) is a rare cancer, characterized by high metastatic potential and poor prognosis, and has limited treatment options. The current standard of care in nonmetastatic settings is neoadjuvant chemotherapy (NACT), but treatment efficacy varies substantially across patients. This heterogeneity is still poorly understood, partly due to the paucity of curated TNBC data. Here we investigate the use of machine learning (ML) leveraging whole-slide images and clinical information to predict, at diagnosis, the histological response to NACT for early TNBC women patients. To overcome the biases of small-scale studies while respecting data privacy, we conducted a multicentric TNBC study using federated learning, in which patient data remain secured behind hospitals' firewalls. We show that local ML models relying on whole-slide images can predict response to NACT but that collaborative training of ML models further improves performance, on par with the best current approaches in which ML models are trained using time-consuming expert annotations. Our ML model is interpretable and is sensitive to specific histological patterns. This proof of concept study, in which federated learning is applied to real-world datasets, paves the way for future biomarker discovery using unprecedentedly large datasets.

Ogier du Terrail Jean, Leopold Armand, Joly Clément, Béguier Constance, Andreux Mathieu, Maussion Charles, Schmauch Benoît, Tramel Eric W, Bendjebbar Etienne, Zaslavskiy Mikhail, Wainrib Gilles, Milder Maud, Gervasoni Julie, Guerin Julien, Durand Thierry, Livartowski Alain, Moutet Kelvin, Gautier Clément, Djafar Inal, Moisson Anne-Laure, Marini Camille, Galtier Mathieu, Balazard Félix, Dubois Rémy, Moreira Jeverson, Simon Antoine, Drubay Damien, Lacroix-Triki Magali, Franchet Camille, Bataillon Guillaume, Heudel Pierre-Etienne

2023-Jan-19

Ophthalmology Ophthalmology

Assessment of Retinopathy of Prematurity Regression and Reactivation Using an Artificial Intelligence-Based Vascular Severity Score.

In JAMA network open

IMPORTANCE : One of the biggest challenges when using anti-vascular endothelial growth factor (VEGF) agents to treat retinopathy of prematurity (ROP) is the need to perform long-term follow-up examinations to identify eyes at risk of ROP reactivation requiring retreatment.

OBJECTIVE : To evaluate whether an artificial intelligence (AI)-based vascular severity score (VSS) can be used to analyze ROP regression and reactivation after anti-VEGF treatment and potentially identify eyes at risk of ROP reactivation requiring retreatment.

DESIGN, SETTING, AND PARTICIPANTS : This prognostic study was a secondary analysis of posterior pole fundus images collected during the multicenter, double-blind, investigator-initiated Comparing Alternative Ranibizumab Dosages for Safety and Efficacy in Retinopathy of Prematurity (CARE-ROP) randomized clinical trial, which compared 2 different doses of ranibizumab (0.12 mg vs 0.20 mg) for the treatment of ROP. The CARE-ROP trial screened and enrolled infants between September 5, 2014, and July 14, 2016. A total of 1046 wide-angle fundus images obtained from 19 infants at predefined study time points were analyzed. The analyses of VSS were performed between January 20, 2021, and November 18, 2022.

INTERVENTIONS : An AI-based algorithm assigned a VSS between 1 (normal) and 9 (most severe) to fundus images.

MAIN OUTCOMES AND MEASURES : Analysis of VSS in infants with ROP over time and VSS comparisons between the 2 treatment groups (0.12 mg vs 0.20 mg of ranibizumab) and between infants who did and did not receive retreatment for ROP reactivation.

RESULTS : Among 19 infants with ROP in the CARE-ROP randomized clinical trial, the median (range) postmenstrual age at first treatment was 36.4 (34.7-39.7) weeks; 10 infants (52.6%) were male, and 18 (94.7%) were White. The mean (SD) VSS was 6.7 (1.9) at baseline and significantly decreased to 2.7 (1.9) at week 1 (P < .001) and 2.9 (1.3) at week 4 (P < .001). The mean (SD) VSS of infants with ROP reactivation requiring retreatment was 6.5 (1.9) at the time of retreatment, which was significantly higher than the VSS at week 4 (P < .001). No significant difference was found in VSS between the 2 treatment groups, but the change in VSS between baseline and week 1 was higher for infants who later required retreatment (mean [SD], 7.8 [1.3] at baseline vs 1.7 [0.7] at week 1) vs infants who did not (mean [SD], 6.4 [1.9] at baseline vs 3.0 [2.0] at week 1). In eyes requiring retreatment, higher baseline VSS was correlated with earlier time of retreatment (Pearson r = -0.9997; P < .001).

CONCLUSIONS AND RELEVANCE : In this study, VSS decreased after ranibizumab treatment, consistent with clinical disease regression. In cases of ROP reactivation requiring retreatment, VSS increased again to values comparable with baseline values. In addition, a greater change in VSS during the first week after initial treatment was found to be associated with a higher risk of later ROP reactivation, and high baseline VSS was correlated with earlier retreatment. These findings may have implications for monitoring ROP regression and reactivation after anti-VEGF treatment.

Eilts Sonja K, Pfeil Johanna M, Poschkamp Broder, Krohne Tim U, Eter Nicole, Barth Teresa, Guthoff Rainer, Lagrèze Wolf, Grundel Milena, Bründer Marie-Christine, Busch Martin, Kalpathy-Cramer Jayashree, Chiang Michael F, Chan R V Paul, Coyner Aaron S, Ostmo Susan, Campbell J Peter, Stahl Andreas

2023-Jan-03

General General

Wearable full-body motion tracking of activities of daily living predicts disease trajectory in Duchenne muscular dystrophy.

In Nature medicine ; h5-index 170.0

Artificial intelligence has the potential to revolutionize healthcare, yet clinical trials in neurological diseases continue to rely on subjective, semiquantitative and motivation-dependent endpoints for drug development. To overcome this limitation, we collected a digital readout of whole-body movement behavior of patients with Duchenne muscular dystrophy (DMD) (n = 21) and age-matched controls (n = 17). Movement behavior was assessed while the participant engaged in everyday activities using a 17-sensor bodysuit during three clinical visits over the course of 12 months. We first defined new movement behavioral fingerprints capable of distinguishing DMD from controls. Then, we used machine learning algorithms that combined the behavioral fingerprints to make cross-sectional and longitudinal disease course predictions, which outperformed predictions derived from currently used clinical assessments. Finally, using Bayesian optimization, we constructed a behavioral biomarker, termed the KineDMD ethomic biomarker, which is derived from daily-life behavioral data and whose value progresses with age in an S-shaped sigmoid curve form. The biomarker developed in this study, derived from digital readouts of daily-life movement behavior, can predict disease progression in patients with muscular dystrophy and can potentially track the response to therapy.

Ricotti Valeria, Kadirvelu Balasundaram, Selby Victoria, Festenstein Richard, Mercuri Eugenio, Voit Thomas, Faisal A Aldo

2023-Jan-19

General General

BMI-adjusted adipose tissue volumes exhibit depot-specific and divergent associations with cardiometabolic diseases.

In Nature communications ; h5-index 260.0

For any given body mass index (BMI), individuals vary substantially in fat distribution, and this variation may have important implications for cardiometabolic risk. Here, we study disease associations with BMI-independent variation in visceral (VAT), abdominal subcutaneous (ASAT), and gluteofemoral (GFAT) fat depots in 40,032 individuals of the UK Biobank with body MRI. We apply deep learning models based on two-dimensional body MRI projections to enable near-perfect estimation of fat depot volumes (R2 in heldout dataset = 0.978-0.991 for VAT, ASAT, and GFAT). Next, we derive BMI-adjusted metrics for each fat depot (e.g. VAT adjusted for BMI, VATadjBMI) to quantify local adiposity burden. VATadjBMI is associated with increased risk of type 2 diabetes and coronary artery disease, ASATadjBMI is largely neutral, and GFATadjBMI is associated with reduced risk. These results - describing three metabolically distinct fat depots at scale - clarify the cardiometabolic impact of BMI-independent differences in body fat distribution.

Agrawal Saaket, Klarqvist Marcus D R, Diamant Nathaniel, Stanley Takara L, Ellinor Patrick T, Mehta Nehal N, Philippakis Anthony, Ng Kenney, Claussnitzer Melina, Grinspoon Steven K, Batra Puneet, Khera Amit V

2023-Jan-17

General General

Transcriptional vulnerabilities of striatal neurons in human and rodent models of Huntington's disease.

In Nature communications ; h5-index 260.0

Striatal projection neurons (SPNs), which progressively degenerate in human patients with Huntington's disease (HD), are classified along two axes: the canonical direct-indirect pathway division and the striosome-matrix compartmentation. It is well established that the indirect-pathway SPNs are susceptible to neurodegeneration and transcriptomic disturbances, but less is known about how the striosome-matrix axis is compromised in HD in relation to the canonical axis. Here we show, using single-nucleus RNA-sequencing data from male Grade 1 HD patient post-mortem brain samples and male zQ175 and R6/2 mouse models, that the two axes are multiplexed and differentially compromised in HD. In human HD, striosomal indirect-pathway SPNs are the most depleted SPN population. In mouse HD models, the transcriptomic distinctiveness of striosome-matrix SPNs is diminished more than that of direct-indirect pathway SPNs. Furthermore, the loss of striosome-matrix distinction is more prominent within indirect-pathway SPNs. These results open the possibility that the canonical direct-indirect pathway and striosome-matrix compartments are differentially compromised in late and early stages of disease progression, respectively, differentially contributing to the symptoms, thus calling for distinct therapeutic strategies.

Matsushima Ayano, Pineda Sergio Sebastian, Crittenden Jill R, Lee Hyeseung, Galani Kyriakitsa, Mantero Julio, Tombaugh Geoffrey, Kellis Manolis, Heiman Myriam, Graybiel Ann M

2023-Jan-17

Radiology Radiology

Deep Learning Reconstruction Enables Prospectively Accelerated Clinical Knee MRI.

In Radiology ; h5-index 91.0

Background MRI is a powerful diagnostic tool with a long acquisition time. Recently, deep learning (DL) methods have provided accelerated high-quality image reconstructions from undersampled data, but it is unclear if DL image reconstruction can be reliably translated to everyday clinical practice. Purpose To determine the diagnostic equivalence of prospectively accelerated DL-reconstructed knee MRI compared with conventional accelerated MRI for evaluating internal derangement of the knee in a clinical setting. Materials and Methods A DL reconstruction model was trained with images from 298 clinical 3-T knee examinations. In a prospective analysis, patients clinically referred for knee MRI underwent a conventional accelerated knee MRI protocol at 3 T followed by an accelerated DL protocol between January 2020 and February 2021. The equivalence of the DL reconstruction of the images relative to the conventional images for the detection of an abnormality was assessed in terms of interchangeability. Each examination was reviewed by six musculoskeletal radiologists. Analyses pertaining to the detection of meniscal or ligament tears and bone marrow or cartilage abnormalities were based on four-point ordinal scores for the likelihood of an abnormality. Additionally, the protocols were compared with use of four-point ordinal scores for each aspect of image quality: overall image quality, presence of artifacts, sharpness, and signal-to-noise ratio. Results A total of 170 participants (mean age ± SD, 45 years ± 16; 76 men) were evaluated. The DL-reconstructed images were determined to be of diagnostic equivalence with the conventional images for detection of abnormalities. The overall image quality score, averaged over six readers, was significantly better (P < .001) for the DL than for the conventional images. Conclusion In a clinical setting, deep learning reconstruction enabled a nearly twofold reduction in scan time for a knee MRI and was diagnostically equivalent with the conventional protocol. © RSNA, 2023 Supplemental material is available for this article. See also the editorial by Roemer in this issue.

Johnson Patricia M, Lin Dana J, Zbontar Jure, Zitnick C Lawrence, Sriram Anuroop, Muckley Matthew, Babb James S, Kline Mitchell, Ciavarra Gina, Alaia Erin, Samim Mohammad, Walter William R, Calderon Liz, Pock Thomas, Sodickson Daniel K, Recht Michael P, Knoll Florian

2023-Jan-17

Radiology Radiology

Deep Learning Analysis of Chest Radiographs to Triage Patients with Acute Chest Pain Syndrome.

In Radiology ; h5-index 91.0

Background Patients presenting to the emergency department (ED) with acute chest pain (ACP) syndrome undergo additional testing to exclude acute coronary syndrome (ACS), pulmonary embolism (PE), or aortic dissection (AD), often yielding negative results. Purpose To assess whether deep learning (DL) analysis of the initial chest radiograph may help triage patients with ACP syndrome more efficiently. Materials and Methods This retrospective study used electronic health records of patients with ACP syndrome at presentation who underwent a combination of chest radiography and additional cardiovascular or pulmonary imaging or stress tests at two hospitals (Massachusetts General Hospital [MGH], Brigham and Women's Hospital [BWH]) between January 2005 and December 2015. A DL model was trained on 23 005 patients from MGH to predict a 30-day composite end point of ACS, PE, AD, and all-cause mortality based on chest radiographs. Area under the receiver operating characteristic curve (AUC) was used to compare performance between models (model 1: age + sex; model 2: model 1 + conventional troponin or d-dimer positivity; model 3: model 2 + DL predictions) in internal and external test sets from MGH and BWH, respectively. Results At MGH, 5750 patients (mean age, 59 years ± 17 [SD]; 3329 men, 2421 women) were evaluated. Model 3, which included DL predictions, significantly improved discrimination of those with the composite outcome compared with models 2 and 1 (AUC, 0.85 [95% CI: 0.84, 0.86] vs 0.76 [95% CI: 0.74, 0.77] vs 0.62 [95% CI: 0.60 0.64], respectively; P < .001 for all). When using a sensitivity threshold of 99%, 14% (813 of 5750) of patients could be deferred from cardiovascular or pulmonary testing for differential diagnosis of ACP syndrome using model 3 compared with 2% (98 of 5750) of patients using model 2 (P < .001). Model 3 maintained its diagnostic performance in different age, sex, race, and ethnicity groups. In external validation at BWH (22 764 patients; mean age, 57 years ± 17; 11 470 women), trends were similar and improved after fine tuning. Conclusion Deep learning analysis of chest radiographs may facilitate more efficient triage of patients with acute chest pain syndrome in the emergency department. © RSNA, 2023 Supplemental material is available for this article. See also the editorial by Goo in this issue.

Kolossváry Márton, Raghu Vineet K, Nagurney John T, Hoffmann Udo, Lu Michael T

2023-Jan-17

General General

Risk Factors for Pediatric Sepsis in the Emergency Department: A Machine Learning Pilot Study.

In Pediatric emergency care

OBJECTIVE : To identify underappreciated sepsis risk factors among children presenting to a pediatric emergency department (ED).

METHODS : A retrospective observational study (2017-2019) of children aged 18 years and younger presenting to a pediatric ED at a tertiary care children's hospital with fever, hypotension, or an infectious disease International Classification of Diseases (ICD)-10 diagnosis. Structured patient data including demographics, problem list, and vital signs were extracted for 35,074 qualifying ED encounters. According to the Improving Pediatric Sepsis Outcomes Classification, confirmed by expert review, 191 patients met clinical sepsis criteria. Five machine learning models were trained to predict sepsis/nonsepsis outcomes. Top features enabling model performance (N = 20) were then extracted to identify patient risk factors.

RESULTS : Machine learning methods reached a performance of up to 93% sensitivity and 84% specificity in identifying patients who received a hospital diagnosis of sepsis. A random forest classifier performed the best, followed by a classification and regression tree. Maximum documented heart rate was the top feature in these models, with importance coefficients (ICs) of 0.09 and 0.21, which represent how much an individual feature contributes to the model. Maximum mean arterial pressure was the second most important feature (IC 0.05, 0.13). Immunization status (IC 0.02), age (IC 0.03), and patient zip code (IC 0.02) were also among the top features enabling models to predict sepsis from ED visit data. Stratified analysis revealed changes in the predictive importance of risk factors by race, ethnicity, oncologic history, and insurance status.

CONCLUSIONS : Machine learning models trained to identify pediatric sepsis using ED clinical and sociodemographic variables confirmed well-established predictors, including heart rate and mean arterial pressure, and identified underappreciated relationships between sepsis and patient age, immunization status, and demographics.

Mercurio Laura, Pou Sovijja, Duffy Susan, Eickhoff Carsten

2023-Jan-17

General General

The next generation of evidence-based medicine.

In Nature medicine ; h5-index 170.0

Recently, advances in wearable technologies, data science and machine learning have begun to transform evidence-based medicine, offering a tantalizing glimpse into a future of next-generation 'deep' medicine. Despite stunning advances in basic science and technology, clinical translations in major areas of medicine are lagging. While the COVID-19 pandemic exposed inherent systemic limitations of the clinical trial landscape, it also spurred some positive changes, including new trial designs and a shift toward a more patient-centric and intuitive evidence-generation system. In this Perspective, I share my heuristic vision of the future of clinical trials and evidence-based medicine.

Subbiah Vivek

2023-Jan-16

General General

Comparative analysis of genome-scale, base-resolution DNA methylation profiles across 580 animal species.

In Nature communications ; h5-index 260.0

Methylation of cytosines is a prototypic epigenetic modification of the DNA. It has been implicated in various regulatory mechanisms across the animal kingdom and particularly in vertebrates. We mapped DNA methylation in 580 animal species (535 vertebrates, 45 invertebrates), resulting in 2443 genome-scale DNA methylation profiles of multiple organs. Bioinformatic analysis of this large dataset quantified the association of DNA methylation with the underlying genomic DNA sequence throughout vertebrate evolution. We observed a broadly conserved link with two major transitions-once in the first vertebrates and again with the emergence of reptiles. Cross-species comparisons focusing on individual organs supported a deeply conserved association of DNA methylation with tissue type, and cross-mapping analysis of DNA methylation at gene promoters revealed evolutionary changes for orthologous genes. In summary, this study establishes a large resource of vertebrate and invertebrate DNA methylomes, it showcases the power of reference-free epigenome analysis in species for which no reference genomes are available, and it contributes an epigenetic perspective to the study of vertebrate evolution.

Klughammer Johanna, Romanovskaia Daria, Nemc Amelie, Posautz Annika, Seid Charlotte A, Schuster Linda C, Keinath Melissa C, Lugo Ramos Juan Sebastian, Kosack Lindsay, Evankow Ann, Printz Dieter, Kirchberger Stefanie, Ergüner Bekir, Datlinger Paul, Fortelny Nikolaus, Schmidl Christian, Farlik Matthias, Skjærven Kaja, Bergthaler Andreas, Liedvogel Miriam, Thaller Denise, Burger Pamela A, Hermann Marcela, Distel Martin, Distel Daniel L, Kübber-Heiss Anna, Bock Christoph

2023-Jan-16

General General

The next generation of evidence-based medicine.

In Nature medicine ; h5-index 170.0

Recently, advances in wearable technologies, data science and machine learning have begun to transform evidence-based medicine, offering a tantalizing glimpse into a future of next-generation 'deep' medicine. Despite stunning advances in basic science and technology, clinical translations in major areas of medicine are lagging. While the COVID-19 pandemic exposed inherent systemic limitations of the clinical trial landscape, it also spurred some positive changes, including new trial designs and a shift toward a more patient-centric and intuitive evidence-generation system. In this Perspective, I share my heuristic vision of the future of clinical trials and evidence-based medicine.

Subbiah Vivek

2023-Jan-16

General General

Transformer for one stop interpretable cell type annotation.

In Nature communications ; h5-index 260.0

Consistent annotation transfer from reference dataset to query dataset is fundamental to the development and reproducibility of single-cell research. Compared with traditional annotation methods, deep learning based methods are faster and more automated. A series of useful single cell analysis tools based on autoencoder architecture have been developed but these struggle to strike a balance between depth and interpretability. Here, we present TOSICA, a multi-head self-attention deep learning model based on Transformer that enables interpretable cell type annotation using biologically understandable entities, such as pathways or regulons. We show that TOSICA achieves fast and accurate one-stop annotation and batch-insensitive integration while providing biologically interpretable insights for understanding cellular behavior during development and disease progressions. We demonstrate TOSICA's advantages by applying it to scRNA-seq data of tumor-infiltrating immune cells, and CD14+ monocytes in COVID-19 to reveal rare cell types, heterogeneity and dynamic trajectories associated with disease progression and severity.

Chen Jiawei, Xu Hao, Tao Wanyu, Chen Zhaoxiong, Zhao Yuxuan, Han Jing-Dong J

2023-Jan-14

General General

Transformer for one stop interpretable cell type annotation.

In Nature communications ; h5-index 260.0

Consistent annotation transfer from reference dataset to query dataset is fundamental to the development and reproducibility of single-cell research. Compared with traditional annotation methods, deep learning based methods are faster and more automated. A series of useful single cell analysis tools based on autoencoder architecture have been developed but these struggle to strike a balance between depth and interpretability. Here, we present TOSICA, a multi-head self-attention deep learning model based on Transformer that enables interpretable cell type annotation using biologically understandable entities, such as pathways or regulons. We show that TOSICA achieves fast and accurate one-stop annotation and batch-insensitive integration while providing biologically interpretable insights for understanding cellular behavior during development and disease progressions. We demonstrate TOSICA's advantages by applying it to scRNA-seq data of tumor-infiltrating immune cells, and CD14+ monocytes in COVID-19 to reveal rare cell types, heterogeneity and dynamic trajectories associated with disease progression and severity.

Chen Jiawei, Xu Hao, Tao Wanyu, Chen Zhaoxiong, Zhao Yuxuan, Han Jing-Dong J

2023-Jan-14

General General

Effect of a Machine Learning Recommender System and Viral Peer Marketing Intervention on Smoking Cessation: A Randomized Clinical Trial.

In JAMA network open

IMPORTANCE : Novel data science and marketing methods of smoking-cessation intervention have not been adequately evaluated.

OBJECTIVE : To compare machine learning recommender (ML recommender) computer tailoring of motivational text messages vs a standard motivational text-based intervention (standard messaging) and a viral peer-recruitment tool kit (viral tool kit) for recruiting friends and family vs no tool kit in a smoking-cessation intervention.

DESIGN, SETTING, AND PARTICIPANTS : This 2 ×2 factorial randomized clinical trial with partial allocation, conducted between July 2017 and September 2019 within an online tobacco intervention, recruited current smokers aged 18 years and older who spoke English from the US via the internet and peer referral. Data were analyzed from March through May 2022.

INTERVENTIONS : Participants registering for the online intervention were randomly assigned to the ML recommender or standard messaging groups followed by partially random allocation to access to viral tool kit or no viral tool kit groups. The ML recommender provided ongoing refinement of message selection based on user feedback and comparison with a growing database of other users, while the standard system selected messages based on participant baseline readiness to quit.

MAIN OUTCOMES AND MEASURES : Our primary outcome was self-reported 7-day point prevalence smoking cessation at 6 months.

RESULTS : Of 1487 participants who smoked (444 aged 19-34 years [29.9%], 508 aged 35-54 years [34.1%], 535 aged ≥55 years [36.0%]; 1101 [74.0%] females; 189 Black [12.7%] and 1101 White [78.5%]; 106 Hispanic [7.1%]), 741 individuals were randomly assigned to the ML recommender group and 746 individuals to the standard messaging group; viral tool kit access was provided to 745 participants, and 742 participants received no such access. There was no significant difference in 6-month smoking cessation between ML recommender (146 of 412 participants [35.4%] with outcome data) and standard messaging (156 of 389 participants [40.1%] with outcome data) groups (adjusted odds ratio, 0.81; 95% CI, 0.61-1.08). Smoking cessation was significantly higher in viral tool kit (177 of 395 participants [44.8%] with outcome data) vs no viral tool kit (125 of 406 participants [30.8%] with outcome data) groups (adjusted odds ratio, 1.48; 95% CI, 1.11-1.98).

CONCLUSIONS AND RELEVANCE : In this study, machine learning-based selection did not improve performance compared with standard message selection, while viral marketing did improve cessation outcomes. These results suggest that in addition to increasing dissemination, viral recruitment may have important implications for improving effectiveness of smoking-cessation interventions.

TRIAL REGISTRATION : ClinicalTrials.gov Identifier: NCT03224520.

Faro Jamie M, Chen Jinying, Flahive Julie, Nagawa Catherine S, Orvek Elizabeth A, Houston Thomas K, Allison Jeroan J, Person Sharina D, Smith Bridget M, Blok Amanda C, Sadasivam Rajani S

2023-Jan-03

Surgery Surgery

Genome-wide analysis of aberrant position and sequence of plasma DNA fragment ends in patients with cancer.

In Science translational medicine ; h5-index 138.0

Genome-wide fragmentation patterns in cell-free DNA (cfDNA) in plasma are strongly influenced by cellular origin due to variation in chromatin accessibility across cell types. Such differences between healthy and cancer cells provide the opportunity for development of novel cancer diagnostics. Here, we investigated whether analysis of cfDNA fragment end positions and their surrounding DNA sequences reveals the presence of tumor-derived DNA in blood. We performed genome-wide analysis of cfDNA from 521 samples and analyzed sequencing data from an additional 2147 samples, including healthy individuals and patients with 11 different cancer types. We developed a metric based on genome-wide differences in fragment positioning, weighted by fragment length and GC content [information-weighted fraction of aberrant fragments (iwFAF)]. We observed that iwFAF strongly correlated with tumor fraction, was higher for DNA fragments carrying somatic mutations, and was higher within genomic regions affected by copy number amplifications. We also calculated sample-level means of nucleotide frequencies observed at genomic positions spanning fragment ends. Using a combination of iwFAF and nine nucleotide frequencies from three positions surrounding fragment ends, we developed a machine learning model to differentiate healthy individuals from patients with cancer. We observed an area under the receiver operative characteristic curve (AUC) of 0.91 for detection of cancer at any stage and an AUC of 0.87 for detection of stage I cancer. Our findings remained robust with as few as 1 million fragments analyzed per sample, demonstrating that analysis of fragment ends can become a cost-effective and accessible approach for cancer detection and monitoring.

Budhraja Karan K, McDonald Bradon R, Stephens Michelle D, Contente-Cuomo Tania, Markus Havell, Farooq Maria, Favaro Patricia F, Connor Sydney, Byron Sara A, Egan Jan B, Ernst Brenda, McDaniel Timothy K, Sekulic Aleksandar, Tran Nhan L, Prados Michael D, Borad Mitesh J, Berens Michael E, Pockaj Barbara A, LoRusso Patricia M, Bryce Alan, Trent Jeffrey M, Murtaza Muhammed

2023-Jan-11

General General

Spikebench: An open benchmark for spike train time-series classification.

In PLoS computational biology

Modern well-performing approaches to neural decoding are based on machine learning models such as decision tree ensembles and deep neural networks. The wide range of algorithms that can be utilized to learn from neural spike trains, which are essentially time-series data, results in the need for diverse and challenging benchmarks for neural decoding, similar to the ones in the fields of computer vision and natural language processing. In this work, we propose a spike train classification benchmark, based on open-access neural activity datasets and consisting of several learning tasks such as stimulus type classification, animal's behavioral state prediction, and neuron type identification. We demonstrate that an approach based on hand-crafted time-series feature engineering establishes a strong baseline performing on par with state-of-the-art deep learning-based models for neural decoding. We release the code allowing to reproduce the reported results.

Lazarevich Ivan, Prokin Ilya, Gutkin Boris, Kazantsev Victor

2023-Jan-10

Surgery Surgery

Multistain deep learning for prediction of prognosis and therapy response in colorectal cancer.

In Nature medicine ; h5-index 170.0

Although it has long been known that the immune cell composition has a strong prognostic and predictive value in colorectal cancer (CRC), scoring systems such as the immunoscore (IS) or quantification of intraepithelial lymphocytes are only slowly being adopted into clinical routine use and have their limitations. To address this we established and evaluated a multistain deep learning model (MSDLM) utilizing artificial intelligence (AI) to determine the AImmunoscore (AIS) in more than 1,000 patients with CRC. Our model had high prognostic capabilities and outperformed other clinical, molecular and immune cell-based parameters. It could also be used to predict the response to neoadjuvant therapy in patients with rectal cancer. Using an explainable AI approach, we confirmed that the MSDLM's decisions were based on established cellular patterns of anti-tumor immunity. Hence, the AIS could provide clinicians with a valuable decision-making tool based on the tumor immune microenvironment.

Foersch Sebastian, Glasner Christina, Woerl Ann-Christin, Eckstein Markus, Wagner Daniel-Christoph, Schulz Stefan, Kellers Franziska, Fernandez Aurélie, Tserea Konstantina, Kloth Michael, Hartmann Arndt, Heintz Achim, Weichert Wilko, Roth Wilfried, Geppert Carol, Kather Jakob Nikolas, Jesinghaus Moritz

2023-Jan-09

oncology Oncology

Galaxy Training: A powerful framework for teaching!

In PLoS computational biology

There is an ongoing explosion of scientific datasets being generated, brought on by recent technological advances in many areas of the natural sciences. As a result, the life sciences have become increasingly computational in nature, and bioinformatics has taken on a central role in research studies. However, basic computational skills, data analysis, and stewardship are still rarely taught in life science educational programs, resulting in a skills gap in many of the researchers tasked with analysing these big datasets. In order to address this skills gap and empower researchers to perform their own data analyses, the Galaxy Training Network (GTN) has previously developed the Galaxy Training Platform (https://training.galaxyproject.org), an open access, community-driven framework for the collection of FAIR (Findable, Accessible, Interoperable, Reusable) training materials for data analysis utilizing the user-friendly Galaxy framework as its primary data analysis platform. Since its inception, this training platform has thrived, with the number of tutorials and contributors growing rapidly, and the range of topics extending beyond life sciences to include topics such as climatology, cheminformatics, and machine learning. While initially aimed at supporting researchers directly, the GTN framework has proven to be an invaluable resource for educators as well. We have focused our efforts in recent years on adding increased support for this growing community of instructors. New features have been added to facilitate the use of the materials in a classroom setting, simplifying the contribution flow for new materials, and have added a set of train-the-trainer lessons. Here, we present the latest developments in the GTN project, aimed at facilitating the use of the Galaxy Training materials by educators, and its usage in different learning environments.

Hiltemann Saskia, Rasche Helena, Gladman Simon, Hotz Hans-Rudolf, Larivière Delphine, Blankenberg Daniel, Jagtap Pratik D, Wollmann Thomas, Bretaudeau Anthony, Goué Nadia, Griffin Timothy J, Royaux Coline, Le Bras Yvan, Mehta Subina, Syme Anna, Coppens Frederik, Droesbeke Bert, Soranzo Nicola, Bacon Wendi, Psomopoulos Fotis, Gallardo-Alba Cristóbal, Davis John, Föll Melanie Christine, Fahrner Matthias, Doyle Maria A, Serrano-Solano Beatriz, Fouilloux Anne Claire, van Heusden Peter, Maier Wolfgang, Clements Dave, Heyl Florian, Grüning Björn, Batut Bérénice

2023-Jan

oncology Oncology

Early response evaluation by single cell signaling profiling in acute myeloid leukemia.

In Nature communications ; h5-index 260.0

Aberrant pro-survival signaling is a hallmark of cancer cells, but the response to chemotherapy is poorly understood. In this study, we investigate the initial signaling response to standard induction chemotherapy in a cohort of 32 acute myeloid leukemia (AML) patients, using 36-dimensional mass cytometry. Through supervised and unsupervised machine learning approaches, we find that reduction of extracellular-signal-regulated kinase (ERK) 1/2 and p38 mitogen-activated protein kinase (MAPK) phosphorylation in the myeloid cell compartment 24 h post-chemotherapy is a significant predictor of patient 5-year overall survival in this cohort. Validation by RNA sequencing shows induction of MAPK target gene expression in patients with high phospho-ERK1/2 24 h post-chemotherapy, while proteomics confirm an increase of the p38 prime target MAPK activated protein kinase 2 (MAPKAPK2). In this study, we demonstrate that mass cytometry can be a valuable tool for early response evaluation in AML and elucidate the potential of functional signaling analyses in precision oncology diagnostics.

Tislevoll Benedicte Sjo, Hellesøy Monica, Fagerholt Oda Helen Eck, Gullaksen Stein-Erik, Srivastava Aashish, Birkeland Even, Kleftogiannis Dimitrios, Ayuda-Durán Pilar, Piechaczyk Laure, Tadele Dagim Shiferaw, Skavland Jørn, Panagiotis Baliakas, Hovland Randi, Andresen Vibeke, Seternes Ole Morten, Tvedt Tor Henrik Anderson, Aghaeepour Nima, Gavasso Sonia, Porkka Kimmo, Jonassen Inge, Fløisand Yngvar, Enserink Jorrit, Blaser Nello, Gjertsen Bjørn Tore

2023-Jan-07

General General

Over-optimism in unsupervised microbiome analysis: Insights from network learning and clustering.

In PLoS computational biology

In recent years, unsupervised analysis of microbiome data, such as microbial network analysis and clustering, has increased in popularity. Many new statistical and computational methods have been proposed for these tasks. This multiplicity of analysis strategies poses a challenge for researchers, who are often unsure which method(s) to use and might be tempted to try different methods on their dataset to look for the "best" ones. However, if only the best results are selectively reported, this may cause over-optimism: the "best" method is overly fitted to the specific dataset, and the results might be non-replicable on validation data. Such effects will ultimately hinder research progress. Yet so far, these topics have been given little attention in the context of unsupervised microbiome analysis. In our illustrative study, we aim to quantify over-optimism effects in this context. We model the approach of a hypothetical microbiome researcher who undertakes four unsupervised research tasks: clustering of bacterial genera, hub detection in microbial networks, differential microbial network analysis, and clustering of samples. While these tasks are unsupervised, the researcher might still have certain expectations as to what constitutes interesting results. We translate these expectations into concrete evaluation criteria that the hypothetical researcher might want to optimize. We then randomly split an exemplary dataset from the American Gut Project into discovery and validation sets multiple times. For each research task, multiple method combinations (e.g., methods for data normalization, network generation, and/or clustering) are tried on the discovery data, and the combination that yields the best result according to the evaluation criterion is chosen. While the hypothetical researcher might only report this result, we also apply the "best" method combination to the validation dataset. The results are then compared between discovery and validation data. In all four research tasks, there are notable over-optimism effects; the results on the validation data set are worse compared to the discovery data, averaged over multiple random splits into discovery/validation data. Our study thus highlights the importance of validation and replication in microbiome analysis to obtain reliable results and demonstrates that the issue of over-optimism goes beyond the context of statistical testing and fishing for significance.

Ullmann Theresa, Peschel Stefanie, Finger Philipp, Müller Christian L, Boulesteix Anne-Laure

2023-Jan-06

General General

Photonic machine learning with on-chip diffractive optics.

In Nature communications ; h5-index 260.0

Machine learning technologies have been extensively applied in high-performance information-processing fields. However, the computation rate of existing hardware is severely circumscribed by conventional Von Neumann architecture. Photonic approaches have demonstrated extraordinary potential for executing deep learning processes that involve complex calculations. In this work, an on-chip diffractive optical neural network (DONN) based on a silicon-on-insulator platform is proposed to perform machine learning tasks with high integration and low power consumption characteristics. To validate the proposed DONN, we fabricated 1-hidden-layer and 3-hidden-layer on-chip DONNs with footprints of 0.15 mm2 and 0.3 mm2 and experimentally verified their performance on the classification task of the Iris plants dataset, yielding accuracies of 86.7% and 90%, respectively. Furthermore, a 3-hidden-layer on-chip DONN is fabricated to classify the Modified National Institute of Standards and Technology handwritten digit images. The proposed passive on-chip DONN provides a potential solution for accelerating future artificial intelligence hardware with enhanced performance.

Fu Tingzhao, Zang Yubin, Huang Yuyao, Du Zhenmin, Huang Honghao, Hu Chengyang, Chen Minghua, Yang Sigang, Chen Hongwei

2023-Jan-05

General General

Microcomb-based integrated photonic processing unit.

In Nature communications ; h5-index 260.0

The emergence of parallel convolution-operation technology has substantially powered the complexity and functionality of optical neural networks (ONN) by harnessing the dimension of optical wavelength. However, this advanced architecture faces remarkable challenges in high-level integration and on-chip operation. In this work, convolution based on time-wavelength plane stretching approach is implemented on a microcomb-driven chip-based photonic processing unit (PPU). To support the operation of this processing unit, we develop a dedicated control and operation protocol, leading to a record high weight precision of 9 bits. Moreover, the compact architecture and high data loading speed enable a preeminent photonic-core compute density of over 1 trillion of operations per second per square millimeter (TOPS mm-2). Two proof-of-concept experiments are demonstrated, including image edge detection and handwritten digit recognition, showing comparable processing capability compared to that of a digital computer. Due to the advanced performance and the great scalability, this parallel photonic processing unit can potentially revolutionize sophisticated artificial intelligence tasks including autonomous driving, video action recognition and image reconstruction.

Bai Bowen, Yang Qipeng, Shu Haowen, Chang Lin, Yang Fenghe, Shen Bitao, Tao Zihan, Wang Jing, Xu Shaofu, Xie Weiqiang, Zou Weiwen, Hu Weiwei, Bowers John E, Wang Xingjun

2023-Jan-05

oncology Oncology

Computational pathology in 2030: a Delphi study forecasting the role of AI in pathology within the next decade.

In EBioMedicine

BACKGROUND : Artificial intelligence (AI) is rapidly fuelling a fundamental transformation in the practice of pathology. However, clinical integration remains challenging, with no AI algorithms to date in routine adoption within typical anatomic pathology (AP) laboratories. This survey gathered current expert perspectives and expectations regarding the role of AI in AP from those with first-hand computational pathology and AI experience.

METHODS : Perspectives were solicited using the Delphi method from 24 subject matter experts between December 2020 and February 2021 regarding the anticipated role of AI in pathology by the year 2030. The study consisted of three consecutive rounds: 1) an open-ended, free response questionnaire generating a list of survey items; 2) a Likert-scale survey scored by experts and analysed for consensus; and 3) a repeat survey of items not reaching consensus to obtain further expert consensus.

FINDINGS : Consensus opinions were reached on 141 of 180 survey items (78.3%). Experts agreed that AI would be routinely and impactfully used within AP laboratory and pathologist clinical workflows by 2030. High consensus was reached on 100 items across nine categories encompassing the impact of AI on (1) pathology key performance indicators (KPIs) and (2) the pathology workforce and specific tasks performed by (3) pathologists and (4) AP lab technicians, as well as (5) specific AI applications and their likelihood of routine use by 2030, (6) AI's role in integrated diagnostics, (7) pathology tasks likely to be fully automated using AI, and (8) regulatory/legal and (9) ethical aspects of AI integration in pathology.

INTERPRETATION : This systematic consensus study details the expected short-to-mid-term impact of AI on pathology practice. These findings provide timely and relevant information regarding future care delivery in pathology and raise key practical, ethical, and legal challenges that must be addressed prior to AI's successful clinical implementation.

FUNDING : No specific funding was provided for this study.

Berbís M Alvaro, McClintock David S, Bychkov Andrey, Van der Laak Jeroen, Pantanowitz Liron, Lennerz Jochen K, Cheng Jerome Y, Delahunt Brett, Egevad Lars, Eloy Catarina, Farris Alton B, Fraggetta Filippo, García Del Moral Raimundo, Hartman Douglas J, Herrmann Markus D, Hollemans Eva, Iczkowski Kenneth A, Karsan Aly, Kriegsmann Mark, Salama Mohamed E, Sinard John H, Tuthill J Mark, Williams Bethany, Casado-Sánchez César, Sánchez-Turrión Víctor, Luna Antonio, Aneiros-Fernández José, Shen Jeanne

2023-Jan-03

Anatomic pathology, Artificial intelligence, Computational pathology, Digital pathology, Machine learning, Pathologist workflow

General General

Ten quick tips for computational analysis of medical images.

In PLoS computational biology

Medical imaging is a great asset for modern medicine, since it allows physicians to spatially interrogate a disease site, resulting in precise intervention for diagnosis and treatment, and to observe particular aspect of patients' conditions that otherwise would not be noticeable. Computational analysis of medical images, moreover, can allow the discovery of disease patterns and correlations among cohorts of patients with the same disease, thus suggesting common causes or providing useful information for better therapies and cures. Machine learning and deep learning applied to medical images, in particular, have produced new, unprecedented results that can pave the way to advanced frontiers of medical discoveries. While computational analysis of medical images has become easier, however, the possibility to make mistakes or generate inflated or misleading results has become easier, too, hindering reproducibility and deployment. In this article, we provide ten quick tips to perform computational analysis of medical images avoiding common mistakes and pitfalls that we noticed in multiple studies in the past. We believe our ten guidelines, if taken into practice, can help the computational-medical imaging community to perform better scientific research that eventually can have a positive impact on the lives of patients worldwide.

Chicco Davide, Shiradkar Rakesh

2023-Jan

Ophthalmology Ophthalmology

Rapid, label-free histopathological diagnosis of liver cancer based on Raman spectroscopy and deep learning.

In Nature communications ; h5-index 260.0

Biopsy is the recommended standard for pathological diagnosis of liver carcinoma. However, this method usually requires sectioning and staining, and well-trained pathologists to interpret tissue images. Here, we utilize Raman spectroscopy to study human hepatic tissue samples, developing and validating a workflow for in vitro and intraoperative pathological diagnosis of liver cancer. We distinguish carcinoma tissues from adjacent non-tumour tissues in a rapid, non-disruptive, and label-free manner by using Raman spectroscopy combined with deep learning, which is validated by tissue metabolomics. This technique allows for detailed pathological identification of the cancer tissues, including subtype, differentiation grade, and tumour stage. 2D/3D Raman images of unprocessed human tissue slices with submicrometric resolution are also acquired based on visualization of molecular composition, which could assist in tumour boundary recognition and clinicopathologic diagnosis. Lastly, the potential for a portable handheld Raman system is illustrated during surgery for real-time intraoperative human liver cancer diagnosis.

Huang Liping, Sun Hongwei, Sun Liangbin, Shi Keqing, Chen Yuzhe, Ren Xueqian, Ge Yuancai, Jiang Danfeng, Liu Xiaohu, Knoll Wolfgang, Zhang Qingwen, Wang Yi

2023-Jan-04

General General

Cerebro-cerebellar networks facilitate learning through feedback decoupling.

In Nature communications ; h5-index 260.0

Behavioural feedback is critical for learning in the cerebral cortex. However, such feedback is often not readily available. How the cerebral cortex learns efficiently despite the sparse nature of feedback remains unclear. Inspired by recent deep learning algorithms, we introduce a systems-level computational model of cerebro-cerebellar interactions. In this model a cerebral recurrent network receives feedback predictions from a cerebellar network, thereby decoupling learning in cerebral networks from future feedback. When trained in a simple sensorimotor task the model shows faster learning and reduced dysmetria-like behaviours, in line with the widely observed functional impact of the cerebellum. Next, we demonstrate that these results generalise to more complex motor and cognitive tasks. Finally, the model makes several experimentally testable predictions regarding cerebro-cerebellar task-specific representations over learning, task-specific benefits of cerebellar predictions and the differential impact of cerebellar and inferior olive lesions. Overall, our work offers a theoretical framework of cerebro-cerebellar networks as feedback decoupling machines.

Boven Ellen, Pemberton Joseph, Chadderton Paul, Apps Richard, Costa Rui Ponte

2023-Jan-04

General General

Development of a Machine Learning Model for Sonographic Assessment of Gestational Age.

In JAMA network open

IMPORTANCE : Fetal ultrasonography is essential for confirmation of gestational age (GA), and accurate GA assessment is important for providing appropriate care throughout pregnancy and for identifying complications, including fetal growth disorders. Derivation of GA from manual fetal biometry measurements (ie, head, abdomen, and femur) is operator dependent and time-consuming.

OBJECTIVE : To develop artificial intelligence (AI) models to estimate GA with higher accuracy and reliability, leveraging standard biometry images and fly-to ultrasonography videos.

DESIGN, SETTING, AND PARTICIPANTS : To improve GA estimates, this diagnostic study used AI to interpret standard plane ultrasonography images and fly-to ultrasonography videos, which are 5- to 10-second videos that can be automatically recorded as part of the standard of care before the still image is captured. Three AI models were developed and validated: (1) an image model using standard plane images, (2) a video model using fly-to videos, and (3) an ensemble model (combining both image and video models). The models were trained and evaluated on data from the Fetal Age Machine Learning Initiative (FAMLI) cohort, which included participants from 2 study sites at Chapel Hill, North Carolina (US), and Lusaka, Zambia. Participants were eligible to be part of this study if they received routine antenatal care at 1 of these sites, were aged 18 years or older, had a viable intrauterine singleton pregnancy, and could provide written consent. They were not eligible if they had known uterine or fetal abnormality, or had any other conditions that would make participation unsafe or complicate interpretation. Data analysis was performed from January to July 2022.

MAIN OUTCOMES AND MEASURES : The primary analysis outcome for GA was the mean difference in absolute error between the GA model estimate and the clinical standard estimate, with the ground truth GA extrapolated from the initial GA estimated at an initial examination.

RESULTS : Of the total cohort of 3842 participants, data were calculated for a test set of 404 participants with a mean (SD) age of 28.8 (5.6) years at enrollment. All models were statistically superior to standard fetal biometry-based GA estimates derived from images captured by expert sonographers. The ensemble model had the lowest mean absolute error compared with the clinical standard fetal biometry (mean [SD] difference, -1.51 [3.96] days; 95% CI, -1.90 to -1.10 days). All 3 models outperformed standard biometry by a more substantial margin on fetuses that were predicted to be small for their GA.

CONCLUSIONS AND RELEVANCE : These findings suggest that AI models have the potential to empower trained operators to estimate GA with higher accuracy.

Lee Chace, Willis Angelica, Chen Christina, Sieniek Marcin, Watters Amber, Stetson Bethany, Uddin Akib, Wong Jonny, Pilgrim Rory, Chou Katherine, Tse Daniel, Shetty Shravya, Gomes Ryan G

2023-Jan-03

Radiology Radiology

Two-dimensional Convolutional Neural Network Using Quantitative US for Noninvasive Assessment of Hepatic Steatosis in NAFLD.

In Radiology ; h5-index 91.0

Background Quantitative US (QUS) using radiofrequency data analysis has been recently introduced for noninvasive evaluation of hepatic steatosis. Deep learning algorithms may improve the diagnostic performance of QUS for hepatic steatosis. Purpose To evaluate a two-dimensional (2D) convolutional neural network (CNN) algorithm using QUS parametric maps and B-mode images for diagnosis of hepatic steatosis, with the MRI-derived proton density fat fraction (PDFF) as the reference standard, in patients with nonalcoholic fatty liver disease (NAFLD). Materials and Methods: Consecutive adult participants with suspected NAFLD were prospectively enrolled at a single academic medical center from July 2020 to June 2021. Using radiofrequency data analysis, two QUS parameters (tissue attenuation imaging [TAI] and tissue scatter-distribution imaging [TSI]) were measured. On B-mode images, hepatic steatosis was graded using visual scoring (none, mild, moderate, or severe). Using B-mode images and two QUS parametric maps (TAI and TSI) as input data, the algorithm estimated the US fat fraction (USFF) as a percentage. The correlation between the USFF and MRI PDFF was evaluated using the Pearson correlation coefficient. The diagnostic performance of the USFF for hepatic steatosis (MRI PDFF ≥5%) was evaluated using receiver operating characteristic curve analysis and compared with that of TAI, TSI, and visual scoring. Results Overall, 173 participants (mean age, 51 years ± 14 [SD]; 96 men) were included, with 126 (73%) having hepatic steatosis (MRI PDFF ≥5%). USFF correlated strongly with MRI PDFF (Pearson r = 0.86, 95% CI: 0.82, 0.90; P < .001). For diagnosing hepatic steatosis (MRI PDFF ≥5%), the USFF yielded an area under the receiver operating characteristic curve of 0.97 (95% CI: 0.93, 0.99), higher than those of TAI, TSI, and visual scoring (P = .015, .006, and < .001, respectively), with a sensitivity of 90% (95% CI: 84, 95 [114 of 126]) and a specificity of 91% (95% CI: 80, 98 [43 of 47]) at a cutoff value of 5.7%. Conclusion A deep learning algorithm using quantitative US parametric maps and B-mode images accurately estimated the hepatic fat fraction and diagnosed hepatic steatosis in participants with nonalcoholic fatty liver disease. ClinicalTrials.gov registration nos. NCT04462562, NCT04180631 © RSNA, 2023 Supplemental material is available for this article. See also the editorial by Sidhu and Fang in this issue.

Jeon Sun Kyung, Lee Jeong Min, Joo Ijin, Yoon Jeong Hee, Lee Gunwoo

2023-Jan-03

Ophthalmology Ophthalmology

Artificial intelligence for the diagnosis of retinopathy of prematurity: A systematic review of current algorithms.

In Eye (London, England) ; h5-index 41.0

BACKGROUND/OBJECTIVES : With the increasing survival of premature infants, there is an increased demand to provide adequate retinopathy of prematurity (ROP) services. Wide field retinal imaging (WFDRI) and artificial intelligence (AI) have shown promise in the field of ROP and have the potential to improve the diagnostic performance and reduce the workload for screening ophthalmologists. The aim of this review is to systematically review and provide a summary of the diagnostic characteristics of existing deep learning algorithms.

SUBJECT/METHODS : Two authors independently searched the literature, and studies using a deep learning system from retinal imaging were included. Data were extracted, assessed and reported using PRISMA guidelines.

RESULTS : Twenty-seven studies were included in this review. Nineteen studies used AI systems to diagnose ROP, classify the staging of ROP, diagnose the presence of pre-plus or plus disease, or assess the quality of retinal images. The included studies reported a sensitivity of 71%-100%, specificity of 74-99% and area under the curve of 91-99% for the primary outcome of the study. AI techniques were comparable to the assessment of ophthalmologists in terms of overall accuracy and sensitivity. Eight studies evaluated vascular severity scores and were able to accurately differentiate severity using an automated classification score.

CONCLUSION : Artificial intelligence for ROP diagnosis is a growing field, and many potential utilities have already been identified, including the presence of plus disease, staging of disease and a new automated severity score. AI has a role as an adjunct to clinical assessment; however, there is insufficient evidence to support its use as a sole diagnostic tool currently.

Ramanathan Ashwin, Athikarisamy Sam Ebenezer, Lam Geoffrey C

2022-Dec-28

Pathology Pathology

An accurate prediction of the origin for bone metastatic cancer using deep learning on digital pathological images.

In EBioMedicine

BACKGROUND : Determining the origin of bone metastatic cancer (OBMC) is of great significance to clinical therapeutics. It is challenging for pathologists to determine the OBMC with limited clinical information and bone biopsy.

METHODS : We designed a regional multiple-instance learning algorithm to predict the OBMC based on hematoxylin-eosin (H&E) staining slides alone. We collected 1041 cases from eight different hospitals and labeled 26,431 regions of interest to train the model. The performance of the model was assessed by ten-fold cross validation and external validation. Under the guidance of top3 predictions, we conducted an IHC test on 175 cases of unknown origins to compare the consistency of the results predicted by the model and indicated by the IHC markers. We also applied the model to identify whether there was tumor or not in a region, as well as distinguishing squamous cell carcinoma, adenocarcinoma, and neuroendocrine tumor.

FINDINGS : In the within-cohort, our model achieved a top1-accuracy of 91.35% and a top3-accuracy of 97.75%. In the external cohort, our model displayed a good generalizability with a top3-accuracy of 97.44%. The top1 consistency between the results of the model and the immunohistochemistry markers was 83.90% and the top3 consistency was 94.33%. The model obtained an accuracy of 98.98% to identify whether there was tumor or not and an accuracy of 93.85% to differentiate three types of cancers.

INTERPRETATION : Our model demonstrated good performance to predict the OBMC from routine histology and had great potential for assisting pathologists with determining the OBMC accurately.

FUNDING : National Science Foundation of China (61875102 and 61975089), Natural Science Foundation of Guangdong province (2021A15-15012379 and 2022A1515 012550), Science and Technology Research Program of Shenzhen City (JCYJ20200109110606054 and WDZC20200821141349001), and Tsinghua University Spring Breeze Fund (2020Z99CFZ023).

Zhu Lianghui, Shi Huijuan, Wei Huiting, Wang Chengjiang, Shi Shanshan, Zhang Fenfen, Yan Renao, Liu Yiqing, He Tingting, Wang Liyuan, Cheng Junru, Duan Hufei, Du Hong, Meng Fengjiao, Zhao Wenli, Gu Xia, Guo Linlang, Ni Yingpeng, He Yonghong, Guan Tian, Han Anjia

2022-Dec-26

Bone metastatic cancer, Deep learning, Digital pathology, Origin, Regional multiple-instance learning

General General

Multi-horizon predictive models for guiding extracorporeal resource allocation in critically ill COVID-19 patients.

In Journal of the American Medical Informatics Association : JAMIA

OBJECTIVE : Extracorporeal membrane oxygenation (ECMO) resource allocation tools are currently lacking. We developed machine learning (ML) models for predicting COVID-19 patients at risk of receiving ECMO to guide patient triage and resource allocation.

MATERIAL AND METHODS : We included COVID-19 patients admitted to intensive care units for >24 hours from March 2020 to October 2021, divided into training and testing development and testing only holdout cohorts. We developed ECMO-deployment timely prediction model ForecastECMO using Gradient Boosting Tree (GBT), with pre-ECMO prediction horizons from 0-48 hours, compared to PaO2/FiO2 (PF) ratio, Sequential Organ Failure Assessment (SOFA) score, PREdiction of Survival on ECMO Therapy-Score (PRESET) score, logistic regression (LR), and 30 pre-selected clinical variables GBT Clinical GBT models, with area under the receiver operator curve (AUROC) and precision recall curve (AUPRC) metrics.

RESULTS : ECMO prevalence was 2.89% and 1.73% in development and holdout cohorts. ForecastECMO had the best performance in both cohorts. At the 18-hour prediction horizon, a potentially clinically actionable pre-ECMO window, ForecastECMO had the highest AUROC (0.94 & 0.95) and AUPRC (0.54 & 0.37) in development and holdout cohorts in identifying ECMO patients without data 18-hours prior to ECMO.

DISCUSSION AND CONCLUSION : We developed a multi-horizon model, ForecastECMO, with high performance in identifying patients receiving ECMO at various prediction horizons. This model has potential to be used as early alert tool to guide ECMO resource allocation for COVID-19 patients. Future prospective multi-center validation would provide evidence for generalizability and real-world application of such models to improve patient outcomes.

Xue Bing, Shah Neel, Yang Hanqing, Kannampallil Thomas, Payne Philip Richard Orrin, Lu Chenyang, Said Ahmed Sameh

2022-Dec-28

COVID-19, ECMO, early alert, machine learning, prediction, resource allocation

Radiology Radiology

Validation of a Deep Learning-Based Model to Predict Lung Cancer Risk Using Chest Radiographs and Electronic Medical Record Data.

In JAMA network open

IMPORTANCE : Lung cancer screening with chest computed tomography (CT) prevents lung cancer death; however, fewer than 5% of eligible Americans are screened. CXR-LC, an open-source deep learning tool that estimates lung cancer risk from existing chest radiograph images and commonly available electronic medical record (EMR) data, may enable automated identification of high-risk patients as a step toward improving lung cancer screening participation.

OBJECTIVE : To validate CXR-LC using EMR data to identify individuals at high-risk for lung cancer to complement 2022 US Centers for Medicare & Medicaid Services (CMS) lung cancer screening eligibility guidelines.

DESIGN, SETTING, AND PARTICIPANTS : This prognostic study compared CXR-LC estimates with CMS screening guidelines using patient data from a large US hospital system. Included participants were persons who currently or formerly smoked cigarettes with an outpatient posterior-anterior chest radiograph between January 1, 2013, and December 31, 2014, with no history of lung cancer or screening CT. Data analysis was performed between May 2021 and June 2022.

EXPOSURES : CXR-LC lung cancer screening eligibility (previously defined as having a 3.297% or greater 12-year risk) based on inputs (chest radiograph image, age, sex, and whether currently smoking) extracted from the EMR.

MAIN OUTCOMES AND MEASURES : 6-year incident lung cancer.

RESULTS : A total of 14 737 persons were included in the study population (mean [SD] age, 62.6 [6.8] years; 7154 [48.5%] male; 204 [1.4%] Asian, 1051 [7.3%] Black, 432 [2.9%] Hispanic, 12 330 [85.2%] White) with a 2.4% rate of incident lung cancer over 6 years (361 patients with cancer). CMS eligibility could be determined in 6277 patients (42.6%) using smoking pack-year and quit-date from the EMR. Patients eligible by both CXR-LC and 2022 CMS criteria had a high rate of lung cancer (83 of 974 patients [8.5%]), higher than those eligible by 2022 CMS criteria alone (5 of 177 patients [2.8%]; P < .001). Patients eligible by CXR-LC but not 2022 CMS criteria also had a high 6-year incidence of lung cancer (121 of 3703 [3.3%]). In the 8460 cases (57.4%) where CMS eligibility was unknown, CXR-LC eligible patients had a 5-fold higher rate of lung cancer than ineligible (127 of 5177 [2.5%] vs 18 of 2283 [0.5%]; P < .001). Similar results were found in subgroups, including female patients and Black persons.

CONCLUSIONS AND RELEVANCE : Using routine chest radiographs and other data automatically extracted from the EMR, CXR-LC identified high-risk individuals who may benefit from lung cancer screening CT.

Raghu Vineet K, Walia Anika S, Zinzuwadia Aniket N, Goiffon Reece J, Shepard Jo-Anne O, Aerts Hugo J W L, Lennes Inga T, Lu Michael T

2022-Dec-01

General General

Multi-horizon predictive models for guiding extracorporeal resource allocation in critically ill COVID-19 patients.

In Journal of the American Medical Informatics Association : JAMIA

OBJECTIVE : Extracorporeal membrane oxygenation (ECMO) resource allocation tools are currently lacking. We developed machine learning (ML) models for predicting COVID-19 patients at risk of receiving ECMO to guide patient triage and resource allocation.

MATERIAL AND METHODS : We included COVID-19 patients admitted to intensive care units for >24 hours from March 2020 to October 2021, divided into training and testing development and testing only holdout cohorts. We developed ECMO-deployment timely prediction model ForecastECMO using Gradient Boosting Tree (GBT), with pre-ECMO prediction horizons from 0-48 hours, compared to PaO2/FiO2 (PF) ratio, Sequential Organ Failure Assessment (SOFA) score, PREdiction of Survival on ECMO Therapy-Score (PRESET) score, logistic regression (LR), and 30 pre-selected clinical variables GBT Clinical GBT models, with area under the receiver operator curve (AUROC) and precision recall curve (AUPRC) metrics.

RESULTS : ECMO prevalence was 2.89% and 1.73% in development and holdout cohorts. ForecastECMO had the best performance in both cohorts. At the 18-hour prediction horizon, a potentially clinically actionable pre-ECMO window, ForecastECMO had the highest AUROC (0.94 & 0.95) and AUPRC (0.54 & 0.37) in development and holdout cohorts in identifying ECMO patients without data 18-hours prior to ECMO.

DISCUSSION AND CONCLUSION : We developed a multi-horizon model, ForecastECMO, with high performance in identifying patients receiving ECMO at various prediction horizons. This model has potential to be used as early alert tool to guide ECMO resource allocation for COVID-19 patients. Future prospective multi-center validation would provide evidence for generalizability and real-world application of such models to improve patient outcomes.

Xue Bing, Shah Neel, Yang Hanqing, Kannampallil Thomas, Payne Philip Richard Orrin, Lu Chenyang, Said Ahmed Sameh

2022-Dec-28

COVID-19, ECMO, early alert, machine learning, prediction, resource allocation

General General

Prediction of designer-recombinases for DNA editing with generative deep learning.

In Nature communications ; h5-index 260.0

Site-specific tyrosine-type recombinases are effective tools for genome engineering, with the first engineered variants having demonstrated therapeutic potential. So far, adaptation to new DNA target site selectivity of designer-recombinases has been achieved mostly through iterative cycles of directed molecular evolution. While effective, directed molecular evolution methods are laborious and time consuming. Here we present RecGen (Recombinase Generator), an algorithm for the intelligent generation of designer-recombinases. We gather the sequence information of over one million Cre-like recombinase sequences evolved for 89 different target sites with which we train Conditional Variational Autoencoders for recombinase generation. Experimental validation demonstrates that the algorithm can predict recombinase sequences with activity on novel target-sites, indicating that RecGen is useful to accelerate the development of future designer-recombinases.

Schmitt Lukas Theo, Paszkowski-Rogacz Maciej, Jug Florian, Buchholz Frank

2022-Dec-27

Public Health Public Health

Prevalence of Asymptomatic SARS-CoV-2 Infection in Japan.

In JAMA network open

IMPORTANCE : Real-world evidence of SARS-CoV-2 transmission is needed to understand the prevalence of infection in the Japanese population.

OBJECTIVE : To conduct sentinel screening of the Japanese population to determine the prevalence of SARS-CoV-2 infection in asymptomatic individuals, with complementary analysis for symptomatic patients as reported by active epidemiologic surveillance used by the government.

DESIGN, SETTING, AND PARTICIPANTS : This cross-sectional study of a sentinel screening program investigated approximately 1 million asymptomatic individuals with polymerase chain reaction (PCR) testing for SARS-CoV-2 infection between February 22 and December 8, 2021. Participants included children, students, employed adults, and older individuals, as well as volunteers to broadly reflect the general Japanese population in the 14 prefectures of Japan that declared a state of emergency. Saliva samples and a cycle threshold (Ct) value of approximately 40 as standard in Japan were used. Polymerase chain reaction testing for symptomatic patients was separately done by public health authorities, and the results were obtained from the Ministry of Health, Labour, and Welfare of Japan to complement data on asymptomatic infections from the present study.

MAIN OUTCOMES AND MEASURES : Temporal trends in positivity and prevalence (including surges of different variants) and demographic associations (eg, age, geographic location, and vaccination status) were assessed.

RESULTS : The positive rate of SARS-CoV-2 infection in 1 082 976 asymptomatic individuals (52.08% males; mean [SD] age 39.4 [15.7] years) was 0.03% (95% CI, 0.02%-0.05%) during periods without surges and a maximum of 0.33% (95% CI, 0.25%-0.43%) during peak surges at the Japanese standard Ct value of approximately 40; however, the positive rate would have been 10-fold less at a Ct value of 25 as used elsewhere in the world (eg, UK). There was an increase in patients with a positive PCR test result with a Ct value of 25 or 30 preceding surges in infection and hotspots of asymptomatic infections.

CONCLUSIONS AND RELEVANCE : In this cross-sectional study of asymptomatic SARS-CoV-2 infection in the general population of Japan in 2021, as investigated by sentinel surveillance, a low rate of infection was seen in the Japanese population compared with reported levels elsewhere in the world. This finding provides real-world data on the state of infection in Japan.

Suzuki Toru, Aizawa Kenichi, Shibuya Kenji, Yamanaka Shinya, Anzai Yuichiro, Kurokawa Kiyoshi, Nagai Ryozo

2022-Dec-01

Public Health Public Health

Prevalence of Asymptomatic SARS-CoV-2 Infection in Japan.

In JAMA network open

IMPORTANCE : Real-world evidence of SARS-CoV-2 transmission is needed to understand the prevalence of infection in the Japanese population.

OBJECTIVE : To conduct sentinel screening of the Japanese population to determine the prevalence of SARS-CoV-2 infection in asymptomatic individuals, with complementary analysis for symptomatic patients as reported by active epidemiologic surveillance used by the government.

DESIGN, SETTING, AND PARTICIPANTS : This cross-sectional study of a sentinel screening program investigated approximately 1 million asymptomatic individuals with polymerase chain reaction (PCR) testing for SARS-CoV-2 infection between February 22 and December 8, 2021. Participants included children, students, employed adults, and older individuals, as well as volunteers to broadly reflect the general Japanese population in the 14 prefectures of Japan that declared a state of emergency. Saliva samples and a cycle threshold (Ct) value of approximately 40 as standard in Japan were used. Polymerase chain reaction testing for symptomatic patients was separately done by public health authorities, and the results were obtained from the Ministry of Health, Labour, and Welfare of Japan to complement data on asymptomatic infections from the present study.

MAIN OUTCOMES AND MEASURES : Temporal trends in positivity and prevalence (including surges of different variants) and demographic associations (eg, age, geographic location, and vaccination status) were assessed.

RESULTS : The positive rate of SARS-CoV-2 infection in 1 082 976 asymptomatic individuals (52.08% males; mean [SD] age 39.4 [15.7] years) was 0.03% (95% CI, 0.02%-0.05%) during periods without surges and a maximum of 0.33% (95% CI, 0.25%-0.43%) during peak surges at the Japanese standard Ct value of approximately 40; however, the positive rate would have been 10-fold less at a Ct value of 25 as used elsewhere in the world (eg, UK). There was an increase in patients with a positive PCR test result with a Ct value of 25 or 30 preceding surges in infection and hotspots of asymptomatic infections.

CONCLUSIONS AND RELEVANCE : In this cross-sectional study of asymptomatic SARS-CoV-2 infection in the general population of Japan in 2021, as investigated by sentinel surveillance, a low rate of infection was seen in the Japanese population compared with reported levels elsewhere in the world. This finding provides real-world data on the state of infection in Japan.

Suzuki Toru, Aizawa Kenichi, Shibuya Kenji, Yamanaka Shinya, Anzai Yuichiro, Kurokawa Kiyoshi, Nagai Ryozo

2022-Dec-01

Public Health Public Health

Development and Validation of a Machine Learning Model to Estimate Risk of Adverse Outcomes Within 30 Days of Opioid Dispensation.

In JAMA network open

IMPORTANCE : Machine learning approaches can assist opioid stewardship by identifying high-risk opioid prescribing for potential interventions.

OBJECTIVE : To develop a machine learning model for deployment that can estimate the risk of adverse outcomes within 30 days of an opioid dispensation as a potential component of prescription drug monitoring programs using access to real-world data.

DESIGN, SETTING, AND PARTICIPANTS : This prognostic study used population-level administrative health data to construct a machine learning model. This study took place in Alberta, Canada (from January 1, 2018, to December 31, 2019), and included all patients 18 years and older who received at least 1 opioid dispensation from a community pharmacy within the province.

EXPOSURES : Each opioid dispensation served as the unit of analysis.

MAIN OUTCOMES AND MEASURES : Opioid-related adverse outcomes were identified from administrative data sets. An XGBoost model was developed on 2018 data to estimate the risk of hospitalization, an emergency department visit, or mortality within 30 days of an opioid dispensation; validation on 2019 data was done to evaluate model performance. Model discrimination, calibration, and other relevant metrics are reported using daily and weekly predictions on both ranked predictions and predicted probability thresholds using all data from 2019.

RESULTS : A total of 853 324 participants represented 6 181 025 opioid dispensations, with 145 016 outcome events reported (2.3%); 46.4% of the participants were men and 53.6% were women, with a mean (SD) age of 49.1 (15.6) years for men and 51.0 (18.0) years for women. Of the outcome events, 77 326 (2.6% pretest probability) occurred within 30 days of a dispensation in the validation set (XGBoost C statistic, 0.82 [95% CI, 0.81-0.82]). The top 0.1 percentile of estimated risk had a positive likelihood ratio (LR) of 28.7, which translated to a posttest probability of 43.1%. In our simulations, the weekly measured predictions had higher positive LRs in both the highest-risk dispensations and percentiles of estimated risk compared with predictions measured daily. Net benefit analysis showed that using machine learning prediction may not add additional benefit over the entire range of probability thresholds.

CONCLUSIONS AND RELEVANCE : These findings suggest that prescription drug monitoring programs can use machine learning classifiers to identify patients at risk of opioid-related adverse outcomes and intervene on high-risk ranked predictions. Better access to available administrative and clinical data could improve the prediction performance of machine learning classifiers and thus expand opioid stewardship efforts.

Sharma Vishal, Kulkarni Vinaykumar, Jess Ed, Gilani Fizza, Eurich Dean, Simpson Scot H, Voaklander Don, Semenchuk Michael, London Connor, Samanani Salim

2022-Dec-01

General General

Physical deep learning with biologically inspired training method: gradient-free approach for physical hardware.

In Nature communications ; h5-index 260.0

Ever-growing demand for artificial intelligence has motivated research on unconventional computation based on physical devices. While such computation devices mimic brain-inspired analog information processing, the learning procedures still rely on methods optimized for digital processing such as backpropagation, which is not suitable for physical implementation. Here, we present physical deep learning by extending a biologically inspired training algorithm called direct feedback alignment. Unlike the original algorithm, the proposed method is based on random projection with alternative nonlinear activation. Thus, we can train a physical neural network without knowledge about the physical system and its gradient. In addition, we can emulate the computation for this training on scalable physical hardware. We demonstrate the proof-of-concept using an optoelectronic recurrent neural network called deep reservoir computer. We confirmed the potential for accelerated computation with competitive performance on benchmarks. Our results provide practical solutions for the training and acceleration of neuromorphic computation.

Nakajima Mitsumasa, Inoue Katsuma, Tanaka Kenji, Kuniyoshi Yasuo, Hashimoto Toshikazu, Nakajima Kohei

2022-Dec-26

Radiology Radiology

Accuracy and efficiency of an artificial intelligence-based pulmonary broncho-vascular three-dimensional reconstruction system supporting thoracic surgery: Retrospective and prospective validation study.

In EBioMedicine

BACKGROUND : Anthropomorphic phantoms are used in surgical planning and intervention. Ideal accuracy and high efficiency are prerequisites for its clinical application. We aimed to develop a fully automated artificial intelligence-based three-dimensional (3D) reconstruction system (AI system) to assist thoracic surgery and to determine its accuracy, efficiency, and safety for clinical use.

METHODS : This AI system was developed based on a 3D convolutional neural network (CNN) and optimized by gradient descent after training with 500 cases, achieving a Dice coefficient of 89.2%. Accuracy was verified by comparing virtual structures predicted by the AI system with anatomical structures of patients in retrospective (n = 113) and prospective cohorts (n = 139) who underwent lobectomy or segmentectomy at the Peking University Cancer Hospital. Operation time and blood loss were compared between the retrospective cohort (without AI assistance) and prospective cohort (with AI assistance) for safety evaluation. The time consumption for reconstruction and the quality score were compared between the AI system and manual reconstruction software (Mimics®) for efficiency validation. This study was registered at https://www.chictr.org.cn as ChiCTR2100050985.

FINDINGS : The AI system reconstructed 13,608 pulmonary segmental branches from retrospective and prospective cohorts, and 1573 branches of interest corresponding to phantoms were detectable during the operation for verification, achieving 100% and 97% accuracy for segmental bronchi, 97.2% and 99.1% for segmental arteries, and 93.2% and 98.8% for segmental veins, respectively. With the assistance of the AI system, the operation time was shortened by 24.5 min for lobectomy (p < 0.001) and 20 min for segmentectomy (p = 0.007). Compared to Mimics®, the AI system reduced the model reconstruction time by 14.2 min (p < 0.001), and it also outperformed Mimics® in model quality scores (p < 0.001).

INTERPRETATION : The AI system can accurately predict thoracic anatomical structures with higher efficiency than manual reconstruction software. Constant optimization and larger population validation are required.

FUNDING : This study was funded by the Beijing Natural Science Foundation (No. L222020) and other sources.

Li Xiang, Zhang Shanyuan, Luo Xiang, Gao Guangming, Luo Xiangfeng, Wang Shansi, Li Shaolei, Zhao Dachuan, Wang Yaqi, Cui Xinrun, Liu Bing, Tao Ye, Xiao Bufan, Tang Lei, Yan Shi, Wu Nan

2022-Dec-22

Accuracy, Anatomy, Artificial intelligence, Efficiency, Safety, Three-dimensional reconstruction model

Pathology Pathology

How, for whom, and in what contexts will artificial intelligence be adopted in pathology? A realist interview study.

In Journal of the American Medical Informatics Association : JAMIA

OBJECTIVE : There is increasing interest in using artificial intelligence (AI) in pathology to improve accuracy and efficiency. Studies of clinicians' perceptions of AI have found only moderate acceptability, suggesting further research is needed regarding integration into clinical practice. This study aimed to explore stakeholders' theories concerning how and in what contexts AI is likely to become integrated into pathology.

MATERIALS AND METHODS : A literature review provided tentative theories that were revised through a realist interview study with 20 pathologists and 5 pathology trainees. Questions sought to elicit whether, and in what ways, the tentative theories fitted with interviewees' perceptions and experiences. Analysis focused on identifying the contextual factors that may support or constrain uptake of AI in pathology.

RESULTS : Interviews highlighted the importance of trust in AI, with interviewees emphasizing evaluation and the opportunity for pathologists to become familiar with AI as means for establishing trust. Interviewees expressed a desire to be involved in design and implementation of AI tools, to ensure such tools address pressing needs, but needs vary by subspecialty. Workflow integration is desired but whether AI tools should work automatically will vary according to the task and the context.

CONCLUSIONS : It must not be assumed that AI tools that provide benefit in one subspecialty will provide benefit in others. Pathologists should be involved in the decision to introduce AI, with opportunity to assess strengths and weaknesses. Further research is needed concerning the evidence required to satisfy pathologists regarding the benefits of AI.

King Henry, Williams Bethany, Treanor Darren, Randell Rebecca

2022-Dec-24

artificial intelligence, implementation, pathology, qualitative research, realist evaluation

General General

Sourcing thermotolerant poly(ethylene terephthalate) hydrolase scaffolds from natural diversity.

In Nature communications ; h5-index 260.0

Enzymatic deconstruction of poly(ethylene terephthalate) (PET) is under intense investigation, given the ability of hydrolase enzymes to depolymerize PET to its constituent monomers near the polymer glass transition temperature. To date, reported PET hydrolases have been sourced from a relatively narrow sequence space. Here, we identify additional PET-active biocatalysts from natural diversity by using bioinformatics and machine learning to mine 74 putative thermotolerant PET hydrolases. We successfully express, purify, and assay 51 enzymes from seven distinct phylogenetic groups; observing PET hydrolysis activity on amorphous PET film from 37 enzymes in reactions spanning pH from 4.5-9.0 and temperatures from 30-70 °C. We conduct PET hydrolysis time-course reactions with the best-performing enzymes, where we observe differences in substrate selectivity as function of PET morphology. We employed X-ray crystallography and AlphaFold to examine the enzyme architectures of all 74 candidates, revealing protein folds and accessory domains not previously associated with PET deconstruction. Overall, this study expands the number and diversity of thermotolerant scaffolds for enzymatic PET deconstruction.

Erickson Erika, Gado Japheth E, Avilán Luisana, Bratti Felicia, Brizendine Richard K, Cox Paul A, Gill Raj, Graham Rosie, Kim Dong-Jin, König Gerhard, Michener William E, Poudel Saroj, Ramirez Kelsey J, Shakespeare Thomas J, Zahn Michael, Boyd Eric S, Payne Christina M, DuBois Jennifer L, Pickford Andrew R, Beckham Gregg T, McGeehan John E

2022-Dec-21

Internal Medicine Internal Medicine

A deep learning approach to identify missing is-a relations in SNOMED CT.

In Journal of the American Medical Informatics Association : JAMIA

OBJECTIVE : SNOMED CT is the largest clinical terminology worldwide. Quality assurance of SNOMED CT is of utmost importance to ensure that it provides accurate domain knowledge to various SNOMED CT-based applications. In this work, we introduce a deep learning-based approach to uncover missing is-a relations in SNOMED CT.

MATERIALS AND METHODS : Our focus is to identify missing is-a relations between concept-pairs exhibiting a containment pattern (ie, the set of words of one concept being a proper subset of that of the other concept). We use hierarchically related containment concept-pairs as positive instances and hierarchically unrelated containment concept-pairs as negative instances to train a model predicting whether an is-a relation exists between 2 concepts with containment pattern. The model is a binary classifier leveraging concept name features, hierarchical features, enriched lexical attribute features, and logical definition features. We introduce a cross-validation inspired approach to identify missing is-a relations among all hierarchically unrelated containment concept-pairs.

RESULTS : We trained and applied our model on the Clinical finding subhierarchy of SNOMED CT (September 2019 US edition). Our model (based on the validation sets) achieved a precision of 0.8164, recall of 0.8397, and F1 score of 0.8279. Applying the model to predict actual missing is-a relations, we obtained a total of 1661 potential candidates. Domain experts performed evaluation on randomly selected 230 samples and verified that 192 (83.48%) are valid.

CONCLUSIONS : The results showed that our deep learning approach is effective in uncovering missing is-a relations between containment concept-pairs in SNOMED CT.

Abeysinghe Rashmie, Zheng Fengbo, Bernstam Elmer V, Shi Jay, Bodenreider Olivier, Cui Licong

2022-Dec-20

SNOMED CT, deep learning, missing is-a relations, ontology quality assurance

Radiology Radiology

Inconsistent Partitioning and Unproductive Feature Associations Yield Idealized Radiomic Models.

In Radiology ; h5-index 91.0

Background Radiomics is the extraction of predefined mathematic features from medical images for the prediction of variables of clinical interest. While some studies report superlative accuracy of radiomic machine learning (ML) models, the published methodology is often incomplete, and the results are rarely validated in external testing data sets. Purpose To characterize the type, prevalence, and statistical impact of methodologic errors present in radiomic ML studies. Materials and Methods Radiomic ML publications were reviewed for the presence of performance-inflating methodologic flaws. Common flaws were subsequently reproduced with randomly generated features interpolated from publicly available radiomic data sets to demonstrate the precarious nature of reported findings. Results In an assessment of radiomic ML publications, the authors uncovered two general categories of data analysis errors: inconsistent partitioning and unproductive feature associations. In simulations, the authors demonstrated that inconsistent partitioning augments radiomic ML accuracy by 1.4 times from unbiased performance and that correcting for flawed methodologic results in areas under the receiver operating characteristic curve approaching a value of 0.5 (random chance). With use of randomly generated features, the authors illustrated that unproductive associations between radiomic features and gene sets can imply false causality for biologic phenomenon. Conclusion Radiomic machine learning studies may contain methodologic flaws that undermine their validity. This study provides a review template to avoid such flaws. © RSNA, 2022 Supplemental material is available for this article. See also the editorial by Jacobs in this issue.

Gidwani Mishka, Chang Ken, Patel Jay Biren, Hoebel Katharina Viktoria, Ahmed Syed Rakin, Singh Praveer, Fuller Clifton David, Kalpathy-Cramer Jayashree

2022-Dec-20

General General

scDSSC: Deep Sparse Subspace Clustering for scRNA-seq Data.

In PLoS computational biology

Single cell RNA sequencing (scRNA-seq) enables researchers to characterize transcriptomic profiles at the single-cell resolution with increasingly high throughput. Clustering is a crucial step in single cell analysis. Clustering analysis of transcriptome profiled by scRNA-seq can reveal the heterogeneity and diversity of cells. However, single cell study still remains great challenges due to its high noise and dimension. Subspace clustering aims at discovering the intrinsic structure of data in unsupervised fashion. In this paper, we propose a deep sparse subspace clustering method scDSSC combining noise reduction and dimensionality reduction for scRNA-seq data, which simultaneously learns feature representation and clustering via explicit modelling of scRNA-seq data generation. Experiments on a variety of scRNA-seq datasets from thousands to tens of thousands of cells have shown that scDSSC can significantly improve clustering performance and facilitate the interpretability of clustering and downstream analysis. Compared to some popular scRNA-deq analysis methods, scDSSC outperformed state-of-the-art methods under various clustering performance metrics.

Wang HaiYun, Zhao JianPing, Zheng ChunHou, Su YanSen

2022-Dec-19

Ophthalmology Ophthalmology

Optical coherence tomography imaging biomarkers associated with neovascular age-related macular degeneration: a systematic review.

In Eye (London, England) ; h5-index 41.0

UNLABELLED : The aim of this systematic literature review is twofold, (1) detail the impact of retinal biomarkers identifiable via optical coherence tomography (OCT) on disease progression and response to treatment in neovascular age-related macular degeneration (nAMD) and (2) establish which biomarkers are currently identifiable by artificial intelligence (AI) models and the utilisation of this technology. Following the PRISMA guidelines, PubMed was searched for peer-reviewed publications dated between January 2016 and January 2022.

POPULATION : Patients diagnosed with nAMD with OCT imaging.

SETTINGS : Comparable settings to NHS hospitals.

STUDY DESIGNS : Randomised controlled trials, prospective/retrospective cohort studies and review articles. From 228 articles, 130 were full-text reviewed, 50 were removed for falling outside the scope of this review with 10 added from the author's inventory, resulting in the inclusion of 90 articles. From 9 biomarkers identified; intraretinal fluid (IRF), subretinal fluid, pigment epithelial detachment, subretinal hyperreflective material (SHRM), retinal pigmental epithelial (RPE) atrophy, drusen, outer retinal tabulation (ORT), hyperreflective foci (HF) and retinal thickness, 5 are considered pertinent to nAMD disease progression; IRF, SHRM, drusen, ORT and HF. A number of these biomarkers can be classified using current AI models. Significant retinal biomarkers pertinent to disease activity and progression in nAMD are identifiable via OCT; IRF being the most important in terms of the significant impact on visual outcome. Incorporating AI into ophthalmology practice is a promising advancement towards automated and reproducible analyses of OCT data with the ability to diagnose disease and predict future disease conversion.

SYSTEMATIC REVIEW REGISTRATION : This review has been registered with PROSPERO (registration ID: CRD42021233200).

Hanson Rachel L W, Airody Archana, Sivaprasad Sobha, Gale Richard P

2022-Dec-16

General General

Deep embedding and alignment of protein sequences.

In Nature methods ; h5-index 152.0

Protein sequence alignment is a key component of most bioinformatics pipelines to study the structures and functions of proteins. Aligning highly divergent sequences remains, however, a difficult task that current algorithms often fail to perform accurately, leaving many proteins or open reading frames poorly annotated. Here we leverage recent advances in deep learning for language modeling and differentiable programming to propose DEDAL (deep embedding and differentiable alignment), a flexible model to align protein sequences and detect homologs. DEDAL is a machine learning-based model that learns to align sequences by observing large datasets of raw protein sequences and of correct alignments. Once trained, we show that DEDAL improves by up to two- or threefold the alignment correctness over existing methods on remote homologs and better discriminates remote homologs from evolutionarily unrelated sequences, paving the way to improvements on many downstream tasks relying on sequence alignment in structural and functional genomics.

Llinares-López Felipe, Berthet Quentin, Blondel Mathieu, Teboul Olivier, Vert Jean-Philippe

2022-Dec-15

oncology Oncology

Estimating diagnostic uncertainty in artificial intelligence assisted pathology using conformal prediction.

In Nature communications ; h5-index 260.0

Unreliable predictions can occur when an artificial intelligence (AI) system is presented with data it has not been exposed to during training. We demonstrate the use of conformal prediction to detect unreliable predictions, using histopathological diagnosis and grading of prostate biopsies as example. We digitized 7788 prostate biopsies from 1192 men in the STHLM3 diagnostic study, used for training, and 3059 biopsies from 676 men used for testing. With conformal prediction, 1 in 794 (0.1%) predictions is incorrect for cancer diagnosis (compared to 14 errors [2%] without conformal prediction) while 175 (22%) of the predictions are flagged as unreliable when the AI-system is presented with new data from the same lab and scanner that it was trained on. Conformal prediction could with small samples (N = 49 for external scanner, N = 10 for external lab and scanner, and N = 12 for external lab, scanner and pathology assessment) detect systematic differences in external data leading to worse predictive performance. The AI-system with conformal prediction commits 3 (2%) errors for cancer detection in cases of atypical prostate tissue compared to 44 (25%) without conformal prediction, while the system flags 143 (80%) unreliable predictions. We conclude that conformal prediction can increase patient safety of AI-systems.

Olsson Henrik, Kartasalo Kimmo, Mulliqi Nita, Capuccini Marco, Ruusuvuori Pekka, Samaratunga Hemamali, Delahunt Brett, Lindskog Cecilia, Janssen Emiel A M, Blilie Anders, Egevad Lars, Spjuth Ola, Eklund Martin

2022-Dec-15

General General

A unifying Bayesian framework for merging X-ray diffraction data.

In Nature communications ; h5-index 260.0

Novel X-ray methods are transforming the study of the functional dynamics of biomolecules. Key to this revolution is detection of often subtle conformational changes from diffraction data. Diffraction data contain patterns of bright spots known as reflections. To compute the electron density of a molecule, the intensity of each reflection must be estimated, and redundant observations reduced to consensus intensities. Systematic effects, however, lead to the measurement of equivalent reflections on different scales, corrupting observation of changes in electron density. Here, we present a modern Bayesian solution to this problem, which uses deep learning and variational inference to simultaneously rescale and merge reflection observations. We successfully apply this method to monochromatic and polychromatic single-crystal diffraction data, as well as serial femtosecond crystallography data. We find that this approach is applicable to the analysis of many types of diffraction experiments, while accurately and sensitively detecting subtle dynamics and anomalous scattering.

Dalton Kevin M, Greisman Jack B, Hekstra Doeke R

2022-Dec-15

General General

Recalibrating probabilistic forecasts of epidemics.

In PLoS computational biology

Distributional forecasts are important for a wide variety of applications, including forecasting epidemics. Often, forecasts are miscalibrated, or unreliable in assigning uncertainty to future events. We present a recalibration method that can be applied to a black-box forecaster given retrospective forecasts and observations, as well as an extension to make this method more effective in recalibrating epidemic forecasts. This method is guaranteed to improve calibration and log score performance when trained and measured in-sample. We also prove that the increase in expected log score of a recalibrated forecaster is equal to the entropy of the PIT distribution. We apply this recalibration method to the 27 influenza forecasters in the FluSight Network and show that recalibration reliably improves forecast accuracy and calibration. This method, available on Github, is effective, robust, and easy to use as a post-processing tool to improve epidemic forecasts.

Rumack Aaron, Tibshirani Ryan J, Rosenfeld Roni

2022-Dec-15

General General

Genome-wide identification and characterization of DNA enhancers with a stacked multivariate fusion framework.

In PLoS computational biology

Enhancers are short non-coding DNA sequences outside of the target promoter regions that can be bound by specific proteins to increase a gene's transcriptional activity, which has a crucial role in the spatiotemporal and quantitative regulation of gene expression. However, enhancers do not have a specific sequence motifs or structures, and their scattered distribution in the genome makes the identification of enhancers from human cell lines particularly challenging. Here we present a novel, stacked multivariate fusion framework called SMFM, which enables a comprehensive identification and analysis of enhancers from regulatory DNA sequences as well as their interpretation. Specifically, to characterize the hierarchical relationships of enhancer sequences, multi-source biological information and dynamic semantic information are fused to represent regulatory DNA enhancer sequences. Then, we implement a deep learning-based sequence network to learn the feature representation of the enhancer sequences comprehensively and to extract the implicit relationships in the dynamic semantic information. Ultimately, an ensemble machine learning classifier is trained based on the refined multi-source features and dynamic implicit relations obtained from the deep learning-based sequence network. Benchmarking experiments demonstrated that SMFM significantly outperforms other existing methods using several evaluation metrics. In addition, an independent test set was used to validate the generalization performance of SMFM by comparing it to other state-of-the-art enhancer identification methods. Moreover, we performed motif analysis based on the contribution scores of different bases of enhancer sequences to the final identification results. Besides, we conducted interpretability analysis of the identified enhancer sequences based on attention weights of EnhancerBERT, a fine-tuned BERT model that provides new insights into exploring the gene semantic information likely to underlie the discovered enhancers in an interpretable manner. Finally, in a human placenta study with 4,562 active distal gene regulatory enhancers, SMFM successfully exposed tissue-related placental development and the differential mechanism, demonstrating the generalizability and stability of our proposed framework.

Wang Yansong, Hou Zilong, Yang Yuning, Wong Ka-Chun, Li Xiangtao

2022-Dec-15

General General

Eleven quick tips for data cleaning and feature engineering.

In PLoS computational biology

Applying computational statistics or machine learning methods to data is a key component of many scientific studies, in any field, but alone might not be sufficient to generate robust and reliable outcomes and results. Before applying any discovery method, preprocessing steps are necessary to prepare the data to the computational analysis. In this framework, data cleaning and feature engineering are key pillars of any scientific study involving data analysis and that should be adequately designed and performed since the first phases of the project. We call "feature" a variable describing a particular trait of a person or an observation, recorded usually as a column in a dataset. Even if pivotal, these data cleaning and feature engineering steps sometimes are done poorly or inefficiently, especially by beginners and unexperienced researchers. For this reason, we propose here our quick tips for data cleaning and feature engineering on how to carry out these important preprocessing steps correctly avoiding common mistakes and pitfalls. Although we designed these guidelines with bioinformatics and health informatics scenarios in mind, we believe they can more in general be applied to any scientific area. We therefore target these guidelines to any researcher or practitioners wanting to perform data cleaning or feature engineering. We believe our simple recommendations can help researchers and scholars perform better computational analyses that can lead, in turn, to more solid outcomes and more reliable discoveries.

Chicco Davide, Oneto Luca, Tavazzi Erica

2022-Dec

Radiology Radiology

Evaluation of an Artificial Intelligence Model for Detection of Pneumothorax and Tension Pneumothorax in Chest Radiographs.

In JAMA network open

IMPORTANCE : Early detection of pneumothorax, most often via chest radiography, can help determine need for emergent clinical intervention. The ability to accurately detect and rapidly triage pneumothorax with an artificial intelligence (AI) model could assist with earlier identification and improve care.

OBJECTIVE : To compare the accuracy of an AI model vs consensus thoracic radiologist interpretations in detecting any pneumothorax (incorporating both nontension and tension pneumothorax) and tension pneumothorax.

DESIGN, SETTING, AND PARTICIPANTS : This diagnostic study was a retrospective standalone performance assessment using a data set of 1000 chest radiographs captured between June 1, 2015, and May 31, 2021. The radiographs were obtained from patients aged at least 18 years at 4 hospitals in the Mass General Brigham hospital network in the United States. Included radiographs were selected using 2 strategies from all chest radiography performed at the hospitals, including inpatient and outpatient. The first strategy identified consecutive radiographs with pneumothorax through a manual review of radiology reports, and the second strategy identified consecutive radiographs with tension pneumothorax using natural language processing. For both strategies, negative radiographs were selected by taking the next negative radiograph acquired from the same radiography machine as each positive radiograph. The final data set was an amalgamation of these processes. Each radiograph was interpreted independently by up to 3 radiologists to establish consensus ground-truth interpretations. Each radiograph was then interpreted by the AI model for the presence of pneumothorax and tension pneumothorax. This study was conducted between July and October 2021, with the primary analysis performed between October and November 2021.

MAIN OUTCOMES AND MEASURES : The primary end points were the areas under the receiver operating characteristic curves (AUCs) for the detection of pneumothorax and tension pneumothorax. The secondary end points were the sensitivities and specificities for the detection of pneumothorax and tension pneumothorax.

RESULTS : The final analysis included radiographs from 985 patients (mean [SD] age, 60.8 [19.0] years; 436 [44.3%] female patients), including 307 patients with nontension pneumothorax, 128 patients with tension pneumothorax, and 550 patients without pneumothorax. The AI model detected any pneumothorax with an AUC of 0.979 (95% CI, 0.970-0.987), sensitivity of 94.3% (95% CI, 92.0%-96.3%), and specificity of 92.0% (95% CI, 89.6%-94.2%) and tension pneumothorax with an AUC of 0.987 (95% CI, 0.980-0.992), sensitivity of 94.5% (95% CI, 90.6%-97.7%), and specificity of 95.3% (95% CI, 93.9%-96.6%).

CONCLUSIONS AND RELEVANCE : These findings suggest that the assessed AI model accurately detected pneumothorax and tension pneumothorax in this chest radiograph data set. The model's use in the clinical workflow could lead to earlier identification and improved care for patients with pneumothorax.

Hillis James M, Bizzo Bernardo C, Mercaldo Sarah, Chin John K, Newbury-Chaet Isabella, Digumarthy Subba R, Gilman Matthew D, Muse Victorine V, Bottrell Georgie, Seah Jarrel C Y, Jones Catherine M, Kalra Mannudeep K, Dreyer Keith J

2022-Dec-01

General General

Accuracy and data efficiency in deep learning models of protein expression.

In Nature communications ; h5-index 260.0

Synthetic biology often involves engineering microbial strains to express high-value proteins. Thanks to progress in rapid DNA synthesis and sequencing, deep learning has emerged as a promising approach to build sequence-to-expression models for strain optimization. But such models need large and costly training data that create steep entry barriers for many laboratories. Here we study the relation between accuracy and data efficiency in an atlas of machine learning models trained on datasets of varied size and sequence diversity. We show that deep learning can achieve good prediction accuracy with much smaller datasets than previously thought. We demonstrate that controlled sequence diversity leads to substantial gains in data efficiency and employed Explainable AI to show that convolutional neural networks can finely discriminate between input DNA sequences. Our results provide guidelines for designing genotype-phenotype screens that balance cost and quality of training data, thus helping promote the wider adoption of deep learning in the biotechnology sector.

Nikolados Evangelos-Marios, Wongprommoon Arin, Aodha Oisin Mac, Cambray Guillaume, Oyarzún Diego A

2022-Dec-15

General General

Interpreting tree ensemble machine learning models with endoR.

In PLoS computational biology

Tree ensemble machine learning models are increasingly used in microbiome science as they are compatible with the compositional, high-dimensional, and sparse structure of sequence-based microbiome data. While such models are often good at predicting phenotypes based on microbiome data, they only yield limited insights into how microbial taxa may be associated. We developed endoR, a method to interpret tree ensemble models. First, endoR simplifies the fitted model into a decision ensemble. Then, it extracts information on the importance of individual features and their pairwise interactions, displaying them as an interpretable network. Both the endoR network and importance scores provide insights into how features, and interactions between them, contribute to the predictive performance of the fitted model. Adjustable regularization and bootstrapping help reduce the complexity and ensure that only essential parts of the model are retained. We assessed endoR on both simulated and real metagenomic data. We found endoR to have comparable accuracy to other common approaches while easing and enhancing model interpretation. Using endoR, we also confirmed published results on gut microbiome differences between cirrhotic and healthy individuals. Finally, we utilized endoR to explore associations between human gut methanogens and microbiome components. Indeed, these hydrogen consumers are expected to interact with fermenting bacteria in a complex syntrophic network. Specifically, we analyzed a global metagenome dataset of 2203 individuals and confirmed the previously reported association between Methanobacteriaceae and Christensenellales. Additionally, we observed that Methanobacteriaceae are associated with a network of hydrogen-producing bacteria. Our method accurately captures how tree ensembles use features and interactions between them to predict a response. As demonstrated by our applications, the resultant visualizations and summary outputs facilitate model interpretation and enable the generation of novel hypotheses about complex systems.

Ruaud Albane, Pfister Niklas, Ley Ruth E, Youngblut Nicholas D

2022-Dec-14

Surgery Surgery

Development and Validation of a Prediction Model for Need for Massive Transfusion During Surgery Using Intraoperative Hemodynamic Monitoring Data.

In JAMA network open

IMPORTANCE : Massive transfusion is essential to prevent complications during uncontrolled intraoperative hemorrhage. As massive transfusion requires time for blood product preparation and additional medical personnel for a team-based approach, early prediction of massive transfusion is crucial for appropriate management.

OBJECTIVE : To evaluate a real-time prediction model for massive transfusion during surgery based on the incorporation of preoperative data and intraoperative hemodynamic monitoring data.

DESIGN, SETTING, AND PARTICIPANTS : This prognostic study used data sets from patients who underwent surgery with invasive blood pressure monitoring at Seoul National University Hospital (SNUH) from 2016 to 2019 and Boramae Medical Center (BMC) from 2020 to 2021. SNUH represented the development and internal validation data sets (n = 17 986 patients), and BMC represented the external validation data sets (n = 494 patients). Data were analyzed from November 2020 to December 2021.

EXPOSURES : A deep learning-based real-time prediction model for massive transfusion.

MAIN OUTCOMES AND MEASURES : Massive transfusion was defined as a transfusion of 3 or more units of red blood cells over an hour. A preoperative prediction model for massive transfusion was developed using preoperative variables. Subsequently, a real-time prediction model using preoperative and intraoperative parameters was constructed to predict massive transfusion 10 minutes in advance. A prediction model, the massive transfusion index, calculated the risk of massive transfusion in real time.

RESULTS : Among 17 986 patients at SNUH (mean [SD] age, 58.65 [14.81] years; 9036 [50.2%] female), 416 patients (2.3%) underwent massive transfusion during the operation (mean [SD] duration of operation, 170.99 [105.03] minutes). The real-time prediction model constructed with the use of preoperative and intraoperative parameters significantly outperformed the preoperative prediction model (area under the receiver characteristic curve [AUROC], 0.972; 95% CI, 0.968-0.976 vs AUROC, 0.824; 95% CI, 0.813-0.834 in the SNUH internal validation data set; P < .001). Patients with the highest massive transfusion index (ie, >90th percentile) had a 47.5-fold increased risk for a massive transfusion compared with those with a lower massive transfusion index (ie, <80th percentile). The real-time prediction model also showed excellent performance in the external validation data set (AUROC of 0.943 [95% CI, 0.919-0.961] in BMC).

CONCLUSIONS AND RELEVANCE : The findings of this prognostic study suggest that the real-time prediction model for massive transfusion showed high accuracy of prediction performance, enabling early intervention for high-risk patients. It suggests strong confidence in artificial intelligence-assisted clinical decision support systems in the operating field.

Lee Seung Mi, Lee Garam, Kim Tae Kyong, Le Trang, Hao Jie, Jung Young Mi, Park Chan-Wook, Park Joong Shin, Jun Jong Kwan, Lee Hyung-Chul, Kim Dokyoon

2022-Dec-01

Surgery Surgery

Factors Associated With Quality Care Among Adults With Rheumatoid Arthritis.

In JAMA network open

IMPORTANCE : Although quality care markers exist for patients with rheumatoid arthritis (RA), the predictors of meeting these markers are unclear.

OBJECTIVE : To explore factors associated with quality care among patients with RA.

DESIGN, SETTING, AND PARTICIPANTS : A retrospective cohort study using insurance claims from 2009 to 2017 was conducted, and 6 sequential logistic regression models were built to evaluate quality care markers. Quality care markers were measured at 1 year post-RA diagnosis for each patient. The MarketScan Research Database, which contains commercial and Medicare Advantage administrative claims data from more than 100 million individuals in the US, was used to identify patients aged 18 to 64 years with a diagnosis claim for RA. Patients with conditions presenting similar to RA and missing demographic characteristics were excluded. Data analysis occurred between February 18 and May 5, 2022.

EXPOSURES : Success or failure to meet selected RA quality care markers within 1 year after RA diagnosis.

MAIN OUTCOMES AND MEASURES : Prevalence of meeting successive quality care markers for RA.

RESULTS : Among 581 770 patients, 430 843 (74.1%) were women and the mean (SD) age was 48.9 (11.3) years. Most patients (236 285 [40.6%]) resided in the South and had an income less than or equal to $45 200 (490 366 [84.3%]). Of the total study population, 399 862 individuals (68.7%) met at least 1 quality care marker and 181 908 (31.3%) met 0 markers. Most commonly, patients met annual laboratory testing (299 323 [51.5%]) and referral to a rheumatologist (256 765 [44.1%]) markers. The least met marker was receiving hepatitis B screening prior to initiation of disease-modifying antirheumatic drug (DMARD) therapy (18 548 [3.2%]). Women were most likely to meet all quality care markers except receiving DMARDs with hepatitis B screening (odds ratio, 1.14; 95% CI, 1.12-1.16). Individuals with lower median household income had lower odds of receiving a rheumatologist referral, an annual physical examination, or annual laboratory testing, but greater odds of receiving the other quality care markers. Patients with Medicare and those with comorbidities were generally less likely to meet quality care markers.

CONCLUSIONS AND RELEVANCE : In this cohort study of patients with RA, findings indicated downstream associations with rheumatologist referral and receiving DMARDs and varied associations between meeting quality care markers and patient characteristics. These findings suggest that prioritizing early care, especially for vulnerable patients, will ensure that quality care continues.

Seyferth Anne V, Cichocki Meghan N, Wang Chien-Wei, Huang Yun-Ju, Huang Yi-Wei, Chen Jung-Sheng, Kuo Chang-Fu, Chung Kevin C

2022-Dec-01

Public Health Public Health

The development of an automatic speech recognition model using interview data from long-term care for older adults.

In Journal of the American Medical Informatics Association : JAMIA

OBJECTIVE : In long-term care (LTC) for older adults, interviews are used to collect client perspectives that are often recorded and transcribed verbatim, which is a time-consuming, tedious task. Automatic speech recognition (ASR) could provide a solution; however, current ASR systems are not effective for certain demographic groups. This study aims to show how data from specific groups, such as older adults or people with accents, can be used to develop an effective ASR.

MATERIALS AND METHODS : An initial ASR model was developed using the Mozilla Common Voice dataset. Audio and transcript data (34 h) from interviews with residents, family, and care professionals on quality of care were used. Interview data were continuously processed to reduce the word error rate (WER).

RESULTS : Due to background noise and mispronunciations, an initial ASR model had a WER of 48.3% on interview data. After finetuning using interview data, the average WER was reduced to 24.3%. When tested on speech data from the interviews, a median WER of 22.1% was achieved, with residents displaying the highest WER (22.7%). The resulting ASR model was at least 6 times faster than manual transcription.

DISCUSSION : The current method decreased the WER substantially, verifying its efficacy. Moreover, using local transcription of audio can be beneficial to the privacy of participants.

CONCLUSIONS : The current study shows that interview data from LTC for older adults can be effectively used to improve an ASR model. While the model output does still contain some errors, researchers reported that it saved much time during transcription.

Hacking Coen, Verbeek Hilde, Hamers Jan P H, Aarts Sil

2022-Dec-10

artificial intelligence, automatic speech recognition, long-term care, nursing homes, older adults

General General

Limited conservation in cross-species comparison of GLK transcription factor binding suggested wide-spread cistrome divergence.

In Nature communications ; h5-index 260.0

Non-coding cis-regulatory variants in animal genomes are an important driving force in the evolution of transcription regulation and phenotype diversity. However, cistrome dynamics in plants remain largely underexplored. Here, we compare the binding of GOLDEN2-LIKE (GLK) transcription factors in tomato, tobacco, Arabidopsis, maize and rice. Although the function of GLKs is conserved, most of their binding sites are species-specific. Conserved binding sites are often found near photosynthetic genes dependent on GLK for expression, but sites near non-differentially expressed genes in the glk mutant are nevertheless under purifying selection. The binding sites' regulatory potential can be predicted by machine learning model using quantitative genome features and TF co-binding information. Our study show that genome cis-variation caused wide-spread TF binding divergence, and most of the TF binding sites are genetically redundant. This poses a major challenge for interpreting the effect of individual sites and highlights the importance of quantitatively measuring TF occupancy.

Tu Xiaoyu, Ren Sibo, Shen Wei, Li Jianjian, Li Yuxiang, Li Chuanshun, Li Yangmeihui, Zong Zhanxiang, Xie Weibo, Grierson Donald, Fei Zhangjun, Giovannoni Jim, Li Pinghua, Zhong Silin

2022-Dec-09

oncology Oncology

The transcription factor DDIT3 is a potential driver of dyserythropoiesis in myelodysplastic syndromes.

In Nature communications ; h5-index 260.0

Myelodysplastic syndromes (MDS) are hematopoietic stem cell (HSC) malignancies characterized by ineffective hematopoiesis, with increased incidence in older individuals. Here we analyze the transcriptome of human HSCs purified from young and older healthy adults, as well as MDS patients, identifying transcriptional alterations following different patterns of expression. While aging-associated lesions seem to predispose HSCs to myeloid transformation, disease-specific alterations may trigger MDS development. Among MDS-specific lesions, we detect the upregulation of the transcription factor DNA Damage Inducible Transcript 3 (DDIT3). Overexpression of DDIT3 in human healthy HSCs induces an MDS-like transcriptional state, and dyserythropoiesis, an effect associated with a failure in the activation of transcriptional programs required for normal erythroid differentiation. Moreover, DDIT3 knockdown in CD34+ cells from MDS patients with anemia is able to restore erythropoiesis. These results identify DDIT3 as a driver of dyserythropoiesis, and a potential therapeutic target to restore the inefficient erythroid differentiation characterizing MDS patients.

Berastegui Nerea, Ainciburu Marina, Romero Juan P, Garcia-Olloqui Paula, Alfonso-Pierola Ana, Philippe Céline, Vilas-Zornoza Amaia, San Martin-Uriz Patxi, Ruiz-Hernández Raquel, Abarrategi Ander, Ordoñez Raquel, Alignani Diego, Sarvide Sarai, Castro-Labrador Laura, Lamo-Espinosa José M, San-Julian Mikel, Jimenez Tamara, López-Cadenas Félix, Muntion Sandra, Sanchez-Guijo Fermin, Molero Antonieta, Montoro Maria Julia, Tazón Bárbara, Serrano Guillermo, Diaz-Mazkiaran Aintzane, Hernaez Mikel, Huerga Sofía, Bewicke-Copley Findlay, Rio-Machin Ana, Maurano Matthew T, Díez-Campelo María, Valcarcel David, Rouault-Pierre Kevin, Lara-Astiaso David, Ezponda Teresa, Prosper Felipe

2022-Dec-09

General General

pGlycoQuant with a deep residual network for quantitative glycoproteomics at intact glycopeptide level.

In Nature communications ; h5-index 260.0

Large-scale intact glycopeptide identification has been advanced by software tools. However, tools for quantitative analysis remain lagging behind, which hinders exploring the differential site-specific glycosylation. Here, we report pGlycoQuant, a generic tool for both primary and tandem mass spectrometry-based intact glycopeptide quantitation. pGlycoQuant advances in glycopeptide matching through applying a deep learning model that reduces missing values by 19-89% compared with Byologic, MSFragger-Glyco, Skyline, and Proteome Discoverer, as well as a Match In Run algorithm for more glycopeptide coverage, greatly expanding the quantitative function of several widely used search engines, including pGlyco 2.0, pGlyco3, Byonic and MSFragger-Glyco. Further application of pGlycoQuant to the N-glycoproteomic study in three different metastatic HCC cell lines quantifies 6435 intact N-glycopeptides and, together with in vitro molecular biology experiments, illustrates site 979-core fucosylation of L1CAM as a potential regulator of HCC metastasis. We expected further applications of the freely available pGlycoQuant in glycoproteomic studies.

Kong Siyuan, Gong Pengyun, Zeng Wen-Feng, Jiang Biyun, Hou Xinhang, Zhang Yang, Zhao Huanhuan, Liu Mingqi, Yan Guoquan, Zhou Xinwen, Qiao Xihua, Wu Mengxi, Yang Pengyuan, Liu Chao, Cao Weiqian

2022-Dec-07

General General

Direct retrieval of Zernike-based pupil functions using integrated diffractive deep neural networks.

In Nature communications ; h5-index 260.0

Retrieving the pupil phase of a beam path is a central problem for optical systems across scales, from telescopes, where the phase information allows for aberration correction, to the imaging of near-transparent biological samples in phase contrast microscopy. Current phase retrieval schemes rely on complex digital algorithms that process data acquired from precise wavefront sensors, reconstructing the optical phase information at great expense of computational resources. Here, we present a compact optical-electronic module based on multi-layered diffractive neural networks printed on imaging sensors, capable of directly retrieving Zernike-based pupil phase distributions from an incident point spread function. We demonstrate this concept numerically and experimentally, showing the direct pupil phase retrieval of superpositions of the first 14 Zernike polynomials. The integrability of the diffractive elements with CMOS sensors shows the potential for the direct extraction of the pupil phase information from a detector module without additional digital post-processing.

Goi Elena, Schoenhardt Steffen, Gu Min

2022-Dec-07

General General

Negotiation and honesty in artificial intelligence methods for the board game of Diplomacy.

In Nature communications ; h5-index 260.0

The success of human civilization is rooted in our ability to cooperate by communicating and making joint plans. We study how artificial agents may use communication to better cooperate in Diplomacy, a long-standing AI challenge. We propose negotiation algorithms allowing agents to agree on contracts regarding joint plans, and show they outperform agents lacking this ability. For humans, misleading others about our intentions forms a barrier to cooperation. Diplomacy requires reasoning about our opponents' future plans, enabling us to study broken commitments between agents and the conditions for honest cooperation. We find that artificial agents face a similar problem as humans: communities of communicating agents are susceptible to peers who deviate from agreements. To defend against this, we show that the inclination to sanction peers who break contracts dramatically reduces the advantage of such deviators. Hence, sanctioning helps foster mostly truthful communication, despite conditions that initially favor deviations from agreements.

Kramár János, Eccles Tom, Gemp Ian, Tacchetti Andrea, McKee Kevin R, Malinowski Mateusz, Graepel Thore, Bachrach Yoram

2022-Dec-06

Radiology Radiology

Emerging and Evolving Concepts in Cancer Immunotherapy Imaging.

In Radiology ; h5-index 91.0

Criteria based on measurements of lesion diameter at CT have guided treatment with historical therapies due to the strong association between tumor size and survival. Clinical experience with immune checkpoint modulators shows that editing immune system function can be effective in various solid tumors. Equally, novel immune-related phenomena accompany this novel therapeutic paradigm. These effects of immunotherapy challenge the association of tumor size with response or progression and include risks and adverse events that present new demands for imaging to guide treatment decisions. Emerging and evolving approaches to immunotherapy highlight further key issues for imaging evaluation, such as dissociated response following local administration of immune checkpoint modulators, pseudoprogression due to immune infiltration in the tumor environment, and premature death due to hyperprogression. Research that may offer tools for radiologists to meet these challenges is reviewed. Different modalities are discussed, including immuno-PET, as well as new applications of CT, MRI, and fluorodeoxyglucose PET, such as radiomics and imaging of hematopoietic tissues or anthropometric characteristics. Multilevel integration of imaging and other biomarkers may improve clinical guidance for immunotherapies and provide theranostic opportunities.

Dercle Laurent, Sun Shawn, Seban Romain-David, Mekki Ahmed, Sun Roger, Tselikas Lambros, Hans Sophie, Bernard-Tessier Alice, Mihoubi Bouvier Fadila, Aide Nicolas, Vercellino Laetitia, Rivas Alexia, Girard Antoine, Mokrane Fatima-Zohra, Manson Guillaume, Houot Roch, Lopci Egesta, Yeh Randy, Ammari Samy, Schwartz Lawrence H

2022-Dec-06

Radiology Radiology

Predicting Hypoperfusion Lesion and Target Mismatch in Stroke from Diffusion-weighted MRI Using Deep Learning.

In Radiology ; h5-index 91.0

Background Perfusion imaging is important to identify a target mismatch in stroke but requires contrast agents and postprocessing software. Purpose To use a deep learning model to predict the hypoperfusion lesion in stroke and identify patients with a target mismatch profile from diffusion-weighted imaging (DWI) and clinical information alone, using perfusion MRI as the reference standard. Materials and Methods Imaging data sets of patients with acute ischemic stroke with baseline perfusion MRI and DWI were retrospectively reviewed from multicenter data available from 2008 to 2019 (Imaging Collaterals in Acute Stroke, Diffusion and Perfusion Imaging Evaluation for Understanding Stroke Evolution 2, and University of California, Los Angeles stroke registry). For perfusion MRI, rapid processing of perfusion and diffusion software automatically segmented the hypoperfusion lesion (time to maximum, ≥6 seconds) and ischemic core (apparent diffusion coefficient [ADC], ≤620 × 10-6 mm2/sec). A three-dimensional U-Net deep learning model was trained using baseline DWI, ADC, National Institutes of Health Stroke Scale score, and stroke symptom sidedness as inputs, with the union of hypoperfusion and ischemic core segmentation serving as the ground truth. Model performance was evaluated using the Dice score coefficient (DSC). Target mismatch classification based on the model was compared with that of the clinical-DWI mismatch approach defined by the DAWN trial by using the McNemar test. Results Overall, 413 patients (mean age, 67 years ± 15 [SD]; 207 men) were included for model development and primary analysis using fivefold cross-validation (247, 83, and 83 patients in the training, validation, and test sets, respectively, for each fold). The model predicted the hypoperfusion lesion with a median DSC of 0.61 (IQR, 0.45-0.71). The model identified patients with target mismatch with a sensitivity of 90% (254 of 283; 95% CI: 86, 93) and specificity of 77% (100 of 130; 95% CI: 69, 83) compared with the clinical-DWI mismatch sensitivity of 50% (140 of 281; 95% CI: 44, 56) and specificity of 89% (116 of 130; 95% CI: 83, 94) (P < .001 for all). Conclusion A three-dimensional U-Net deep learning model predicted the hypoperfusion lesion from diffusion-weighted imaging (DWI) and clinical information and identified patients with a target mismatch profile with higher sensitivity than the clinical-DWI mismatch approach. ClinicalTrials.gov registration nos. NCT02225730, NCT01349946, NCT02586415 © RSNA, 2022 Online supplemental material is available for this article. See also the editorial by Kallmes and Rabinstein in this issue.

Yu Yannan, Christensen Soren, Ouyang Jiahong, Scalzo Fabien, Liebeskind David S, Lansberg Maarten G, Albers Gregory W, Zaharchuk Greg

2022-Dec-06

General General

POPDx: an automated framework for patient phenotyping across 392 246 individuals in the UK Biobank study.

In Journal of the American Medical Informatics Association : JAMIA

OBJECTIVE : For the UK Biobank, standardized phenotype codes are associated with patients who have been hospitalized but are missing for many patients who have been treated exclusively in an outpatient setting. We describe a method for phenotype recognition that imputes phenotype codes for all UK Biobank participants.

MATERIALS AND METHODS : POPDx (Population-based Objective Phenotyping by Deep Extrapolation) is a bilinear machine learning framework for simultaneously estimating the probabilities of 1538 phenotype codes. We extracted phenotypic and health-related information of 392 246 individuals from the UK Biobank for POPDx development and evaluation. A total of 12 803 ICD-10 diagnosis codes of the patients were converted to 1538 phecodes as gold standard labels. The POPDx framework was evaluated and compared to other available methods on automated multiphenotype recognition.

RESULTS : POPDx can predict phenotypes that are rare or even unobserved in training. We demonstrate substantial improvement of automated multiphenotype recognition across 22 disease categories, and its application in identifying key epidemiological features associated with each phenotype.

CONCLUSIONS : POPDx helps provide well-defined cohorts for downstream studies. It is a general-purpose method that can be applied to other biobanks with diverse but incomplete data.

Yang Lu, Wang Sheng, Altman Russ B

2022-Dec-05

AI in healthcare, UK Biobank, deep learning, machine learning, patient phenotyping, rare disease

Public Health Public Health

Secuer: Ultrafast, scalable and accurate clustering of single-cell RNA-seq data.

In PLoS computational biology

Identifying cell clusters is a critical step for single-cell transcriptomics study. Despite the numerous clustering tools developed recently, the rapid growth of scRNA-seq volumes prompts for a more (computationally) efficient clustering method. Here, we introduce Secuer, a Scalable and Efficient speCtral clUstERing algorithm for scRNA-seq data. By employing an anchor-based bipartite graph representation algorithm, Secuer enjoys reduced runtime and memory usage over one order of magnitude for datasets with more than 1 million cells. Meanwhile, Secuer also achieves better or comparable accuracy than competing methods in small and moderate benchmark datasets. Furthermore, we showcase that Secuer can also serve as a building block for a new consensus clustering method, Secuer-consensus, which again improves the runtime and scalability of state-of-the-art consensus clustering methods while also maintaining the accuracy. Overall, Secuer is a versatile, accurate, and scalable clustering framework suitable for small to ultra-large single-cell clustering tasks.

Wei Nana, Nie Yating, Liu Lin, Zheng Xiaoqi, Wu Hua-Jun

2022-Dec-05

General General

Regional mutational signature activities in cancer genomes.

In PLoS computational biology

Cancer genomes harbor a catalog of somatic mutations. The type and genomic context of these mutations depend on their causes and allow their attribution to particular mutational signatures. Previous work has shown that mutational signature activities change over the course of tumor development, but investigations of genomic region variability in mutational signatures have been limited. Here, we expand upon this work by constructing regional profiles of mutational signature activities over 2,203 whole genomes across 25 tumor types, using data aggregated by the Pan-Cancer Analysis of Whole Genomes (PCAWG) consortium. We present GenomeTrackSig as an extension to the TrackSig R package to construct regional signature profiles using optimal segmentation and the expectation-maximization (EM) algorithm. We find that 426 genomes from 20 tumor types display at least one change in mutational signature activities (changepoint), and 306 genomes contain at least one of 54 recurrent changepoints shared by seven or more genomes of the same tumor type. Five recurrent changepoint locations are shared by multiple tumor types. Within these regions, the particular signature changes are often consistent across samples of the same type and some, but not all, are characterized by signatures associated with subclonal expansion. The changepoints we found cannot strictly be explained by gene density, mutation density, or cell-of-origin chromatin state. We hypothesize that they reflect a confluence of factors including evolutionary timing of mutational processes, regional differences in somatic mutation rate, large-scale changes in chromatin state that may be tissue type-specific, and changes in chromatin accessibility during subclonal expansion. These results provide insight into the regional effects of DNA damage and repair processes, and may help us localize genomic and epigenomic changes that occur during cancer development.

Timmons Caitlin, Morris Quaid, Harrigan Caitlin F

2022-Dec-05

Internal Medicine Internal Medicine

Prediction of type 2 diabetes using genome-wide polygenic risk score and metabolic profiles: A machine learning analysis of population-based 10-year prospective cohort study.

In EBioMedicine

BACKGROUND : Previous work on predicting type 2 diabetes by integrating clinical and genetic factors has mostly focused on the Western population. In this study, we use genome-wide polygenic risk score (gPRS) and serum metabolite data for type 2 diabetes risk prediction in the Asian population.

METHODS : Data of 1425 participants from the Korean Genome and Epidemiology Study (KoGES) Ansan-Ansung cohort were used in this study. For gPRS analysis, genotypic and clinical information from KoGES health examinee (n = 58,701) and KoGES cardiovascular disease association (n = 8105) sub-cohorts were included. Linkage disequilibrium analysis identified 239,062 genetic variants that were used to determine the gPRS, while the metabolites were selected using the Boruta algorithm. We used bootstrapped cross-validation to evaluate logistic regression and random forest (RF)-based machine learning models. Finally, associations of gPRS and selected metabolites with the values of homeostatic model assessment of beta-cell function (HOMA-B) and insulin resistance (HOMA-IR) were further estimated.

FINDINGS : During the follow-up period (8.3 ± 2.8 years), 331 participants (23.2%) were diagnosed with type 2 diabetes. The areas under the curves of the RF-based models were 0.844, 0.876, and 0.883 for the model using only demographic and clinical factors, model including the gPRS, and model with both gPRS and metabolites, respectively. Incorporation of additional parameters in the latter two models improved the classification by 11.7% and 4.2% respectively. While gPRS was significantly associated with HOMA-B value, most metabolites had a significant association with HOMA-IR value.

INTERPRETATION : Incorporating both gPRS and metabolite data led to enhanced type 2 diabetes risk prediction by capturing distinct etiologies of type 2 diabetes development. An RF-based model using clinical factors, gPRS, and metabolites predicted type 2 diabetes risk more accurately than the logistic regression-based model.

FUNDING : This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korean government (MEST) (No. 2019M3E5D1A02070863 and 2022R1C1C1005458). This work was also supported by the 2020 Research Fund (1.200098.01) of UNIST (Ulsan National Institute of Science & Technology).

Hahn Seok-Ju, Kim Suhyeon, Choi Young Sik, Lee Junghye, Kang Jihun

2022-Nov-30

East Asian, Genome-wide polygenic risk score, KoGES, Machine learning, Serum metabolites, Type 2 diabetes

Ophthalmology Ophthalmology

Transforming ophthalmology in the digital century-new care models with added value for patients.

In Eye (London, England) ; h5-index 41.0

Ophthalmology faces many challenges in providing effective and meaningful eye care to an ever-increasing group of people. Even health systems that have so far been able to cope with the quantitative patient increase, due to their funding and the availability of highly qualified professionals, and improvements in practice routine efficiency, will be pushed to their limits. Further pressure on care will also be caused by new active substances for the largest group of patients with AMD, the so-called dry form. Treatment availability for this so far untreated group will increase the volume of patients 2-3 times. Without the adaptation of the care structures, this quantitative and qualitative expansion in therapy will inevitably lead to an undersupply.There is increasing scientific evidence that significant efficiency gains in the care of chronic diseases can be achieved through better networking of stakeholders in the healthcare system and greater patient involvement. Digitalization can make an important contribution here. Many technological solutions have been developed in recent years and the time is now ready to exploit this potential. The exceptional setting during the SARS-CoV-2 pandemic has shown many that new technology is available safely, quickly, and effectively. The emergency has catalyzed innovation processes and shown for post-pandemic time after that we are equipped to tackle the challenges in ophthalmic healthcare - ultimately for the benefit of patients and society.

Faes Livia, Maloca Peter M, Hatz Katja, Wolfensberger Thomas J, Munk Marion R, Sim Dawn A, Bachmann Lucas M, Schmid Martin K

2022-Dec-03

General General

Reference panel guided topological structure annotation of Hi-C data.

In Nature communications ; h5-index 260.0

Accurately annotating topological structures (e.g., loops and topologically associating domains) from Hi-C data is critical for understanding the role of 3D genome organization in gene regulation. This is a challenging task, especially at high resolution, in part due to the limited sequencing coverage of Hi-C data. Current approaches focus on the analysis of individual Hi-C data sets of interest, without taking advantage of the facts that (i) several hundred Hi-C contact maps are publicly available, and (ii) the vast majority of topological structures are conserved across multiple cell types. Here, we present RefHiC, an attention-based deep learning framework that uses a reference panel of Hi-C datasets to facilitate topological structure annotation from a given study sample. We compare RefHiC against tools that do not use reference samples and find that RefHiC outperforms other programs at both topological associating domain and loop annotation across different cell types, species, and sequencing depths.

Zhang Yanlin, Blanchette Mathieu

2022-Dec-02

General General

Deciphering clinical abbreviations with a privacy protecting machine learning system.

In Nature communications ; h5-index 260.0

Physicians write clinical notes with abbreviations and shorthand that are difficult to decipher. Abbreviations can be clinical jargon (writing "HIT" for "heparin induced thrombocytopenia"), ambiguous terms that require expertise to disambiguate (using "MS" for "multiple sclerosis" or "mental status"), or domain-specific vernacular ("cb" for "complicated by"). Here we train machine learning models on public web data to decode such text by replacing abbreviations with their meanings. We report a single translation model that simultaneously detects and expands thousands of abbreviations in real clinical notes with accuracies ranging from 92.1%-97.1% on multiple external test datasets. The model equals or exceeds the performance of board-certified physicians (97.6% vs 88.7% total accuracy). Our results demonstrate a general method to contextually decipher abbreviations and shorthand that is built without any privacy-compromising data.

Rajkomar Alvin, Loreaux Eric, Liu Yuchen, Kemp Jonas, Li Benny, Chen Ming-Jun, Zhang Yi, Mohiuddin Afroz, Gottweis Juraj

2022-Dec-02

Ophthalmology Ophthalmology

Transforming ophthalmology in the digital century-new care models with added value for patients.

In Eye (London, England) ; h5-index 41.0

Ophthalmology faces many challenges in providing effective and meaningful eye care to an ever-increasing group of people. Even health systems that have so far been able to cope with the quantitative patient increase, due to their funding and the availability of highly qualified professionals, and improvements in practice routine efficiency, will be pushed to their limits. Further pressure on care will also be caused by new active substances for the largest group of patients with AMD, the so-called dry form. Treatment availability for this so far untreated group will increase the volume of patients 2-3 times. Without the adaptation of the care structures, this quantitative and qualitative expansion in therapy will inevitably lead to an undersupply.There is increasing scientific evidence that significant efficiency gains in the care of chronic diseases can be achieved through better networking of stakeholders in the healthcare system and greater patient involvement. Digitalization can make an important contribution here. Many technological solutions have been developed in recent years and the time is now ready to exploit this potential. The exceptional setting during the SARS-CoV-2 pandemic has shown many that new technology is available safely, quickly, and effectively. The emergency has catalyzed innovation processes and shown for post-pandemic time after that we are equipped to tackle the challenges in ophthalmic healthcare - ultimately for the benefit of patients and society.

Faes Livia, Maloca Peter M, Hatz Katja, Wolfensberger Thomas J, Munk Marion R, Sim Dawn A, Bachmann Lucas M, Schmid Martin K

2022-Dec-03

General General

Using artificial intelligence to improve pain assessment and pain management: a scoping review.

In Journal of the American Medical Informatics Association : JAMIA

CONTEXT : Over 20% of US adults report they experience pain on most days or every day. Uncontrolled pain has led to increased healthcare utilization, hospitalization, emergency visits, and financial burden. Recognizing, assessing, understanding, and treating pain using artificial intelligence (AI) approaches may improve patient outcomes and healthcare resource utilization. A comprehensive synthesis of the current use and outcomes of AI-based interventions focused on pain assessment and management will guide the development of future research.

OBJECTIVES : This review aims to investigate the state of the research on AI-based interventions designed to improve pain assessment and management for adult patients. We also ascertain the actual outcomes of Al-based interventions for adult patients.

METHODS : The electronic databases searched include Web of Science, CINAHL, PsycINFO, Cochrane CENTRAL, Scopus, IEEE Xplore, and ACM Digital Library. The search initially identified 6946 studies. After screening, 30 studies met the inclusion criteria. The Critical Appraisals Skills Programme was used to assess study quality.

RESULTS : This review provides evidence that machine learning, data mining, and natural language processing were used to improve efficient pain recognition and pain assessment, analyze self-reported pain data, predict pain, and help clinicians and patients to manage chronic pain more effectively.

CONCLUSIONS : Findings from this review suggest that using AI-based interventions has a positive effect on pain recognition, pain prediction, and pain self-management; however, most reports are only pilot studies. More pilot studies with physiological pain measures are required before these approaches are ready for large clinical trial.

Zhang Meina, Zhu Linzee, Lin Shih-Yin, Herr Keela, Chi Chih-Lin, Demir Ibrahim, Dunn Lopez Karen, Chi Nai-Ching

2022-Dec-02

artificial intelligence, pain, pain assessment, pain control, pain management

Public Health Public Health

Data-driven identification of post-acute SARS-CoV-2 infection subphenotypes.

In Nature medicine ; h5-index 170.0

The post-acute sequelae of SARS-CoV-2 infection (PASC) refers to a broad spectrum of symptoms and signs that are persistent, exacerbated or newly incident in the period after acute SARS-CoV-2 infection. Most studies have examined these conditions individually without providing evidence on co-occurring conditions. In this study, we leveraged the electronic health record data of two large cohorts, INSIGHT and OneFlorida+, from the national Patient-Centered Clinical Research Network. We created a development cohort from INSIGHT and a validation cohort from OneFlorida+ including 20,881 and 13,724 patients, respectively, who were SARS-CoV-2 infected, and we investigated their newly incident diagnoses 30-180 days after a documented SARS-CoV-2 infection. Through machine learning analysis of over 137 symptoms and conditions, we identified four reproducible PASC subphenotypes, dominated by cardiac and renal (including 33.75% and 25.43% of the patients in the development and validation cohorts); respiratory, sleep and anxiety (32.75% and 38.48%); musculoskeletal and nervous system (23.37% and 23.35%); and digestive and respiratory system (10.14% and 12.74%) sequelae. These subphenotypes were associated with distinct patient demographics, underlying conditions before SARS-CoV-2 infection and acute infection phase severity. Our study provides insights into the heterogeneity of PASC and may inform stratified decision-making in the management of PASC conditions.

Zhang Hao, Zang Chengxi, Xu Zhenxing, Zhang Yongkang, Xu Jie, Bian Jiang, Morozyuk Dmitry, Khullar Dhruv, Zhang Yiye, Nordvig Anna S, Schenck Edward J, Shenkman Elizabeth A, Rothman Russell L, Block Jason P, Lyman Kristin, Weiner Mark G, Carton Thomas W, Wang Fei, Kaushal Rainu

2022-Dec-01

General General

An integrated resource for functional and structural connectivity of the marmoset brain.

In Nature communications ; h5-index 260.0

Comprehensive integration of structural and functional connectivity data is required to model brain functions accurately. While resources for studying the structural connectivity of non-human primate brains already exist, their integration with functional connectivity data has remained unavailable. Here we present a comprehensive resource that integrates the most extensive awake marmoset resting-state fMRI data available to date (39 marmoset monkeys, 710 runs, 12117 mins) with previously published cellular-level neuronal tracing data (52 marmoset monkeys, 143 injections) and multi-resolution diffusion MRI datasets. The combination of these data allowed us to (1) map the fine-detailed functional brain networks and cortical parcellations, (2) develop a deep-learning-based parcellation generator that preserves the topographical organization of functional connectivity and reflects individual variabilities, and (3) investigate the structural basis underlying functional connectivity by computational modeling. This resource will enable modeling structure-function relationships and facilitate future comparative and translational studies of primate brains.

Tian Xiaoguang, Chen Yuyan, Majka Piotr, Szczupak Diego, Perl Yonatan Sanz, Yen Cecil Chern-Chyi, Tong Chuanjun, Feng Furui, Jiang Haiteng, Glen Daniel, Deco Gustavo, Rosa Marcello G P, Silva Afonso C, Liang Zhifeng, Liu Cirong

2022-Dec-01

Public Health Public Health

Data-driven identification of post-acute SARS-CoV-2 infection subphenotypes.

In Nature medicine ; h5-index 170.0

The post-acute sequelae of SARS-CoV-2 infection (PASC) refers to a broad spectrum of symptoms and signs that are persistent, exacerbated or newly incident in the period after acute SARS-CoV-2 infection. Most studies have examined these conditions individually without providing evidence on co-occurring conditions. In this study, we leveraged the electronic health record data of two large cohorts, INSIGHT and OneFlorida+, from the national Patient-Centered Clinical Research Network. We created a development cohort from INSIGHT and a validation cohort from OneFlorida+ including 20,881 and 13,724 patients, respectively, who were SARS-CoV-2 infected, and we investigated their newly incident diagnoses 30-180 days after a documented SARS-CoV-2 infection. Through machine learning analysis of over 137 symptoms and conditions, we identified four reproducible PASC subphenotypes, dominated by cardiac and renal (including 33.75% and 25.43% of the patients in the development and validation cohorts); respiratory, sleep and anxiety (32.75% and 38.48%); musculoskeletal and nervous system (23.37% and 23.35%); and digestive and respiratory system (10.14% and 12.74%) sequelae. These subphenotypes were associated with distinct patient demographics, underlying conditions before SARS-CoV-2 infection and acute infection phase severity. Our study provides insights into the heterogeneity of PASC and may inform stratified decision-making in the management of PASC conditions.

Zhang Hao, Zang Chengxi, Xu Zhenxing, Zhang Yongkang, Xu Jie, Bian Jiang, Morozyuk Dmitry, Khullar Dhruv, Zhang Yiye, Nordvig Anna S, Schenck Edward J, Shenkman Elizabeth A, Rothman Russell L, Block Jason P, Lyman Kristin, Weiner Mark G, Carton Thomas W, Wang Fei, Kaushal Rainu

2022-Dec-01

Public Health Public Health

Ten quick tips for sequence-based prediction of protein properties using machine learning.

In PLoS computational biology

The ubiquitous availability of genome sequencing data explains the popularity of machine learning-based methods for the prediction of protein properties from their amino acid sequences. Over the years, while revising our own work, reading submitted manuscripts as well as published papers, we have noticed several recurring issues, which make some reported findings hard to understand and replicate. We suspect this may be due to biologists being unfamiliar with machine learning methodology, or conversely, machine learning experts may miss some of the knowledge needed to correctly apply their methods to proteins. Here, we aim to bridge this gap for developers of such methods. The most striking issues are linked to a lack of clarity: how were annotations of interest obtained; which benchmark metrics were used; how are positives and negatives defined. Others relate to a lack of rigor: If you sneak in structural information, your method is not sequence-based; if you compare your own model to "state-of-the-art," take the best methods; if you want to conclude that some method is better than another, obtain a significance estimate to support this claim. These, and other issues, we will cover in detail. These points may have seemed obvious to the authors during writing; however, they are not always clear-cut to the readers. We also expect many of these tips to hold for other machine learning-based applications in biology. Therefore, many computational biologists who develop methods in this particular subject will benefit from a concise overview of what to avoid and what to do instead.

Hou Qingzhen, Waury Katharina, Gogishvili Dea, Feenstra K Anton

2022-Dec

Cardiology Cardiology

A comparative study of pretrained language models for long clinical text.

In Journal of the American Medical Informatics Association : JAMIA

OBJECTIVE : Clinical knowledge-enriched transformer models (eg, ClinicalBERT) have state-of-the-art results on clinical natural language processing (NLP) tasks. One of the core limitations of these transformer models is the substantial memory consumption due to their full self-attention mechanism, which leads to the performance degradation in long clinical texts. To overcome this, we propose to leverage long-sequence transformer models (eg, Longformer and BigBird), which extend the maximum input sequence length from 512 to 4096, to enhance the ability to model long-term dependencies in long clinical texts.

MATERIALS AND METHODS : Inspired by the success of long-sequence transformer models and the fact that clinical notes are mostly long, we introduce 2 domain-enriched language models, Clinical-Longformer and Clinical-BigBird, which are pretrained on a large-scale clinical corpus. We evaluate both language models using 10 baseline tasks including named entity recognition, question answering, natural language inference, and document classification tasks.

RESULTS : The results demonstrate that Clinical-Longformer and Clinical-BigBird consistently and significantly outperform ClinicalBERT and other short-sequence transformers in all 10 downstream tasks and achieve new state-of-the-art results.

DISCUSSION : Our pretrained language models provide the bedrock for clinical NLP using long texts. We have made our source code available at https://github.com/luoyuanlab/Clinical-Longformer, and the pretrained models available for public download at: https://huggingface.co/yikuan8/Clinical-Longformer.

CONCLUSION : This study demonstrates that clinical knowledge-enriched long-sequence transformers are able to learn long-term dependencies in long clinical text. Our methods can also inspire the development of other domain-enriched long-sequence transformers.

Li Yikuan, Wehbe Ramsey M, Ahmad Faraz S, Wang Hanyin, Luo Yuan

2022-Nov-30

clinical natural language processing, named entity recognition, natural language inference, question answering, text classification

General General

Dendrocentric learning for synthetic intelligence.

In Nature ; h5-index 368.0

Artificial intelligence now advances by performing twice as many floating-point multiplications every two months, but the semiconductor industry tiles twice as many multipliers on a chip every two years. Moreover, the returns from tiling these multipliers ever more densely now diminish because signals must travel relatively farther and farther. Although travel can be shortened by stacking tiled multipliers in a three-dimensional chip, such a solution acutely reduces the available surface area for dissipating heat. Here I propose to transcend this three-dimensional thermal constraint by moving away from learning with synapses to learning with dendrites. Synaptic inputs are not weighted precisely but rather ordered meticulously along a short stretch of dendrite, termed dendrocentric learning. With the help of a computational model of a dendrite and a conceptual model of a ferroelectric device that emulates it, I illustrate how dendrocentric learning artificial intelligence-or synthetic intelligence for short-could run not with megawatts in the cloud but rather with watts on a smartphone.

Boahen Kwabena

2022-Dec

Public Health Public Health

Analysis of the first genetic engineering attribution challenge.

In Nature communications ; h5-index 260.0

The ability to identify the designer of engineered biological sequences-termed genetic engineering attribution (GEA)-would help ensure due credit for biotechnological innovation, while holding designers accountable to the communities they affect. Here, we present the results of the first Genetic Engineering Attribution Challenge, a public data-science competition to advance GEA techniques. Top-scoring teams dramatically outperformed previous models at identifying the true lab-of-origin of engineered plasmid sequences, including an increase in top-1 and top-10 accuracy of 10 percentage points. A simple ensemble of prizewinning models further increased performance. New metrics, designed to assess a model's ability to confidently exclude candidate labs, also showed major improvements, especially for the ensemble. Most winning teams adopted CNN-based machine-learning approaches; however, one team achieved very high accuracy with an extremely fast neural-network-free approach. Future work, including future competitions, should further explore a wide diversity of approaches for bringing GEA technology into practical use.

Crook Oliver M, Warmbrod Kelsey Lane, Lipstein Greg, Chung Christine, Bakerlee Christopher W, McKelvey T Greg, Holland Shelly R, Swett Jacob L, Esvelt Kevin M, Alley Ethan C, Bradshaw William J

2022-Nov-30

General General

The E3 ubiquitin ligase WWP2 regulates pro-fibrogenic monocyte infiltration and activity in heart fibrosis.

In Nature communications ; h5-index 260.0

Non-ischemic cardiomyopathy (NICM) can cause left ventricular dysfunction through interstitial fibrosis, which corresponds to the failure of cardiac tissue remodeling. Recent evidence implicates monocytes/macrophages in the etiopathology of cardiac fibrosis, but giving their heterogeneity and the antagonizing roles of macrophage subtypes in fibrosis, targeting these cells has been challenging. Here we focus on WWP2, an E3 ubiquitin ligase that acts as a positive genetic regulator of human and murine cardiac fibrosis, and show that myeloid specific deletion of WWP2 reduces cardiac fibrosis in hypertension-induced NICM. By using single cell RNA sequencing analysis of immune cells in the same model, we establish the functional heterogeneity of macrophages and define an early pro-fibrogenic phase of NICM that is driven by Ccl5-expressing Ly6chigh monocytes. Among cardiac macrophage subtypes, WWP2 dysfunction primarily affects Ly6chigh monocytes via modulating Ccl5, and consequentially macrophage infiltration and activation, which contributes to reduced myofibroblast trans-differentiation. WWP2 interacts with transcription factor IRF7, promoting its non-degradative mono-ubiquitination, nuclear translocation and transcriptional activity, leading to upregulation of Ccl5 at transcriptional level. We identify a pro-fibrogenic macrophage subtype in non-ischemic cardiomyopathy, and demonstrate that WWP2 is a key regulator of IRF7-mediated Ccl5/Ly6chigh monocyte axis in heart fibrosis.

Chen Huimei, Chew Gabriel, Devapragash Nithya, Loh Jui Zhi, Huang Kevin Y, Guo Jing, Liu Shiyang, Tan Elisabeth Li Sa, Chen Shuang, Tee Nicole Gui Zhen, Mia Masum M, Singh Manvendra K, Zhang Aihua, Behmoaras Jacques, Petretto Enrico

2022-Nov-30

General General

Chronic stress causes striatal disinhibition mediated by SOM-interneurons in male mice.

In Nature communications ; h5-index 260.0

Chronic stress (CS) is associated with a number of neuropsychiatric disorders, and it may also contribute to or exacerbate motor function. However, the mechanisms by which stress triggers motor symptoms are not fully understood. Here, we report that CS functionally alters dorsomedial striatum (DMS) circuits in male mice, by affecting GABAergic interneuron populations and somatostatin positive (SOM) interneurons in particular. Specifically, we show that CS impairs communication between SOM interneurons and medium spiny neurons, promoting striatal overactivation/disinhibition and increased motor output. Using probabilistic machine learning to analyze animal behavior, we demonstrate that in vivo chemogenetic manipulation of SOM interneurons in DMS modulates motor phenotypes in stressed mice. Altogether, we propose a causal link between dysfunction of striatal SOM interneurons and motor symptoms in models of chronic stress.

Rodrigues Diana, Jacinto Luis, Falcão Margarida, Castro Ana Carolina, Cruz Alexandra, Santa Cátia, Manadas Bruno, Marques Fernanda, Sousa Nuno, Monteiro Patricia

2022-Nov-29

General General

A large-scale neural network training framework for generalized estimation of single-trial population dynamics.

In Nature methods ; h5-index 152.0

Achieving state-of-the-art performance with deep neural population dynamics models requires extensive hyperparameter tuning for each dataset. AutoLFADS is a model-tuning framework that automatically produces high-performing autoencoding models on data from a variety of brain areas and tasks, without behavioral or task information. We demonstrate its broad applicability on several rhesus macaque datasets: from motor cortex during free-paced reaching, somatosensory cortex during reaching with perturbations, and dorsomedial frontal cortex during a cognitive timing task.

Keshtkaran Mohammad Reza, Sedler Andrew R, Chowdhury Raeed H, Tandon Raghav, Basrai Diya, Nguyen Sarah L, Sohn Hansem, Jazayeri Mehrdad, Miller Lee E, Pandarinath Chethan

2022-Nov-28

General General

AlphaFill: enriching AlphaFold models with ligands and cofactors.

In Nature methods ; h5-index 152.0

Artificial intelligence-based protein structure prediction approaches have had a transformative effect on biomolecular sciences. The predicted protein models in the AlphaFold protein structure database, however, all lack coordinates for small molecules, essential for molecular structure or function: hemoglobin lacks bound heme; zinc-finger motifs lack zinc ions essential for structural integrity and metalloproteases lack metal ions needed for catalysis. Ligands important for biological function are absent too; no ADP or ATP is bound to any of the ATPases or kinases. Here we present AlphaFill, an algorithm that uses sequence and structure similarity to 'transplant' such 'missing' small molecules and ions from experimentally determined structures to predicted protein models. The algorithm was successfully validated against experimental structures. A total of 12,029,789 transplants were performed on 995,411 AlphaFold models and are available together with associated validation metrics in the alphafill.eu databank, a resource to help scientists make new hypotheses and design targeted experiments.

Hekkelman Maarten L, de Vries Ida, Joosten Robbie P, Perrakis Anastassis

2022-Nov-24

Radiology Radiology

Automated deidentification of radiology reports combining transformer and "hide in plain sight" rule-based methods.

In Journal of the American Medical Informatics Association : JAMIA

OBJECTIVE : To develop an automated deidentification pipeline for radiology reports that detect protected health information (PHI) entities and replaces them with realistic surrogates "hiding in plain sight."

MATERIALS AND METHODS : In this retrospective study, 999 chest X-ray and CT reports collected between November 2019 and November 2020 were annotated for PHI at the token level and combined with 3001 X-rays and 2193 medical notes previously labeled, forming a large multi-institutional and cross-domain dataset of 6193 documents. Two radiology test sets, from a known and a new institution, as well as i2b2 2006 and 2014 test sets, served as an evaluation set to estimate model performance and to compare it with previously released deidentification tools. Several PHI detection models were developed based on different training datasets, fine-tuning approaches and data augmentation techniques, and a synthetic PHI generation algorithm. These models were compared using metrics such as precision, recall and F1 score, as well as paired samples Wilcoxon tests.

RESULTS : Our best PHI detection model achieves 97.9 F1 score on radiology reports from a known institution, 99.6 from a new institution, 99.5 on i2b2 2006, and 98.9 on i2b2 2014. On reports from a known institution, it achieves 99.1 recall of detecting the core of each PHI span.

DISCUSSION : Our model outperforms all deidentifiers it was compared to on all test sets as well as human labelers on i2b2 2014 data. It enables accurate and automatic deidentification of radiology reports.

CONCLUSIONS : A transformer-based deidentification pipeline can achieve state-of-the-art performance for deidentifying radiology reports and other medical documents.

Chambon Pierre J, Wu Christopher, Steinkamp Jackson M, Adleberg Jason, Cook Tessa S, Langlotz Curtis P

2022-Nov-23

NLP, deidentification, machine learning, radiology, transformer

General General

DeepPROTACs is a deep learning-based targeted degradation predictor for PROTACs.

In Nature communications ; h5-index 260.0

The rational design of PROTACs is difficult due to their obscure structure-activity relationship. This study introduces a deep neural network model - DeepPROTACs to help design potent PROTACs molecules. It can predict the degradation capacity of a proposed PROTAC molecule based on structures of given target protein and E3 ligase. The experimental dataset is mainly collected from PROTAC-DB and appropriately labeled according to the DC50 and Dmax values. In the model of DeepPROTACs, the ligands as well as the ligand binding pockets are generated and represented with graphs and fed into Graph Convolutional Networks for feature extraction. While SMILES representations of linkers are fed into a Bidirectional Long Short-Term Memory layer to generate the features. Experiments show that DeepPROTACs model achieves 77.95% average prediction accuracy and 0.8470 area under receiver operating characteristic curve on the test set. DeepPROTACs is available online at a web server ( https://bailab.siais.shanghaitech.edu.cn/services/deepprotacs/ ) and at github ( https://github.com/fenglei104/DeepPROTACs ).

Li Fenglei, Hu Qiaoyu, Zhang Xianglei, Sun Renhong, Liu Zhuanghua, Wu Sanan, Tian Siyuan, Ma Xinyue, Dai Zhizhuo, Yang Xiaobao, Gao Shenghua, Bai Fang

2022-Nov-21

Radiology Radiology

Emerging Technology in Musculoskeletal MRI and CT.

In Radiology ; h5-index 91.0

This article provides a focused overview of emerging technology in musculoskeletal MRI and CT. These technological advances have primarily focused on decreasing examination times, obtaining higher quality images, providing more convenient and economical imaging alternatives, and improving patient safety through lower radiation doses. New MRI acceleration methods using deep learning and novel reconstruction algorithms can reduce scanning times while maintaining high image quality. New synthetic techniques are now available that provide multiple tissue contrasts from a limited amount of MRI and CT data. Modern low-field-strength MRI scanners can provide a more convenient and economical imaging alternative in clinical practice, while clinical 7.0-T scanners have the potential to maximize image quality. Three-dimensional MRI curved planar reformation and cinematic rendering can provide improved methods for image representation. Photon-counting detector CT can provide lower radiation doses, higher spatial resolution, greater tissue contrast, and reduced noise in comparison with currently used energy-integrating detector CT scanners. Technological advances have also been made in challenging areas of musculoskeletal imaging, including MR neurography, imaging around metal, and dual-energy CT. While the preliminary results of these emerging technologies have been encouraging, whether they result in higher diagnostic performance requires further investigation.

Kijowski Richard, Fritz Jan

2022-Nov-22

Radiology Radiology

χ-Separation Imaging for Diagnosis of Multiple Sclerosis versus Neuromyelitis Optica Spectrum Disorder.

In Radiology ; h5-index 91.0

Background Use of χ-separation imaging can provide surrogates for iron and myelin that relate closely to abnormal changes in multiple sclerosis (MS) lesions. Purpose To evaluate the appearances of MS and neuromyelitis optica spectrum disorder (NMOSD) brain lesions on χ-separation maps and explore their diagnostic value in differentiating the two diseases in comparison with previously reported diagnostic criteria. Materials and Methods This prospective study included individuals with MS or NMOSD who underwent χ-separation imaging from October 2017 to October 2020. Positive (χpos) and negative (χneg) susceptibility were estimated separately by using local frequency shifts and calculating R2' (R2' = R2* - R2). R2 mapping was performed with a machine learning approach. For each lesion, presence of the central vein sign (CVS) and paramagnetic rim sign (PRS) and signal characteristics on χneg and χpos maps were assessed and compared. For each participant, the proportion of lesions with CVS, PRS, and hypodiamagnetism was calculated. Diagnostic performances were assessed using receiver operating characteristic (ROC) curve analysis. Results A total of 32 participants with MS (mean age, 34 years ± 10 [SD]; 25 women, seven men) and 15 with NMOSD (mean age, 52 years ± 17; 14 women, one man) were evaluated, with a total of 611 MS and 225 NMOSD brain lesions. On the χneg maps, 80.2% (490 of 611) of MS lesions were categorized as hypodiamagnetic versus 13.8% (31 of 225) of NMOSD lesions (P < .001). Lesion appearances on the χpos maps showed no evidence of a difference between the two diseases. In per-participant analysis, participants with MS showed a higher proportion of hypodiamagnetic lesions (83%; IQR, 72-93) than those with NMOSD (6%; IQR, 0-14; P < .001). The proportion of hypodiamagnetic lesions achieved excellent diagnostic performance (area under the ROC curve, 0.96; 95% CI: 0.91, 1.00). Conclusion On χ-separation maps, multiple sclerosis (MS) lesions tend to be hypodiamagnetic, which can serve as an important hallmark to differentiate MS from neuromyelitis optica spectrum disorder. © RSNA, 2022 Online supplemental material is available for this article.

Kim Woojun, Shin Hyeong-Geol, Lee Hyebin, Park Dohoon, Kang Junghwa, Nam Yoonho, Lee Jongho, Jang Jinhee

2022-Nov-22

General General

Machine learning approaches for electronic health records phenotyping: a methodical review.

In Journal of the American Medical Informatics Association : JAMIA

OBJECTIVE : Accurate and rapid phenotyping is a prerequisite to leveraging electronic health records for biomedical research. While early phenotyping relied on rule-based algorithms curated by experts, machine learning (ML) approaches have emerged as an alternative to improve scalability across phenotypes and healthcare settings. This study evaluates ML-based phenotyping with respect to (1) the data sources used, (2) the phenotypes considered, (3) the methods applied, and (4) the reporting and evaluation methods used.

MATERIALS AND METHODS : We searched PubMed and Web of Science for articles published between 2018 and 2022. After screening 850 articles, we recorded 37 variables on 100 studies.

RESULTS : Most studies utilized data from a single institution and included information in clinical notes. Although chronic conditions were most commonly considered, ML also enabled the characterization of nuanced phenotypes such as social determinants of health. Supervised deep learning was the most popular ML paradigm, while semi-supervised and weakly supervised learning were applied to expedite algorithm development and unsupervised learning to facilitate phenotype discovery. ML approaches did not uniformly outperform rule-based algorithms, but deep learning offered a marginal improvement over traditional ML for many conditions.

DISCUSSION : Despite the progress in ML-based phenotyping, most articles focused on binary phenotypes and few articles evaluated external validity or used multi-institution data. Study settings were infrequently reported and analytic code was rarely released.

CONCLUSION : Continued research in ML-based phenotyping is warranted, with emphasis on characterizing nuanced phenotypes, establishing reporting and evaluation standards, and developing methods to accommodate misclassified phenotypes due to algorithm errors in downstream applications.

Yang Siyue, Varghese Paul, Stephenson Ellen, Tu Karen, Gronsbell Jessica

2022-Nov-22

cohort identification, electronic health records, machine learning, phenotyping

General General

Performance drift in a mortality prediction algorithm among patients with cancer during the SARS-CoV-2 pandemic.

In Journal of the American Medical Informatics Association : JAMIA

Sudden changes in health care utilization during the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) pandemic may have impacted the performance of clinical predictive models that were trained prior to the pandemic. In this study, we evaluated the performance over time of a machine learning, electronic health record-based mortality prediction algorithm currently used in clinical practice to identify patients with cancer who may benefit from early advance care planning conversations. We show that during the pandemic period, algorithm identification of high-risk patients had a substantial and sustained decline. Decreases in laboratory utilization during the peak of the pandemic may have contributed to drift. Calibration and overall discrimination did not markedly decline during the pandemic. This argues for careful attention to the performance and retraining of predictive algorithms that use inputs from the pandemic period.

Parikh Ravi B, Zhang Yichen, Kolla Likhitha, Chivers Corey, Courtright Katherine R, Zhu Jingsan, Navathe Amol S, Chen Jinbo

2022-Nov-21

SARS-CoV-2, algorithm drift, cancer, machine learning, mortality

General General

Performance drift in a mortality prediction algorithm among patients with cancer during the SARS-CoV-2 pandemic.

In Journal of the American Medical Informatics Association : JAMIA

Sudden changes in health care utilization during the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) pandemic may have impacted the performance of clinical predictive models that were trained prior to the pandemic. In this study, we evaluated the performance over time of a machine learning, electronic health record-based mortality prediction algorithm currently used in clinical practice to identify patients with cancer who may benefit from early advance care planning conversations. We show that during the pandemic period, algorithm identification of high-risk patients had a substantial and sustained decline. Decreases in laboratory utilization during the peak of the pandemic may have contributed to drift. Calibration and overall discrimination did not markedly decline during the pandemic. This argues for careful attention to the performance and retraining of predictive algorithms that use inputs from the pandemic period.

Parikh Ravi B, Zhang Yichen, Kolla Likhitha, Chivers Corey, Courtright Katherine R, Zhu Jingsan, Navathe Amol S, Chen Jinbo

2022-Nov-21

SARS-CoV-2, algorithm drift, cancer, machine learning, mortality

General General

Deep reinforcement learning for optimal experimental design in biology.

In PLoS computational biology

The field of optimal experimental design uses mathematical techniques to determine experiments that are maximally informative from a given experimental setup. Here we apply a technique from artificial intelligence-reinforcement learning-to the optimal experimental design task of maximizing confidence in estimates of model parameter values. We show that a reinforcement learning approach performs favourably in comparison with a one-step ahead optimisation algorithm and a model predictive controller for the inference of bacterial growth parameters in a simulated chemostat. Further, we demonstrate the ability of reinforcement learning to train over a distribution of parameters, indicating that this approach is robust to parametric uncertainty.

Treloar Neythen J, Braniff Nathan, Ingalls Brian, Barnes Chris P

2022-Nov-21

Radiology Radiology

External Validation of an Ensemble Model for Automated Mammography Interpretation by Artificial Intelligence.

In JAMA network open

Importance : With a shortfall in fellowship-trained breast radiologists, mammography screening programs are looking toward artificial intelligence (AI) to increase efficiency and diagnostic accuracy. External validation studies provide an initial assessment of how promising AI algorithms perform in different practice settings.

Objective : To externally validate an ensemble deep-learning model using data from a high-volume, distributed screening program of an academic health system with a diverse patient population.

Design, Setting, and Participants : In this diagnostic study, an ensemble learning method, which reweights outputs of the 11 highest-performing individual AI models from the Digital Mammography Dialogue on Reverse Engineering Assessment and Methods (DREAM) Mammography Challenge, was used to predict the cancer status of an individual using a standard set of screening mammography images. This study was conducted using retrospective patient data collected between 2010 and 2020 from women aged 40 years and older who underwent a routine breast screening examination and participated in the Athena Breast Health Network at the University of California, Los Angeles (UCLA).

Main Outcomes and Measures : Performance of the challenge ensemble method (CEM) and the CEM combined with radiologist assessment (CEM+R) were compared with diagnosed ductal carcinoma in situ and invasive cancers within a year of the screening examination using performance metrics, such as sensitivity, specificity, and area under the receiver operating characteristic curve (AUROC).

Results : Evaluated on 37 317 examinations from 26 817 women (mean [SD] age, 58.4 [11.5] years), individual model AUROC estimates ranged from 0.77 (95% CI, 0.75-0.79) to 0.83 (95% CI, 0.81-0.85). The CEM model achieved an AUROC of 0.85 (95% CI, 0.84-0.87) in the UCLA cohort, lower than the performance achieved in the Kaiser Permanente Washington (AUROC, 0.90) and Karolinska Institute (AUROC, 0.92) cohorts. The CEM+R model achieved a sensitivity (0.813 [95% CI, 0.781-0.843] vs 0.826 [95% CI, 0.795-0.856]; P = .20) and specificity (0.925 [95% CI, 0.916-0.934] vs 0.930 [95% CI, 0.929-0.932]; P = .18) similar to the radiologist performance. The CEM+R model had significantly lower sensitivity (0.596 [95% CI, 0.466-0.717] vs 0.850 [95% CI, 0.766-0.923]; P < .001) and specificity (0.803 [95% CI, 0.734-0.861] vs 0.945 [95% CI, 0.936-0.954]; P < .001) than the radiologist in women with a prior history of breast cancer and Hispanic women (0.894 [95% CI, 0.873-0.910] vs 0.926 [95% CI, 0.919-0.933]; P = .004).

Conclusions and Relevance : This study found that the high performance of an ensemble deep-learning model for automated screening mammography interpretation did not generalize to a more diverse screening cohort, suggesting that the model experienced underspecification. This study suggests the need for model transparency and fine-tuning of AI models for specific target populations prior to their clinical adoption.

Hsu William, Hippe Daniel S, Nakhaei Noor, Wang Pin-Chieh, Zhu Bing, Siu Nathan, Ahsen Mehmet Eren, Lotter William, Sorensen A Gregory, Naeim Arash, Buist Diana S M, Schaffter Thomas, Guinney Justin, Elmore Joann G, Lee Christoph I

2022-Nov-01

Radiology Radiology

Cardiovascular CT, MRI, and PET/CT in 2021: Review of Key Articles.

In Radiology ; h5-index 91.0

This review focuses on three key noninvasive cardiac imaging modalities-cardiac CT angiography (CTA), MRI, and PET/CT-and summarizes key publications in 2021 relevant to radiologists in clinical practice. Although this review focuses primarily on articles published in Radiology, important studies from other major journals are included to highlight "must-know" articles in the field of cardiovascular imaging. Cardiac CTA has been established as the first-line test for patients with stable chest pain and no known coronary artery disease, and its value remains central to the assessment of surgical or transcatheter aortic valve replacement. Artificial intelligence continues to evolve in a number of applications in cardiovascular disease. In cardiac MRI studies, 2021 has seen an emphasis on nonischemic cardiomyopathies, valvular heart disease, and COVID-19 disease cardiac manifestations and the authors highlight the key articles on these topics. A section featuring the increasing role of cardiac PET/CT in the assessment of cardiac sarcoidosis and prosthetic valves is also provided.

Tzimas Georgios, Ryan David T, Murphy David J, Leipsic Jonathon A, Dodd Jonathan D

2022-Nov-15

Radiology Radiology

Cardiovascular CT, MRI, and PET/CT in 2021: Review of Key Articles.

In Radiology ; h5-index 91.0

This review focuses on three key noninvasive cardiac imaging modalities-cardiac CT angiography (CTA), MRI, and PET/CT-and summarizes key publications in 2021 relevant to radiologists in clinical practice. Although this review focuses primarily on articles published in Radiology, important studies from other major journals are included to highlight "must-know" articles in the field of cardiovascular imaging. Cardiac CTA has been established as the first-line test for patients with stable chest pain and no known coronary artery disease, and its value remains central to the assessment of surgical or transcatheter aortic valve replacement. Artificial intelligence continues to evolve in a number of applications in cardiovascular disease. In cardiac MRI studies, 2021 has seen an emphasis on nonischemic cardiomyopathies, valvular heart disease, and COVID-19 disease cardiac manifestations and the authors highlight the key articles on these topics. A section featuring the increasing role of cardiac PET/CT in the assessment of cardiac sarcoidosis and prosthetic valves is also provided.

Tzimas Georgios, Ryan David T, Murphy David J, Leipsic Jonathon A, Dodd Jonathan D

2022-Nov-15

Radiology Radiology

Deep Learning to Simulate Contrast-enhanced Breast MRI of Invasive Breast Cancer.

In Radiology ; h5-index 91.0

Background There is increasing interest in noncontrast breast MRI alternatives for tumor visualization to increase the accessibility of breast MRI. Purpose To evaluate the feasibility and accuracy of generating simulated contrast-enhanced T1-weighted breast MRI scans from precontrast MRI sequences in biopsy-proven invasive breast cancer with use of deep learning. Materials and Methods Women with invasive breast cancer and a contrast-enhanced breast MRI examination that was performed for initial evaluation of the extent of disease between January 2015 and December 2019 at a single academic institution were retrospectively identified. A three-dimensional, fully convolutional deep neural network simulated contrast-enhanced T1-weighted breast MRI scans from five precontrast sequences (T1-weighted non-fat-suppressed [FS], T1-weighted FS, T2-weighted FS, apparent diffusion coefficient, and diffusion-weighted imaging). For qualitative assessment, four breast radiologists (with 3-15 years of experience) blinded to whether the method of contrast was real or simulated assessed image quality (excellent, acceptable, good, poor, or unacceptable), presence of tumor enhancement, and maximum index mass size by using 22 pairs of real and simulated contrast-enhanced MRI scans. Quantitative comparison was performed using whole-breast similarity and error metrics and Dice coefficient analysis of enhancing tumor overlap. Results Ninety-six MRI examinations in 96 women (mean age, 52 years ± 12 [SD]) were evaluated. The readers assessed all simulated MRI scans as having the appearance of a real MRI scan with tumor enhancement. Index mass sizes on real and simulated MRI scans demonstrated good to excellent agreement (intraclass correlation coefficient, 0.73-0.86; P < .001) without significant differences (mean differences, -0.8 to 0.8 mm; P = .36-.80). Almost all simulated MRI scans (84 of 88 [95%]) were considered of diagnostic quality (ratings of excellent, acceptable, or good). Quantitative analysis demonstrated strong similarity (structural similarity index, 0.88 ± 0.05), low voxel-wise error (symmetric mean absolute percent error, 3.26%), and Dice coefficient of enhancing tumor overlap of 0.75 ± 0.25. Conclusion It is feasible to generate simulated contrast-enhanced breast MRI scans with use of deep learning. Simulated and real contrast-enhanced MRI scans demonstrated comparable tumor sizes, areas of tumor enhancement, and image quality without significant qualitative or quantitative differences. © RSNA, 2022 Online supplemental material is available for this article. See also the editorial by Slanetz in this issue.

Chung Maggie, Calabrese Evan, Mongan John, Ray Kimberly M, Hayward Jessica H, Kelil Tatiana, Sieberg Ryan, Hylton Nola, Joe Bonnie N, Lee Amie Y

2022-Nov-15

General General

Diachronic and synchronic variation in the performance of adaptive machine learning systems: the ethical challenges.

In Journal of the American Medical Informatics Association : JAMIA

OBJECTIVES : Machine learning (ML) has the potential to facilitate "continual learning" in medicine, in which an ML system continues to evolve in response to exposure to new data over time, even after being deployed in a clinical setting. In this article, we provide a tutorial on the range of ethical issues raised by the use of such "adaptive" ML systems in medicine that have, thus far, been neglected in the literature.

TARGET AUDIENCE : The target audiences for this tutorial are the developers of ML AI systems, healthcare regulators, the broader medical informatics community, and practicing clinicians.

SCOPE : Discussions of adaptive ML systems to date have overlooked the distinction between 2 sorts of variance that such systems may exhibit-diachronic evolution (change over time) and synchronic variation (difference between cotemporaneous instantiations of the algorithm at different sites)-and underestimated the significance of the latter. We highlight the challenges that diachronic evolution and synchronic variation present for the quality of patient care, informed consent, and equity, and discuss the complex ethical trade-offs involved in the design of such systems.

Hatherley Joshua, Sparrow Robert

2022-Nov-15

artificial intelligence, bioethics, federated learning, medicine, update problem

General General

Using automated methods to detect safety problems with health information technology: a scoping review.

In Journal of the American Medical Informatics Association : JAMIA

OBJECTIVE : To summarize the research literature evaluating automated methods for early detection of safety problems with health information technology (HIT).

MATERIALS AND METHODS : We searched bibliographic databases including MEDLINE, ACM Digital, Embase, CINAHL Complete, PsycINFO, and Web of Science from January 2010 to June 2021 for studies evaluating the performance of automated methods to detect HIT problems. HIT problems were reviewed using an existing classification for safety concerns. Automated methods were categorized into rule-based, statistical, and machine learning methods, and their performance in detecting HIT problems was assessed. The review was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta Analyses extension for Scoping Reviews statement.

RESULTS : Of the 45 studies identified, the majority (n = 27, 60%) focused on detecting use errors involving electronic health records and order entry systems. Machine learning (n = 22) and statistical modeling (n = 17) were the most common methods. Unsupervised learning was used to detect use errors in laboratory test results, prescriptions, and patient records while supervised learning was used to detect technical errors arising from hardware or software issues. Statistical modeling was used to detect use errors, unauthorized access, and clinical decision support system malfunctions while rule-based methods primarily focused on use errors.

CONCLUSIONS : A wide variety of rule-based, statistical, and machine learning methods have been applied to automate the detection of safety problems with HIT. Many opportunities remain to systematically study their application and effectiveness in real-world settings.

Surian Didi, Wang Ying, Coiera Enrico, Magrabi Farah

2022-Nov-14

equipment failure analysis, health information technology, patient safety, review

Radiology Radiology

Methods for Clinical Evaluation of Artificial Intelligence Algorithms for Medical Diagnosis.

In Radiology ; h5-index 91.0

Adequate clinical evaluation of artificial intelligence (AI) algorithms before adoption in practice is critical. Clinical evaluation aims to confirm acceptable AI performance through adequate external testing and confirm the benefits of AI-assisted care compared with conventional care through appropriately designed and conducted studies, for which prospective studies are desirable. This article explains some of the fundamental methodological points that should be considered when designing and appraising the clinical evaluation of AI algorithms for medical diagnosis. The specific topics addressed include the following: (a) the importance of external testing of AI algorithms and strategies for conducting the external testing effectively, (b) the various metrics and graphical methods for evaluating the AI performance as well as essential methodological points to note in using and interpreting them, (c) paired study designs primarily for comparative performance evaluation of conventional and AI-assisted diagnoses, (d) parallel study designs primarily for evaluating the effect of AI intervention with an emphasis on randomized clinical trials, and (e) up-to-date guidelines for reporting clinical studies on AI, with an emphasis on guidelines registered in the EQUATOR Network library. Sound methodological knowledge of these topics will aid the design, execution, reporting, and appraisal of clinical evaluation of AI.

Park Seong Ho, Han Kyunghwa, Jang Hye Young, Park Ji Eun, Lee June-Goo, Kim Dong Wook, Choi Jaesoon

2022-Nov-08

General General

Using similar patients to predict complication in patients with diabetes, hypertension, and lipid disorder: a domain knowledge-infused convolutional neural network approach.

In Journal of the American Medical Informatics Association : JAMIA

OBJECTIVE : This study aims to develop a convolutional neural network-based learning framework called domain knowledge-infused convolutional neural network (DK-CNN) for retrieving clinically similar patient and to personalize the prediction of macrovascular complication using the retrieved patients.

MATERIALS AND METHODS : We use the electronic health records of 169 434 patients with diabetes, hypertension, and/or lipid disorder. Patients are partitioned into 7 subcohorts based on their comorbidities. DK-CNN integrates both domain knowledge and disease trajectory of patients over multiple visits to retrieve similar patients. We use normalized discounted cumulative gain (nDCG) and macrovascular complication prediction performance to evaluate the effectiveness of DK-CNN compared to state-of-the-art models. Ablation studies are conducted to compare DK-CNN with reduced models that do not use domain knowledge as well as models that do not consider short-term, medium-term, and long-term trajectory over multiple visits.

RESULTS : Key findings from this study are: (1) DK-CNN is able to retrieve clinically similar patients and achieves the highest nDCG values in all 7 subcohorts; (2) DK-CNN outperforms other state-of-the-art approaches in terms of complication prediction performance in all 7 subcohorts; and (3) the ablation studies show that the full model achieves the highest nDCG compared with other 2 reduced models.

DISCUSSION AND CONCLUSIONS : DK-CNN is a deep learning-based approach which incorporates domain knowledge and patient trajectory data to retrieve clinically similar patients. It can be used to assist physicians who may refer to the outcomes and past treatments of similar patients as a guide for choosing an effective treatment for patients.

Oei Ronald Wihal, Hsu Wynne, Lee Mong Li, Tan Ngiap Chuan

2022-Nov-07

chronic diseases, convolutional neural network, domain knowledge, patient similarity

Ophthalmology Ophthalmology

Systematic analysis and prediction of genes associated with monogenic disorders on human chromosome X.

In Nature communications ; h5-index 260.0

Disease gene discovery on chromosome (chr) X is challenging owing to its unique modes of inheritance. We undertook a systematic analysis of human chrX genes. We observe a higher proportion of disorder-associated genes and an enrichment of genes involved in cognition, language, and seizures on chrX compared to autosomes. We analyze gene constraints, exon and promoter conservation, expression, and paralogues, and report 127 genes sharing one or more attributes with known chrX disorder genes. Using machine learning classifiers trained to distinguish disease-associated from dispensable genes, we classify 247 genes, including 115 of the 127, as having high probability of being disease-associated. We provide evidence of an excess of variants in predicted genes in existing databases. Finally, we report damaging variants in CDK16 and TRPC5 in patients with intellectual disability or autism spectrum disorders. This study predicts large-scale gene-disease associations that could be used for prioritization of X-linked pathogenic variants.

Leitão Elsa, Schröder Christopher, Parenti Ilaria, Dalle Carine, Rastetter Agnès, Kühnel Theresa, Kuechler Alma, Kaya Sabine, Gérard Bénédicte, Schaefer Elise, Nava Caroline, Drouot Nathalie, Engel Camille, Piard Juliette, Duban-Bedu Bénédicte, Villard Laurent, Stegmann Alexander P A, Vanhoutte Els K, Verdonschot Job A J, Kaiser Frank J, Tran Mau-Them Frédéric, Scala Marcello, Striano Pasquale, Frints Suzanna G M, Argilli Emanuela, Sherr Elliott H, Elder Fikret, Buratti Julien, Keren Boris, Mignot Cyril, Héron Delphine, Mandel Jean-Louis, Gecz Jozef, Kalscheuer Vera M, Horsthemke Bernhard, Piton Amélie, Depienne Christel

2022-Nov-02

oncology Oncology

Deep learning empowered volume delineation of whole-body organs-at-risk for accelerated radiotherapy.

In Nature communications ; h5-index 260.0

In radiotherapy for cancer patients, an indispensable process is to delineate organs-at-risk (OARs) and tumors. However, it is the most time-consuming step as manual delineation is always required from radiation oncologists. Herein, we propose a lightweight deep learning framework for radiotherapy treatment planning (RTP), named RTP-Net, to promote an automatic, rapid, and precise initialization of whole-body OARs and tumors. Briefly, the framework implements a cascade coarse-to-fine segmentation, with adaptive module for both small and large organs, and attention mechanisms for organs and boundaries. Our experiments show three merits: 1) Extensively evaluates on 67 delineation tasks on a large-scale dataset of 28,581 cases; 2) Demonstrates comparable or superior accuracy with an average Dice of 0.95; 3) Achieves near real-time delineation in most tasks with <2 s. This framework could be utilized to accelerate the contouring process in the All-in-One radiotherapy scheme, and thus greatly shorten the turnaround time of patients.

Shi Feng, Hu Weigang, Wu Jiaojiao, Han Miaofei, Wang Jiazhou, Zhang Wei, Zhou Qing, Zhou Jingjie, Wei Ying, Shao Ying, Chen Yanbo, Yu Yue, Cao Xiaohuan, Zhan Yiqiang, Zhou Xiang Sean, Gao Yaozong, Shen Dinggang

2022-Nov-02

oncology Oncology

Uncertainty-informed deep learning models enable high-confidence predictions for digital histopathology.

In Nature communications ; h5-index 260.0

A model's ability to express its own predictive uncertainty is an essential attribute for maintaining clinical user confidence as computational biomarkers are deployed into real-world medical settings. In the domain of cancer digital histopathology, we describe a clinically-oriented approach to uncertainty quantification for whole-slide images, estimating uncertainty using dropout and calculating thresholds on training data to establish cutoffs for low- and high-confidence predictions. We train models to identify lung adenocarcinoma vs. squamous cell carcinoma and show that high-confidence predictions outperform predictions without uncertainty, in both cross-validation and testing on two large external datasets spanning multiple institutions. Our testing strategy closely approximates real-world application, with predictions generated on unsupervised, unannotated slides using predetermined thresholds. Furthermore, we show that uncertainty thresholding remains reliable in the setting of domain shift, with accurate high-confidence predictions of adenocarcinoma vs. squamous cell carcinoma for out-of-distribution, non-lung cancer cohorts.

Dolezal James M, Srisuwananukorn Andrew, Karpeyev Dmitry, Ramesh Siddhi, Kochanny Sara, Cody Brittany, Mansfield Aaron S, Rakshit Sagar, Bansal Radhika, Bois Melanie C, Bungum Aaron O, Schulte Jefree J, Vokes Everett E, Garassino Marina Chiara, Husain Aliya N, Pearson Alexander T

2022-Nov-02

Public Health Public Health

An immune dysfunction score for stratification of patients with acute infection based on whole-blood gene expression.

In Science translational medicine ; h5-index 138.0

Dysregulated host responses to infection can lead to organ dysfunction and sepsis, causing millions of global deaths each year. To alleviate this burden, improved prognostication and biomarkers of response are urgently needed. We investigated the use of whole-blood transcriptomics for stratification of patients with severe infection by integrating data from 3149 samples from patients with sepsis due to community-acquired pneumonia or fecal peritonitis admitted to intensive care and healthy individuals into a gene expression reference map. We used this map to derive a quantitative sepsis response signature (SRSq) score reflective of immune dysfunction and predictive of clinical outcomes, which can be estimated using a 7- or 12-gene signature. Last, we built a machine learning framework, SepstratifieR, to deploy SRSq in adult and pediatric bacterial and viral sepsis, H1N1 influenza, and COVID-19, demonstrating clinically relevant stratification across diseases and revealing some of the physiological alterations linking immune dysregulation to mortality. Our method enables early identification of individuals with dysfunctional immune profiles, bringing us closer to precision medicine in infection.

Cano-Gamez Eddie, Burnham Katie L, Goh Cyndi, Allcock Alice, Malick Zunaira H, Overend Lauren, Kwok Andrew, Smith David A, Peters-Sengers Hessel, Antcliffe David, McKechnie Stuart, Scicluna Brendon P, van der Poll Tom, Gordon Anthony C, Hinds Charles J, Davenport Emma E, Knight Julian C, Webster Nigel, Galley Helen, Taylor Jane, Hall Sally, Addison Jenni, Roughton Sian, Tennant Heather, Guleri Achyut, Waddington Natalia, Arawwawala Dilshan, Durcan John, Short Alasdair, Swan Karen, Williams Sarah, Smolen Susan, Mitchell-Inwang Christine, Gordon Tony, Errington Emily, Templeton Maie, Venatesh Pyda, Ward Geraldine, McCauley Marie, Baudouin Simon, Higham Charley, Soar Jasmeet, Grier Sally, Hall Elaine, Brett Stephen, Kitson David, Wilson Robert, Mountford Laura, Moreno Juan, Hall Peter, Hewlett Jackie, McKechnie Stuart, Garrard Christopher, Millo Julian, Young Duncan, Hutton Paula, Parsons Penny, Smiths Alex, Faras-Arraya Roser, Soar Jasmeet, Raymode Parizade, Thompson Jonathan, Bowrey Sarah, Kazembe Sandra, Rich Natalie, Andreou Prem, Hales Dawn, Roberts Emma, Fletcher Simon, Rosbergen Melissa, Glister Georgina, Cuesta Jeronimo Moreno, Bion Julian, Millar Joanne, Perry Elsa Jane, Willis Heather, Mitchell Natalie, Ruel Sebastian, Carrera Ronald, Wilde Jude, Nilson Annette, Lees Sarah, Kapila Atul, Jacques Nicola, Atkinson Jane, Brown Abby, Prowse Heather, Krige Anton, Bland Martin, Bullock Lynne, Harrison Donna, Mills Gary, Humphreys John, Armitage Kelsey, Laha Shond, Baldwin Jacqueline, Walsh Angela, Doherty Nicola, Drage Stephen, Ortiz-Ruiz de Gordoa Laura, Lowes Sarah, Higham Charley, Walsh Helen, Calder Verity, Swan Catherine, Payne Heather, Higgins David, Andrews Sarah, Mappleback Sarah, Hind Charles, Garrard Chris, Watson D, McLees Eleanor, Purdy Alice, Stotz Martin, Ochelli-Okpue Adaeze, Bonner Stephen, Whitehead Iain, Hugil Keith, Goodridge Victoria, Cawthor Louisa, Kuper Martin, Pahary Sheik, Bellingan Geoffrey, Marshall Richard, Montgomery Hugh, Ryu Jung Hyun, Bercades Georgia, Boluda Susan, Bentley Andrew, Mccalman Katie, Jefferies Fiona, Knight Julian, Davenport Emma, Burnham Katie, Maugeri Narelle, Radhakrishnan Jayachandran, Mi Yuxin, Allcock Alice, Goh Cyndi

2022-Nov-02

Public Health Public Health

An immune dysfunction score for stratification of patients with acute infection based on whole-blood gene expression.

In Science translational medicine ; h5-index 138.0

Dysregulated host responses to infection can lead to organ dysfunction and sepsis, causing millions of global deaths each year. To alleviate this burden, improved prognostication and biomarkers of response are urgently needed. We investigated the use of whole-blood transcriptomics for stratification of patients with severe infection by integrating data from 3149 samples from patients with sepsis due to community-acquired pneumonia or fecal peritonitis admitted to intensive care and healthy individuals into a gene expression reference map. We used this map to derive a quantitative sepsis response signature (SRSq) score reflective of immune dysfunction and predictive of clinical outcomes, which can be estimated using a 7- or 12-gene signature. Last, we built a machine learning framework, SepstratifieR, to deploy SRSq in adult and pediatric bacterial and viral sepsis, H1N1 influenza, and COVID-19, demonstrating clinically relevant stratification across diseases and revealing some of the physiological alterations linking immune dysregulation to mortality. Our method enables early identification of individuals with dysfunctional immune profiles, bringing us closer to precision medicine in infection.

Cano-Gamez Eddie, Burnham Katie L, Goh Cyndi, Allcock Alice, Malick Zunaira H, Overend Lauren, Kwok Andrew, Smith David A, Peters-Sengers Hessel, Antcliffe David, McKechnie Stuart, Scicluna Brendon P, van der Poll Tom, Gordon Anthony C, Hinds Charles J, Davenport Emma E, Knight Julian C, Webster Nigel, Galley Helen, Taylor Jane, Hall Sally, Addison Jenni, Roughton Sian, Tennant Heather, Guleri Achyut, Waddington Natalia, Arawwawala Dilshan, Durcan John, Short Alasdair, Swan Karen, Williams Sarah, Smolen Susan, Mitchell-Inwang Christine, Gordon Tony, Errington Emily, Templeton Maie, Venatesh Pyda, Ward Geraldine, McCauley Marie, Baudouin Simon, Higham Charley, Soar Jasmeet, Grier Sally, Hall Elaine, Brett Stephen, Kitson David, Wilson Robert, Mountford Laura, Moreno Juan, Hall Peter, Hewlett Jackie, McKechnie Stuart, Garrard Christopher, Millo Julian, Young Duncan, Hutton Paula, Parsons Penny, Smiths Alex, Faras-Arraya Roser, Soar Jasmeet, Raymode Parizade, Thompson Jonathan, Bowrey Sarah, Kazembe Sandra, Rich Natalie, Andreou Prem, Hales Dawn, Roberts Emma, Fletcher Simon, Rosbergen Melissa, Glister Georgina, Cuesta Jeronimo Moreno, Bion Julian, Millar Joanne, Perry Elsa Jane, Willis Heather, Mitchell Natalie, Ruel Sebastian, Carrera Ronald, Wilde Jude, Nilson Annette, Lees Sarah, Kapila Atul, Jacques Nicola, Atkinson Jane, Brown Abby, Prowse Heather, Krige Anton, Bland Martin, Bullock Lynne, Harrison Donna, Mills Gary, Humphreys John, Armitage Kelsey, Laha Shond, Baldwin Jacqueline, Walsh Angela, Doherty Nicola, Drage Stephen, Ortiz-Ruiz de Gordoa Laura, Lowes Sarah, Higham Charley, Walsh Helen, Calder Verity, Swan Catherine, Payne Heather, Higgins David, Andrews Sarah, Mappleback Sarah, Hind Charles, Garrard Chris, Watson D, McLees Eleanor, Purdy Alice, Stotz Martin, Ochelli-Okpue Adaeze, Bonner Stephen, Whitehead Iain, Hugil Keith, Goodridge Victoria, Cawthor Louisa, Kuper Martin, Pahary Sheik, Bellingan Geoffrey, Marshall Richard, Montgomery Hugh, Ryu Jung Hyun, Bercades Georgia, Boluda Susan, Bentley Andrew, Mccalman Katie, Jefferies Fiona, Knight Julian, Davenport Emma, Burnham Katie, Maugeri Narelle, Radhakrishnan Jayachandran, Mi Yuxin, Allcock Alice, Goh Cyndi

2022-Nov-02

General General

Unsupervised learning of aging principles from longitudinal data.

In Nature communications ; h5-index 260.0

Age is the leading risk factor for prevalent diseases and death. However, the relation between age-related physiological changes and lifespan is poorly understood. We combined analytical and machine learning tools to describe the aging process in large sets of longitudinal measurements. Assuming that aging results from a dynamic instability of the organism state, we designed a deep artificial neural network, including auto-encoder and auto-regression (AR) components. The AR model tied the dynamics of physiological state with the stochastic evolution of a single variable, the "dynamic frailty indicator" (dFI). In a subset of blood tests from the Mouse Phenome Database, dFI increased exponentially and predicted the remaining lifespan. The observation of the limiting dFI was consistent with the late-life mortality deceleration. dFI changed along with hallmarks of aging, including frailty index, molecular markers of inflammation, senescent cell accumulation, and responded to life-shortening (high-fat diet) and life-extending (rapamycin) treatments.

Avchaciov Konstantin, Antoch Marina P, Andrianova Ekaterina L, Tarkhov Andrei E, Menshikov Leonid I, Burmistrova Olga, Gudkov Andrei V, Fedichev Peter O

2022-Nov-01

Radiology Radiology

Deep Learning Reconstruction for Accelerated Spine MRI: Prospective Analysis of Interchangeability.

In Radiology ; h5-index 91.0

Background Deep learning (DL)-based MRI reconstructions can reduce examination times for turbo spin-echo (TSE) acquisitions. Studies that prospectively employ DL-based reconstructions of rapidly acquired, undersampled spine MRI are needed. Purpose To investigate the diagnostic interchangeability of an unrolled DL-reconstructed TSE (hereafter, TSEDL) T1- and T2-weighted acquisition method with standard TSE and to test their impact on acquisition time, image quality, and diagnostic confidence. Materials and Methods This prospective single-center study included participants with various spinal abnormalities who gave written consent from November 2020 to July 2021. Each participant underwent two MRI examinations: standard fully sampled T1- and T2-weighted TSE acquisitions (reference standard) and prospectively undersampled TSEDL acquisitions with threefold and fourfold acceleration. Image evaluation was performed by five readers. Interchangeability analysis and an image quality-based analysis were used to compare the TSE and TSEDL images. Acquisition time and diagnostic confidence were also compared. Interchangeability was tested using the individual equivalence index regarding various degenerative and nondegenerative entities, which were analyzed on each vertebra and defined as discordant clinical judgments of less than 5%. Interreader and intrareader agreement and concordance (κ and Kendall τ and W statistics) were computed and Wilcoxon and McNemar tests were used. Results Overall, 50 participants were evaluated (mean age, 46 years ± 18 [SD]; 26 men). The TSEDL method enabled up to a 70% reduction in total acquisition time (100 seconds for TSEDL vs 328 seconds for TSE, P < .001). All individual equivalence indexes were less than 4%. TSEDL acquisition was rated as having superior image noise by all readers (P < .001). No evidence of a difference was found between standard TSE and TSEDL regarding frequency of major findings, overall image quality, or diagnostic confidence. Conclusion The deep learning (DL)-reconstructed turbo spin-echo (TSE) method was found to be interchangeable with standard TSE for detecting various abnormalities of the spine at MRI. DL-reconstructed TSE acquisition provided excellent image quality, with a 70% reduction in examination time. German Clinical Trials Register no. DRKS00023278 © RSNA, 2022 Online supplemental material is available for this article. See also the editorial by Hallinan in this issue.

Almansour Haidara, Herrmann Judith, Gassenmaier Sebastian, Afat Saif, Jacoby Johann, Koerzdoerfer Gregor, Nickel Dominik, Mostapha Mahmoud, Nadar Mariappan, Othman Ahmed E

2022-Nov-01

General General

De novo analysis of bulk RNA-seq data at spatially resolved single-cell resolution.

In Nature communications ; h5-index 260.0

Uncovering the tissue molecular architecture at single-cell resolution could help better understand organisms' biological and pathological processes. However, bulk RNA-seq can only measure gene expression in cell mixtures, without revealing the transcriptional heterogeneity and spatial patterns of single cells. Herein, we introduce Bulk2Space ( https://github.com/ZJUFanLab/bulk2space ), a deep learning framework-based spatial deconvolution algorithm that can simultaneously disclose the spatial and cellular heterogeneity of bulk RNA-seq data using existing single-cell and spatial transcriptomics references. The use of bulk transcriptomics to validate Bulk2Space unveils, in particular, the spatial variance of immune cells in different tumor regions, the molecular and spatial heterogeneity of tissues during inflammation-induced tumorigenesis, and spatial patterns of novel genes in different cell types. Moreover, Bulk2Space is utilized to perform spatial deconvolution analysis on bulk transcriptome data from two different mouse brain regions derived from our in-house developed sequencing approach termed Spatial-seq. We have not only reconstructed the hierarchical structure of the mouse isocortex but also further annotated cell types that were not identified by original methods in the mouse hypothalamus.

Liao Jie, Qian Jingyang, Fang Yin, Chen Zhuo, Zhuang Xiang, Zhang Ningyu, Shao Xin, Hu Yining, Yang Penghui, Cheng Junyun, Hu Yang, Yu Lingqi, Yang Haihong, Zhang Jinlu, Lu Xiaoyan, Shao Li, Wu Dan, Gao Yue, Chen Huajun, Fan Xiaohui

2022-Oct-30

General General

G2Φnet: Relating genotype and biomechanical phenotype of tissues with deep learning.

In PLoS computational biology

Many genetic mutations adversely affect the structure and function of load-bearing soft tissues, with clinical sequelae often responsible for disability or death. Parallel advances in genetics and histomechanical characterization provide significant insight into these conditions, but there remains a pressing need to integrate such information. We present a novel genotype-to-biomechanical phenotype neural network (G2Φnet) for characterizing and classifying biomechanical properties of soft tissues, which serve as important functional readouts of tissue health or disease. We illustrate the utility of our approach by inferring the nonlinear, genotype-dependent constitutive behavior of the aorta for four mouse models involving defects or deficiencies in extracellular constituents. We show that G2Φnet can infer the biomechanical response while simultaneously ascribing the associated genotype by utilizing limited, noisy, and unstructured experimental data. More broadly, G2Φnet provides a powerful method and a paradigm shift for correlating genotype and biomechanical phenotype quantitatively, promising a better understanding of their interplay in biological tissues.

Zhang Enrui, Spronck Bart, Humphrey Jay D, Karniadakis George Em

2022-Oct-31

General General

Mutual influence between language and perception in multi-agent communication games.

In PLoS computational biology

Language interfaces with many other cognitive domains. This paper explores how interactions at these interfaces can be studied with deep learning methods, focusing on the relation between language emergence and visual perception. To model the emergence of language, a sender and a receiver agent are trained on a reference game. The agents are implemented as deep neural networks, with dedicated vision and language modules. Motivated by the mutual influence between language and perception in cognition, we apply systematic manipulations to the agents' (i) visual representations, to analyze the effects on emergent communication, and (ii) communication protocols, to analyze the effects on visual representations. Our analyses show that perceptual biases shape semantic categorization and communicative content. Conversely, if the communication protocol partitions object space along certain attributes, agents learn to represent visual information about these attributes more accurately, and the representations of communication partners align. Finally, an evolutionary analysis suggests that visual representations may be shaped in part to facilitate the communication of environmentally relevant distinctions. Aside from accounting for co-adaptation effects between language and perception, our results point out ways to modulate and improve visual representation learning and emergent communication in artificial agents.

Ohmer Xenia, Marino Michael, Franke Michael, König Peter

2022-Oct-31

Surgery Surgery

Assessing the carbon footprint of digital health interventions: a scoping review.

In Journal of the American Medical Informatics Association : JAMIA

OBJECTIVE : Integration of environmentally sustainable digital health interventions requires robust evaluation of their carbon emission life-cycle before implementation in healthcare. This scoping review surveys the evidence on available environmental assessment frameworks, methods, and tools to evaluate the carbon footprint of digital health interventions for environmentally sustainable healthcare.

MATERIALS AND METHODS : Medline (Ovid), Embase (Ovid). PsycINFO (Ovid), CINAHL, Web of Science, Scopus (which indexes IEEE Xplore, Springer Lecture Notes in Computer Science and ACM databases), Compendex, and Inspec databases were searched with no time or language constraints. The Systematic Reviews and Meta-analyses Extension for Scoping Reviews (PRISMA_SCR), Joanna Briggs Scoping Review Framework, and template for intervention description and replication (TiDiER) checklist were used to structure and report the findings.

RESULTS : From 3299 studies screened, data was extracted from 13 full-text studies. No standardised methods or validated tools were identified to systematically determine the environmental sustainability of a digital health intervention over its full life-cycle from conception to realisation. Most studies (n = 8) adapted publicly available carbon calculators to estimate telehealth travel-related emissions. Others adapted these tools to examine the environmental impact of electronic health records (n = 2), e-prescriptions and e-referrals (n = 1), and robotic surgery (n = 1). One study explored optimising the information system electricity consumption of telemedicine. No validated systems-based approach to evaluation and validation of digital health interventions could be identified.

CONCLUSION : There is a need to develop standardised, validated methods and tools for healthcare environments to assist stakeholders to make informed decisions about reduction of carbon emissions from digital health interventions.

Lokmic-Tomkins Zerina, Davies Shauna, Block Lorraine J, Cochrane Lindy, Dorin Alan, von Gerich Hanna, Lozada-Perezmitre Erika, Reid Lisa, Peltonen Laura-Maria

2022-Oct-31

assessment methods and tools, carbon emissions, digital health technologies, environmental sustainability

Radiology Radiology

Deep-learning-based hepatic fat assessment (DeHFt) on non-contrast chest CT and its association with disease severity in COVID-19 infections: A multi-site retrospective study.

In EBioMedicine

BACKGROUND : Hepatic steatosis (HS) identified on CT may provide an integrated cardiometabolic and COVID-19 risk assessment. This study presents a deep-learning-based hepatic fat assessment (DeHFt) pipeline for (a) more standardised measurements and (b) investigating the association between HS (liver-to-spleen attenuation ratio <1 in CT) and COVID-19 infections severity, wherein severity is defined as requiring invasive mechanical ventilation, extracorporeal membrane oxygenation, death.

METHODS : DeHFt comprises two steps. First, a deep-learning-based segmentation model (3D residual-UNet) is trained (N = 80) to segment the liver and spleen. Second, CT attenuation is estimated using slice-based and volumetric-based methods. DeHFt-based mean liver and liver-to-spleen attenuation are compared with an expert's ROI-based measurements. We further obtained the liver-to-spleen attenuation ratio in a large multi-site cohort of patients with COVID-19 infections (D1, N = 805; D2, N = 1917; D3, N = 169) using the DeHFt pipeline and investigated the association between HS and COVID-19 infections severity.

FINDINGS : The DeHFt pipeline achieved a dice coefficient of 0.95, 95% CI [0.93-0.96] on the independent validation cohort (N = 49). The automated slice-based and volumetric-based liver and liver-to-spleen attenuation estimations strongly correlated with expert's measurement. In the COVID-19 cohorts, severe infections had a higher proportion of patients with HS than non-severe infections (pooled OR = 1.50, 95% CI [1.20-1.88], P < .001).

INTERPRETATION : The DeHFt pipeline enabled accurate segmentation of liver and spleen on non-contrast CTs and automated estimation of liver and liver-to-spleen attenuation ratio. In three cohorts of patients with COVID-19 infections (N = 2891), HS was associated with disease severity. Pending validation, DeHFt provides an automated CT-based metabolic risk assessment.

FUNDING : For a full list of funding bodies, please see the Acknowledgements.

Modanwal Gourav, Al-Kindi Sadeer, Walker Jonathan, Dhamdhere Rohan, Yuan Lei, Ji Mengyao, Lu Cheng, Fu Pingfu, Rajagopalan Sanjay, Madabhushi Anant

2022-Oct-26

COVID-19, Hepatic steatosis, NAFLD

Radiology Radiology

Deep-learning-based hepatic fat assessment (DeHFt) on non-contrast chest CT and its association with disease severity in COVID-19 infections: A multi-site retrospective study.

In EBioMedicine

BACKGROUND : Hepatic steatosis (HS) identified on CT may provide an integrated cardiometabolic and COVID-19 risk assessment. This study presents a deep-learning-based hepatic fat assessment (DeHFt) pipeline for (a) more standardised measurements and (b) investigating the association between HS (liver-to-spleen attenuation ratio <1 in CT) and COVID-19 infections severity, wherein severity is defined as requiring invasive mechanical ventilation, extracorporeal membrane oxygenation, death.

METHODS : DeHFt comprises two steps. First, a deep-learning-based segmentation model (3D residual-UNet) is trained (N = 80) to segment the liver and spleen. Second, CT attenuation is estimated using slice-based and volumetric-based methods. DeHFt-based mean liver and liver-to-spleen attenuation are compared with an expert's ROI-based measurements. We further obtained the liver-to-spleen attenuation ratio in a large multi-site cohort of patients with COVID-19 infections (D1, N = 805; D2, N = 1917; D3, N = 169) using the DeHFt pipeline and investigated the association between HS and COVID-19 infections severity.

FINDINGS : The DeHFt pipeline achieved a dice coefficient of 0.95, 95% CI [0.93-0.96] on the independent validation cohort (N = 49). The automated slice-based and volumetric-based liver and liver-to-spleen attenuation estimations strongly correlated with expert's measurement. In the COVID-19 cohorts, severe infections had a higher proportion of patients with HS than non-severe infections (pooled OR = 1.50, 95% CI [1.20-1.88], P < .001).

INTERPRETATION : The DeHFt pipeline enabled accurate segmentation of liver and spleen on non-contrast CTs and automated estimation of liver and liver-to-spleen attenuation ratio. In three cohorts of patients with COVID-19 infections (N = 2891), HS was associated with disease severity. Pending validation, DeHFt provides an automated CT-based metabolic risk assessment.

FUNDING : For a full list of funding bodies, please see the Acknowledgements.

Modanwal Gourav, Al-Kindi Sadeer, Walker Jonathan, Dhamdhere Rohan, Yuan Lei, Ji Mengyao, Lu Cheng, Fu Pingfu, Rajagopalan Sanjay, Madabhushi Anant

2022-Oct-26

COVID-19, Hepatic steatosis, NAFLD

General General

Multisite Evaluation of Prediction Models for Emergency Department Crowding Before and During the COVID-19 Pandemic.

In Journal of the American Medical Informatics Association : JAMIA

OBJECTIVE : To develop a machine learning framework to forecast emergency department (ED) crowding and to evaluate model performance under spatial and temporal data drift.

MATERIALS AND METHODS : We obtained four datasets, identified by the location: 1-large academic hospital and 2-rural hospital, and time period: pre-COVID (Jan 1, 2019-Feb 1, 2020) and COVID-era (May 15, 2020-Feb 1, 2021). Our primary target was a binary outcome that is equal to 1 if the number of patients with acute respiratory illness that were ED boarding for more than four hours was above a prescribed historical percentile. We trained a random forest and used the area under the curve (AUC) to evaluate out-of-sample performance for two experiments: 1) we evaluated the impact of sudden temporal drift by training models using pre-COVID data and testing them during the COVID-era, 2) we evaluated the impact of spatial drift by testing models trained at Location 1 on data from Location 2, and vice versa.

RESULTS : The baseline AUC values for ED boarding ranged from 0.54 (pre-COVID at Location 2) to 0.81 (COVID-era at Location 1). Models trained with pre-COVID data performed similarly to COVID-era models (0.82 vs. 0.78 at Location 1). Models that were transferred from Location 2 to Location 1 performed worse than models trained at Location 1 (0.51 vs. 0.78).

DISCUSSION AND CONCLUSION : Our results demonstrate that ED boarding is a predictable metric for ED crowding, models were not significantly impacted by temporal data drift, and any attempts at implementation must consider spatial data drift.

Smith Ari J, Patterson Brian W, Pulia Michael S, Mayer John, Schwei Rebecca J, Nagarajan Radha, Liao Frank, Shah Manish N, Boutilier Justin J

2022-Oct-29

COVID-19, Emergency medicine, data drift, emergency department boarding, machine learning

General General

Multisite Evaluation of Prediction Models for Emergency Department Crowding Before and During the COVID-19 Pandemic.

In Journal of the American Medical Informatics Association : JAMIA

OBJECTIVE : To develop a machine learning framework to forecast emergency department (ED) crowding and to evaluate model performance under spatial and temporal data drift.

MATERIALS AND METHODS : We obtained four datasets, identified by the location: 1-large academic hospital and 2-rural hospital, and time period: pre-COVID (Jan 1, 2019-Feb 1, 2020) and COVID-era (May 15, 2020-Feb 1, 2021). Our primary target was a binary outcome that is equal to 1 if the number of patients with acute respiratory illness that were ED boarding for more than four hours was above a prescribed historical percentile. We trained a random forest and used the area under the curve (AUC) to evaluate out-of-sample performance for two experiments: 1) we evaluated the impact of sudden temporal drift by training models using pre-COVID data and testing them during the COVID-era, 2) we evaluated the impact of spatial drift by testing models trained at Location 1 on data from Location 2, and vice versa.

RESULTS : The baseline AUC values for ED boarding ranged from 0.54 (pre-COVID at Location 2) to 0.81 (COVID-era at Location 1). Models trained with pre-COVID data performed similarly to COVID-era models (0.82 vs. 0.78 at Location 1). Models that were transferred from Location 2 to Location 1 performed worse than models trained at Location 1 (0.51 vs. 0.78).

DISCUSSION AND CONCLUSION : Our results demonstrate that ED boarding is a predictable metric for ED crowding, models were not significantly impacted by temporal data drift, and any attempts at implementation must consider spatial data drift.

Smith Ari J, Patterson Brian W, Pulia Michael S, Mayer John, Schwei Rebecca J, Nagarajan Radha, Liao Frank, Shah Manish N, Boutilier Justin J

2022-Oct-29

COVID-19, Emergency medicine, data drift, emergency department boarding, machine learning

oncology Oncology

Influences of rare copy-number variation on human complex traits.

In Cell ; h5-index 250.0

The human genome contains hundreds of thousands of regions harboring copy-number variants (CNV). However, the phenotypic effects of most such polymorphisms are unknown because only larger CNVs have been ascertainable from SNP-array data generated by large biobanks. We developed a computational approach leveraging haplotype sharing in biobank cohorts to more sensitively detect CNVs. Applied to UK Biobank, this approach accounted for approximately half of all rare gene inactivation events produced by genomic structural variation. This CNV call set enabled a detailed analysis of associations between CNVs and 56 quantitative traits, identifying 269 independent associations (p < 5 × 10-8) likely to be causally driven by CNVs. Putative target genes were identifiable for nearly half of the loci, enabling insights into dosage sensitivity of these genes and uncovering several gene-trait relationships. These results demonstrate the ability of haplotype-informed analysis to provide insights into the genetic basis of human complex traits.

Hujoel Margaux L A, Sherman Maxwell A, Barton Alison R, Mukamel Ronen E, Sankaran Vijay G, Terao Chikashi, Loh Po-Ru

2022-Oct-27

Complex traits, Copy-number variants, Genetic associations, Haplotypes, Structural variation

General General

New onset delirium prediction using machine learning and long short-term memory (LSTM) in electronic health record.

In Journal of the American Medical Informatics Association : JAMIA

OBJECTIVE : To develop and test an accurate deep learning model for predicting new onset delirium in hospitalized adult patients.

METHODS : Using electronic health record (EHR) data extracted from a large academic medical center, we developed a model combining long short-term memory (LSTM) and machine learning to predict new onset delirium and compared its performance with machine-learning-only models (logistic regression, random forest, support vector machine, neural network, and LightGBM). The labels of models were confusion assessment method (CAM) assessments. We evaluated models on a hold-out dataset. We calculated Shapley additive explanations (SHAP) measures to gauge the feature impact on the model.

RESULTS : A total of 331 489 CAM assessments with 896 features from 34 035 patients were included. The LightGBM model achieved the best performance (AUC 0.927 [0.924, 0.929] and F1 0.626 [0.618, 0.634]) among the machine learning models. When combined with the LSTM model, the final model's performance improved significantly (P = .001) with AUC 0.952 [0.950, 0.955] and F1 0.759 [0.755, 0.765]. The precision value of the combined model improved from 0.497 to 0.751 with a fixed recall of 0.8. Using the mean absolute SHAP values, we identified the top 20 features, including age, heart rate, Richmond Agitation-Sedation Scale score, Morse fall risk score, pulse, respiratory rate, and level of care.

CONCLUSION : Leveraging LSTM to capture temporal trends and combining it with the LightGBM model can significantly improve the prediction of new onset delirium, providing an algorithmic basis for the subsequent development of clinical decision support tools for proactive delirium interventions.

Liu Siru, Schlesinger Joseph J, McCoy Allison B, Reese Thomas J, Steitz Bryan, Russo Elise, Koh Brian, Wright Adam

2022-Oct-27

deep learning, delirium, explainable machine learning, predictive models

Radiology Radiology

An analysis of the effects of limited training data in distributed learning scenarios for brain age prediction.

In Journal of the American Medical Informatics Association : JAMIA

OBJECTIVE : Distributed learning avoids problems associated with central data collection by training models locally at each site. This can be achieved by federated learning (FL) aggregating multiple models that were trained in parallel or training a single model visiting sites sequentially, the traveling model (TM). While both approaches have been applied to medical imaging tasks, their performance in limited local data scenarios remains unknown. In this study, we specifically analyze FL and TM performances when very small sample sizes are available per site.

MATERIALS AND METHODS : 2025 T1-weighted magnetic resonance imaging scans were used to investigate the effect of sample sizes on FL and TM for brain age prediction. We evaluated models across 18 scenarios varying the number of samples per site (1, 2, 5, 10, and 20) and the number of training rounds (20, 40, and 200).

RESULTS : Our results demonstrate that the TM outperforms FL, for every sample size examined. In the extreme case when each site provided only one sample, FL achieved a mean absolute error (MAE) of 18.9 ± 0.13 years, while the TM achieved a MAE of 6.21 ± 0.50 years, comparable to central learning (MAE = 5.99 years).

DISCUSSION : Although FL is more commonly used, our study demonstrates that TM is the best implementation for small sample sizes.

CONCLUSION : The TM offers new opportunities to apply machine learning models in rare diseases and pediatric research but also allows even small hospitals to contribute small datasets.

Souza Raissa, Mouches Pauline, Wilms Matthias, Tuladhar Anup, Langner Sönke, Forkert Nils D

2022-Oct-26

brain age prediction, distributed learning, machine learning

Surgery Surgery

A time-aware attention model for prediction of acute kidney injury after pediatric cardiac surgery.

In Journal of the American Medical Informatics Association : JAMIA

OBJECTIVE : Acute kidney injury (AKI) is a common complication after pediatric cardiac surgery, and the early detection of AKI may allow for timely preventive or therapeutic measures. However, current AKI prediction researches pay less attention to time information among time-series clinical data and model building strategies that meet complex clinical application scenario. This study aims to develop and validate a model for predicting postoperative AKI that operates sequentially over individual time-series clinical data.

MATERIALS AND METHODS : A retrospective cohort of 3386 pediatric patients extracted from PIC database was used for training, calibrating, and testing purposes. A time-aware deep learning model was developed and evaluated from 3 clinical perspectives that use different data collection windows and prediction windows to answer different AKI prediction questions encountered in clinical practice. We compared our model with existing state-of-the-art models from 3 clinical perspectives using the area under the receiver operating characteristic curve (ROC AUC) and the area under the precision-recall curve (PR AUC).

RESULTS : Our proposed model significantly outperformed the existing state-of-the-art models with an improved average performance for any AKI prediction from the 3 evaluation perspectives. This model predicted 91% of all AKI episodes using data collected at 24 h after surgery, resulting in a ROC AUC of 0.908 and a PR AUC of 0.898. On average, our model predicted 83% of all AKI episodes that occurred within the different time windows in the 3 evaluation perspectives. The calibration performance of the proposed model was substantially higher than the existing state-of-the-art models.

CONCLUSIONS : This study showed that a deep learning model can accurately predict postoperative AKI using perioperative time-series data. It has the potential to be integrated into real-time clinical decision support systems to support postoperative care planning.

Zeng Xian, Shi Shanshan, Sun Yuhan, Feng Yuqing, Tan Linhua, Lin Ru, Li Jianhua, Duan Huilong, Shu Qiang, Li Haomin

2022-Oct-26

acute kidney injury, deep machine learning, multi-perspectives evaluation, pediatric cardiac surgery, prediction model

Surgery Surgery

Comparison of Machine Learning Models Including Preoperative, Intraoperative, and Postoperative Data and Mortality After Cardiac Surgery.

In JAMA network open

Importance : A variety of perioperative risk factors are associated with postoperative mortality risk. However, the relative contribution of routinely collected intraoperative clinical parameters to short-term and long-term mortality remains understudied.

Objective : To examine the performance of multiple machine learning models with data from different perioperative periods to predict 30-day, 1-year, and 5-year mortality and investigate factors that contribute to these predictions.

Design, Setting, and Participants : In this prognostic study using prospectively collected data, risk prediction models were developed for short-term and long-term mortality after cardiac surgery. Included participants were adult patients undergoing a first-time valve operation, coronary artery bypass grafting, or a combination of both between 1997 and 2017 in a single center, the University Medical Centre Groningen in the Netherlands. Mortality data were obtained in November 2017. Data analysis took place between February 2020 and August 2021.

Exposure : Cardiac surgery.

Main Outcomes and Measures : Postoperative mortality rates at 30 days, 1 year, and 5 years were the primary outcomes. The area under the receiver operating characteristic curve (AUROC) was used to assess discrimination. The contribution of all preoperative, intraoperative hemodynamic and temperature, and postoperative factors to mortality was investigated using Shapley additive explanations (SHAP) values.

Results : Data from 9415 patients who underwent cardiac surgery (median [IQR] age, 68 [60-74] years; 2554 [27.1%] women) were included. Overall mortality rates at 30 days, 1 year, and 5 years were 268 patients (2.8%), 420 patients (4.5%), and 612 patients (6.5%), respectively. Models including preoperative, intraoperative, and postoperative data achieved AUROC values of 0.82 (95% CI, 0.78-0.86), 0.81 (95% CI, 0.77-0.85), and 0.80 (95% CI, 0.75-0.84) for 30-day, 1-year, and 5-year mortality, respectively. Models including only postoperative data performed similarly (30 days: 0.78 [95% CI, 0.73-0.82]; 1 year: 0.79 [95% CI, 0.74-0.83]; 5 years: 0.77 [95% CI, 0.73-0.82]). However, models based on all perioperative data provided less clinically usable predictions, with lower detection rates; for example, postoperative models identified a high-risk group with a 2.8-fold increase in risk for 5-year mortality (4.1 [95% CI, 3.3-5.1]) vs an increase of 11.3 (95% CI, 6.8-18.7) for the high-risk group identified by the full perioperative model. Postoperative markers associated with metabolic dysfunction and decreased kidney function were the main factors contributing to mortality risk.

Conclusions and Relevance : This study found that the addition of continuous intraoperative hemodynamic and temperature data to postoperative data was not associated with improved machine learning-based identification of patients at increased risk of short-term and long-term mortality after cardiac operations.

Castela Forte José, Yeshmagambetova Galiya, van der Grinten Maureen L, Scheeren Thomas W L, Nijsten Maarten W N, Mariani Massimo A, Henning Robert H, Epema Anne H

2022-Oct-03

Radiology Radiology

Deep Learning Segmentation and Reconstruction for CT of Chronic Total Coronary Occlusion.

In Radiology ; h5-index 91.0

Background CT imaging of chronic total occlusion (CTO) is useful in guiding revascularization, but manual reconstruction and quantification are time consuming. Purpose To develop and validate a deep learning (DL) model for automated CTO reconstruction. Materials and Methods In this retrospective study, a DL model for automated CTO segmentation and reconstruction was developed using coronary CT angiography images from a training set of 6066 patients (582 with CTO, 5484 without CTO) and a validation set of 1962 patients (208 with CTO, 1754 without CTO). The algorithm was validated using an external test set of 211 patients with CTO. The consistency and measurement agreement of CTO quantification were compared between the DL model and the conventional manual protocol using the intraclass correlation coefficient, Cohen κ coefficient, and Bland-Altman plot. The predictive values of CT-derived Multicenter CTO Registry of Japan (J-CTO) score for revascularization success were evaluated. Results In the external test set, 211 patients (mean age, 66 years ± 11 [SD]; 164 men) with 240 CTO lesions were evaluated. Automated segmentation and reconstruction of CTOs by DL was successful in 95% of lesions (228 of 240) without manual editing and in 48% of lesions (116 of 240) with the conventional manual protocol (P < .001). The total postprocessing and measurement time was shorter for DL than for manual reconstruction (mean, 121 seconds ± 20 vs 456 seconds ± 68; P < .001). The quantitative and qualitative CTO parameters evaluated with the two methods showed excellent correlation (all correlation coefficients > 0.85, all P < .001) and minimal measurement difference. The predictive values of J-CTO score derived from DL and conventional manual quantification for procedure success showed no difference (area under the receiver operating characteristic curve, 0.76 [95% CI: 0.69, 0.82] and 0.76 [95% CI: 0.69, 0.82], respectively; P = .55). Conclusion When compared with manual reconstruction, the deep learning model considerably reduced postprocessing time for chronic total occlusion quantification and had excellent correlation and agreement in the anatomic assessment of occlusion features. © RSNA, 2022 Online supplemental material is available for this article. See also the editorial by Loewe in this issue.

Li Meiling, Ling Runjianya, Yu Lihua, Yang Wenyi, Chen Zirong, Wu Dijia, Zhang Jiayin

2022-Oct-25

Radiology Radiology

Deep Learning for Estimating Lung Capacity on Chest Radiographs Predicts Survival in Idiopathic Pulmonary Fibrosis.

In Radiology ; h5-index 91.0

Background Total lung capacity (TLC) has been estimated with use of chest radiographs based on time-consuming methods, such as planimetric techniques and manual measurements. Purpose To develop a deep learning-based, multidimensional model capable of estimating TLC from chest radiographs and demographic variables and validate its technical performance and clinical utility with use of multicenter retrospective data sets. Materials and Methods A deep learning model was pretrained with use of 50 000 consecutive chest CT scans performed between January 2015 and June 2017. The model was fine-tuned on 3523 pairs of posteroanterior chest radiographs and plethysmographic TLC measurements from consecutive patients who underwent pulmonary function testing on the same day. The model was tested with multicenter retrospective data sets from two tertiary care centers and one community hospital, including (a) an external test set 1 (n = 207) and external test set 2 (n = 216) for technical performance and (b) patients with idiopathic pulmonary fibrosis (n = 217) for clinical utility. Technical performance was evaluated with use of various agreement measures, and clinical utility was assessed in terms of the prognostic value for overall survival with use of multivariable Cox regression. Results The mean absolute difference and within-subject SD between observed and estimated TLC were 0.69 L and 0.73 L, respectively, in the external test set 1 (161 men; median age, 70 years [IQR: 61-76 years]) and 0.52 L and 0.53 L in the external test set 2 (113 men; median age, 63 years [IQR: 51-70 years]). In patients with idiopathic pulmonary fibrosis (145 men; median age, 67 years [IQR: 61-73 years]), greater estimated TLC percentage was associated with lower mortality risk (adjusted hazard ratio, 0.97 per percent; 95% CI: 0.95, 0.98; P < .001). Conclusion A fully automatic, deep learning-based model estimated total lung capacity from chest radiographs, and the model predicted survival in idiopathic pulmonary fibrosis. © RSNA, 2022 Online supplemental material is available for this article. See also the editorial by Sorkness in this issue.

Kim Hyungjin, Jin Kwang Nam, Yoo Seung-Jin, Lee Chang Hoon, Lee Sang-Min, Hong Hyunsook, Witanto Joseph Nathanael, Yoon Soon Ho

2022-Oct-25

Radiology Radiology

Molecular Characterization and Therapeutic Approaches to Small Cell Lung Cancer: Imaging Implications.

In Radiology ; h5-index 91.0

Small cell lung cancer (SCLC) is a highly aggressive malignancy with exceptionally poor prognosis, comprising approximately 15% of lung cancers. Emerging knowledge of the molecular and genomic landscape of SCLC and recent successful clinical applications of new systemic agents have allowed for precision oncology treatment approaches. Imaging is essential for the diagnosis, staging, and treatment monitoring of patients with SCLC. The role of imaging is increasing with the approval of new treatment agents, including immune checkpoint inhibitors, which lead to novel imaging manifestations of response and toxicities. The purpose of this state-of-the-art review is to provide the reader with the latest information about SCLC, focusing on the subtyping of this malignancy (molecular characterization) and the emerging systemic therapeutic approaches and their implications for imaging. The review will also discuss the future directions of SCLC imaging, radiomics and machine learning.

Park Hyesun, Tseng Shu-Chi, Sholl Lynette M, Hatabu Hiroto, Awad Mark M, Nishino Mizuki

2022-Oct-25

Radiology Radiology

Virtual Biopsy by Using Artificial Intelligence-based Multimodal Modeling of Binational Mammography Data.

In Radiology ; h5-index 91.0

Background Computational models based on artificial intelligence (AI) are increasingly used to diagnose malignant breast lesions. However, assessment from radiologic images of the specific pathologic lesion subtypes, as detailed in the results of biopsy procedures, remains a challenge. Purpose To develop an AI-based model to identify breast lesion subtypes with mammograms and linked electronic health records labeled with histopathologic information. Materials and Methods In this retrospective study, 26 569 images were collected in 9234 women who underwent digital mammography to pretrain the algorithms. The training data included individuals who had at least 1 year of clinical and imaging history followed by biopsy-based histopathologic diagnosis from March 2013 to November 2018. A model that combined convolutional neural networks with supervised learning algorithms was independently trained to make breast lesion predictions with data from 2120 women in Israel and 1642 women in the United States. Results were reported using the area under the receiver operating characteristic curve (AUC) with the 95% DeLong approach to estimate CIs. Significance was tested with bootstrapping. Results The Israeli model was validated in 456 women and tested in 441 women (mean age, 51 years ± 11 [SD]). The U.S. model was validated in 350 women and tested in 344 women (mean age, 60 years ± 12). For predicting malignancy in the test sets (consisting of 220 Israeli patient examinations and 126 U.S. patient examinations with ductal carcinoma in situ or invasive cancer), the algorithms obtained an AUC of 0.88 (95% CI: 0.85, 0.91) and 0.80 (95% CI: 0.74, 0.85) for Israeli and U.S. patients, respectively (P = .006). These results may not hold for other cohorts of patients, and generalizability across populations should be further investigated. Conclusion The results offer supporting evidence that artificial intelligence applied to clinical and mammographic images can identify breast lesion subtypes when the data are sufficiently large, which may help assess diagnostic workflow and reduce biopsy sampling errors. Published under a CC BY 4.0 license. Online supplemental material is available for this article.

Barros Vesna, Tlusty Tal, Barkan Ella, Hexter Efrat, Gruen David, Guindy Michal, Rosen-Zvi Michal

2022-Oct-25

Pathology Pathology

Annotation of spatially resolved single-cell data with STELLAR.

In Nature methods ; h5-index 152.0

Accurate cell-type annotation from spatially resolved single cells is crucial to understand functional spatial biology that is the basis of tissue organization. However, current computational methods for annotating spatially resolved single-cell data are typically based on techniques established for dissociated single-cell technologies and thus do not take spatial organization into account. Here we present STELLAR, a geometric deep learning method for cell-type discovery and identification in spatially resolved single-cell datasets. STELLAR automatically assigns cells to cell types present in the annotated reference dataset and discovers novel cell types and cell states. STELLAR transfers annotations across different dissection regions, different tissues and different donors, and learns cell representations that capture higher-order tissue structures. We successfully applied STELLAR to CODEX multiplexed fluorescent microscopy data and multiplexed RNA imaging datasets. Within the Human BioMolecular Atlas Program, STELLAR has annotated 2.6 million spatially resolved single cells with dramatic time savings.

Brbić Maria, Cao Kaidi, Hickey John W, Tan Yuqi, Snyder Michael P, Nolan Garry P, Leskovec Jure

2022-Oct-24

General General

Deep mendelian randomization: Investigating the causal knowledge of genomic deep learning models.

In PLoS computational biology

Multi-task deep learning (DL) models can accurately predict diverse genomic marks from sequence, but whether these models learn the causal relationships between genomic marks is unknown. Here, we describe Deep Mendelian Randomization (DeepMR), a method for estimating causal relationships between genomic marks learned by genomic DL models. By combining Mendelian randomization with in silico mutagenesis, DeepMR obtains local (locus specific) and global estimates of (an assumed) linear causal relationship between marks. In a simulation designed to test recovery of pairwise causal relations between transcription factors (TFs), DeepMR gives accurate and unbiased estimates of the 'true' global causal effect, but its coverage decays in the presence of sequence-dependent confounding. We then apply DeepMR to examine the global relationships learned by a state-of-the-art DL model, BPNet, between TFs involved in reprogramming. DeepMR's causal effect estimates validate previously hypothesized relationships between TFs and suggest new relationships for future investigation.

Malina Stephen, Cizin Daniel, Knowles David A

2022-Oct-20

General General

Ab initio predictions for 3D structure and stability of single- and double-stranded DNAs in ion solutions.

In PLoS computational biology

The three-dimensional (3D) structure and stability of DNA are essential to understand/control their biological functions and aid the development of novel materials. In this work, we present a coarse-grained (CG) model for DNA based on the RNA CG model proposed by us, to predict 3D structures and stability for both dsDNA and ssDNA from the sequence. Combined with a Monte Carlo simulated annealing algorithm and CG force fields involving the sequence-dependent base-pairing/stacking interactions and an implicit electrostatic potential, the present model successfully folds 20 dsDNAs (≤52nt) and 20 ssDNAs (≤74nt) into the corresponding native-like structures just from their sequences, with an overall mean RMSD of 3.4Å from the experimental structures. For DNAs with various lengths and sequences, the present model can make reliable predictions on stability, e.g., for 27 dsDNAs with/without bulge/internal loops and 24 ssDNAs including pseudoknot, the mean deviation of predicted melting temperatures from the corresponding experimental data is only ~2.0°C. Furthermore, the model also quantificationally predicts the effects of monovalent or divalent ions on the structure stability of ssDNAs/dsDNAs.

Mu Zi-Chun, Tan Ya-Lan, Zhang Ben-Gong, Liu Jie, Shi Ya-Zhou

2022-Oct-19

General General

Clinical and Temporal Characterization of COVID-19 Subgroups Using Patient Vector Embeddings of Electronic Health Records.

In Journal of the American Medical Informatics Association : JAMIA

OBJECTIVE : To identify and characterize clinical subgroups of hospitalized COVID-19 patients.

MATERIALS AND METHODS : Electronic health records of hospitalized COVID-19 patients at NewYork-Presbyterian/Columbia University Irving Medical Center were temporally sequenced and transformed into patient vector representations using Paragraph Vector models. K-means clustering was performed to identify subgroups.

RESULTS : A diverse cohort of 11,313 patients with COVID-19 and hospitalizations between March 2, 2020 and December 1, 2021 were identified; median [IQR] age: 61.2 [40.3-74.3]; 51.5% female. Twenty subgroups of hospitalized COVID-19 patients, labeled by increasing severity, were characterized by their demographics, conditions, outcomes, and severity (mild-moderate/severe/critical). Subgroup temporal patterns were characterized by the durations in each subgroup, transitions between subgroups, and the complete paths throughout the course of hospitalization.

DISCUSSION : Several subgroups had mild-moderate SARS-CoV-2 infections but were hospitalized for underlying conditions (pregnancy, cardiovascular disease (CVD), etc.). Subgroup 7 included solid organ transplant recipients who mostly developed mild-moderate or severe disease. Subgroup 9 had a history of type-2 diabetes, kidney and CVD, and suffered the highest rates of heart failure (45.2%) and end-stage renal disease (80.6%). Subgroup 13 was the oldest (median: 82.7 years) and had mixed severity but high mortality (33.3%). Subgroup 17 had critical disease and the highest mortality (64.6%), with age (median: 68.1 years) being the only notable risk factor. Subgroups 18-20 had critical disease with high complication rates and long hospitalizations (median: 40+ days). All subgroups are detailed in the full text. A chord diagram depicts the most common transitions, and paths with the highest prevalence, longest hospitalizations, lowest and highest mortalities are presented. Understanding these subgroups and their pathways may aid clinicians in their decisions for better management and earlier intervention for patients.

Ta Casey N, Zucker Jason E, Chiu Po-Hsiang, Fang Yilu, Natarajan Karthik, Weng Chunhua

2022-Oct-18

COVID-19, Cluster analysis, SARS-CoV-2, Unsupervised machine learning

General General

Clinical and Temporal Characterization of COVID-19 Subgroups Using Patient Vector Embeddings of Electronic Health Records.

In Journal of the American Medical Informatics Association : JAMIA

OBJECTIVE : To identify and characterize clinical subgroups of hospitalized COVID-19 patients.

MATERIALS AND METHODS : Electronic health records of hospitalized COVID-19 patients at NewYork-Presbyterian/Columbia University Irving Medical Center were temporally sequenced and transformed into patient vector representations using Paragraph Vector models. K-means clustering was performed to identify subgroups.

RESULTS : A diverse cohort of 11,313 patients with COVID-19 and hospitalizations between March 2, 2020 and December 1, 2021 were identified; median [IQR] age: 61.2 [40.3-74.3]; 51.5% female. Twenty subgroups of hospitalized COVID-19 patients, labeled by increasing severity, were characterized by their demographics, conditions, outcomes, and severity (mild-moderate/severe/critical). Subgroup temporal patterns were characterized by the durations in each subgroup, transitions between subgroups, and the complete paths throughout the course of hospitalization.

DISCUSSION : Several subgroups had mild-moderate SARS-CoV-2 infections but were hospitalized for underlying conditions (pregnancy, cardiovascular disease (CVD), etc.). Subgroup 7 included solid organ transplant recipients who mostly developed mild-moderate or severe disease. Subgroup 9 had a history of type-2 diabetes, kidney and CVD, and suffered the highest rates of heart failure (45.2%) and end-stage renal disease (80.6%). Subgroup 13 was the oldest (median: 82.7 years) and had mixed severity but high mortality (33.3%). Subgroup 17 had critical disease and the highest mortality (64.6%), with age (median: 68.1 years) being the only notable risk factor. Subgroups 18-20 had critical disease with high complication rates and long hospitalizations (median: 40+ days). All subgroups are detailed in the full text. A chord diagram depicts the most common transitions, and paths with the highest prevalence, longest hospitalizations, lowest and highest mortalities are presented. Understanding these subgroups and their pathways may aid clinicians in their decisions for better management and earlier intervention for patients.

Ta Casey N, Zucker Jason E, Chiu Po-Hsiang, Fang Yilu, Natarajan Karthik, Weng Chunhua

2022-Oct-18

COVID-19, Cluster analysis, SARS-CoV-2, Unsupervised machine learning

Radiology Radiology

How does the artificial intelligence-based image-assisted technique help physicians in diagnosis of pulmonary adenocarcinoma? A randomized controlled experiment of multicenter physicians in China.

In Journal of the American Medical Informatics Association : JAMIA

OBJECTIVE : Although artificial intelligence (AI) has achieved high levels of accuracy in the diagnosis of various diseases, its impact on physicians' decision-making performance in clinical practice is uncertain. This study aims to assess the impact of AI on the diagnostic performance of physicians with differing levels of self-efficacy under working conditions involving different time pressures.

MATERIALS AND METHODS : A 2 (independent diagnosis vs AI-assisted diagnosis) × 2 (no time pressure vs 2-minute time limit) randomized controlled experiment of multicenter physicians was conducted. Participants diagnosed 10 pulmonary adenocarcinoma cases and their diagnostic accuracy, sensitivity, and specificity were evaluated. Data analysis was performed using multilevel logistic regression.

RESULTS : One hundred and four radiologists from 102 hospitals completed the experiment. The results reveal (1) AI greatly increases physicians' diagnostic accuracy, either with or without time pressure; (2) when no time pressure, AI significantly improves physicians' diagnostic sensitivity but no significant change in specificity, while under time pressure, physicians' diagnostic sensitivity and specificity are both improved with the aid of AI; (3) when no time pressure, physicians with low self-efficacy benefit from AI assistance thus improving diagnostic accuracy but those with high self-efficacy do not, whereas physicians with low and high levels of self-efficacy both benefit from AI under time pressure.

DISCUSSION : This study is one of the first to provide real-world evidence regarding the impact of AI on physicians' decision-making performance, taking into account 2 boundary factors: clinical time pressure and physicians' self-efficacy.

CONCLUSION : AI-assisted diagnosis should be prioritized for physicians working under time pressure or with low self-efficacy.

Li Jiaoyang, Zhou Lingxiao, Zhan Yi, Xu Haifeng, Zhang Cheng, Shan Fei, Liu Lei

2022-Oct-13

artificial intelligence, diagnostic performance, multicenter physicians, self-efficacy, time pressure

General General

Calibrating spatiotemporal models of microbial communities to microscopy data: A review.

In PLoS computational biology

Spatiotemporal models that account for heterogeneity within microbial communities rely on single-cell data for calibration and validation. Such data, commonly collected via microscopy and flow cytometry, have been made more accessible by recent advances in microfluidics platforms and data processing pipelines. However, validating models against such data poses significant challenges. Validation practices vary widely between modelling studies; systematic and rigorous methods have not been widely adopted. Similar challenges are faced by the (macrobial) ecology community, in which systematic calibration approaches are often employed to improve quantitative predictions from computational models. Here, we review single-cell observation techniques that are being applied to study microbial communities and the calibration strategies that are being employed for accompanying spatiotemporal models. To facilitate future calibration efforts, we have compiled a list of summary statistics relevant for quantifying spatiotemporal patterns in microbial communities. Finally, we highlight some recently developed techniques that hold promise for improved model calibration, including algorithmic guidance of summary statistic selection and machine learning approaches for efficient model simulation.

Yip Aaron, Smith-Roberge Julien, Khorasani Sara Haghayegh, Aucoin Marc G, Ingalls Brian P

2022-Oct

Dermatology Dermatology

Translational gaps and opportunities for medical wearables in digital health.

In Science translational medicine ; h5-index 138.0

A confluence of advances in biosensor technologies, enhancements in health care delivery mechanisms, and improvements in machine learning, together with an increased awareness of remote patient monitoring, has accelerated the impact of digital health across nearly every medical discipline. Medical grade wearables-noninvasive, on-body sensors operating with clinical accuracy-will play an increasingly central role in medicine by providing continuous, cost-effective measurement and interpretation of physiological data relevant to patient status and disease trajectory, both inside and outside of established health care settings. Here, we review current digital health technologies and highlight critical gaps to clinical translation and adoption.

Xu Shuai, Kim Joohee, Walter Jessica R, Ghaffari Roozbeh, Rogers John A

2022-Oct-12

Public Health Public Health

Associations of Early-Life Exposure to Submicron Particulate Matter With Childhood Asthma and Wheeze in China.

In JAMA network open

Importance : Exposure to particulate matter (PM) has been associated with childhood asthma and wheeze. However, the specific associations between asthma and PM with an aerodynamic equivalent diameter of 1 μm or less (ie, PM1), which is a contributor to PM2.5 and potentially more toxic than PM2.5, remain unclear.

Objective : To investigate the association of early-life (prenatal and first year) exposure to size-segregated PM, including PM1, PM1-2.5, PM2.5, PM2.5-10, and PM10, with childhood asthma and wheeze.

Design, Setting, and Participants : This cross-sectional study was based on a questionnaire administered between June 2019 and June 2020 to caregivers of children aged 3 to 6 years in 7 Chinese cities (Wuhan, Changsha, Taiyuan, Nanjing, Shanghai, Chongqing, and Urumqi) as the second phase of the China, Children, Homes, Health study.

Exposures : Exposure to PM1, PM1-2.5, PM2.5, PM2.5-10, and PM10 during the prenatal period and first year of life.

Main Outcomes and Measures : The main outcomes were caregiver-reported childhood asthma and wheeze. A machine learning-based space-time model was applied to estimate early-life PM1, PM2.5, and PM10 exposure at 1 × 1-km resolution. Concentrations of PM1-2.5 and PM2.5-10 were calculated by subtracting PM1 from PM2.5 and PM2.5 from PM10, respectively. Multilevel (city and child) logistic regression models were applied to assess associations.

Results : Of 29 418 children whose caregivers completed the survey (15 320 boys [52.1%]; mean [SD] age, 4.9 [0.9] years), 2524 (8.6%) ever had wheeze and 1161 (3.9%) were diagnosed with asthma. Among all children, 18 514 (62.9%) were breastfed for more than 6 months and 787 (2.7%) had parental history of atopy. A total of 22 250 children (75.6%) had a mother with an educational level of university or above. Of the 25 422 children for whom information about cigarette smoking exposure was collected, 576 (2.3%) had a mother who was a current or former smoker during pregnancy and 7525 (29.7%) had passive household cigarette smoke exposure in early life. Early-life PM1, PM2.5, and PM10 exposure were significantly associated with increased risk of childhood asthma, with higher estimates per 10-μg/m3 increase in PM1 (OR, 1.55; 95% CI, 1.27-1.89) than in PM2.5 (OR, 1.14; 95% CI, 1.03-1.26) and PM10 (OR, 1.11; 95% CI, 1.02-1.20). No association was observed between asthma and PM1-2.5 exposure, suggesting that PM1 rather than PM1-2.5 contributed to the association between PM2.5 and childhood asthma. There were significant associations between childhood wheeze and early-life PM1 exposure (OR, 1.23; 95% CI, 1.07-1.41) and PM2.5 exposure (OR, 1.08; 95% CI, 1.01-1.16) per 10-μg/m3 increase in PM1 and PM2.5, respectively.

Conclusions and Relevance : In this cross-sectional study, higher estimates were observed for the association between PM with smaller particles, such as PM1, vs PM with larger particles and childhood asthma. The results suggest that the association between PM2.5 and childhood asthma was mainly attributable to PM1.

Wu Chuansha, Zhang Yunquan, Wei Jing, Zhao Zhuohui, Norbäck Dan, Zhang Xin, Lu Chan, Yu Wei, Wang Tingting, Zheng Xiaohong, Zhang Ling

2022-Oct-03

General General

Rapid detection and recognition of whole brain activity in a freely behaving Caenorhabditis elegans.

In PLoS computational biology

Advanced volumetric imaging methods and genetically encoded activity indicators have permitted a comprehensive characterization of whole brain activity at single neuron resolution in Caenorhabditis elegans. The constant motion and deformation of the nematode nervous system, however, impose a great challenge for consistent identification of densely packed neurons in a behaving animal. Here, we propose a cascade solution for long-term and rapid recognition of head ganglion neurons in a freely moving C. elegans. First, potential neuronal regions from a stack of fluorescence images are detected by a deep learning algorithm. Second, 2-dimensional neuronal regions are fused into 3-dimensional neuron entities. Third, by exploiting the neuronal density distribution surrounding a neuron and relative positional information between neurons, a multi-class artificial neural network transforms engineered neuronal feature vectors into digital neuronal identities. With a small number of training samples, our bottom-up approach is able to process each volume-1024 × 1024 × 18 in voxels-in less than 1 second and achieves an accuracy of 91% in neuronal detection and above 80% in neuronal tracking over a long video recording. Our work represents a step towards rapid and fully automated algorithms for decoding whole brain activity underlying naturalistic behaviors.

Wu Yuxiang, Wu Shang, Wang Xin, Lang Chengtian, Zhang Quanshi, Wen Quan, Xu Tianqi

2022-Oct-10

Radiology Radiology

Evaluation of Federated Learning Variations for COVID-19 diagnosis using Chest Radiographs from 42 US and European hospitals.

In Journal of the American Medical Informatics Association : JAMIA

OBJECTIVE : Federated learning (FL) allows multiple distributed data holders to collaboratively learn a shared model without data sharing. However, individual health system data are heterogeneous. "Personalized" FL variations have been developed to counter data heterogeneity, but few have been evaluated using real-world healthcare data. The purpose of this study is to investigate the performance of a single-site versus a 3-client federated model using a previously described COVID-19 diagnostic model. Additionally, to investigate the effect of system heterogeneity, we evaluate the performance of 4 FL variations.

MATERIALS AND METHODS : We leverage a FL healthcare collaborative including data from 5 international healthcare systems (US and Europe) encompassing 42 hospitals. We implemented a COVID-19 computer vision diagnosis system using the FedAvg algorithm implemented on Clara Train SDK 4.0. To study the effect of data heterogeneity, training data was pooled from 3 systems locally and federation was simulated. We compared a centralized/pooled model, versus FedAvg, and 3 personalized FL variations (FedProx, FedBN, FedAMP).

RESULTS : We observed comparable model performance with respect to internal validation (local model: AUROC 0.94 vs FedAvg: 0.95, p = 0.5) and improved model generalizability with the FedAvg model (p < 0.05). When investigating the effects of model heterogeneity, we observed poor performance with FedAvg on internal validation as compared to personalized FL algorithms. FedAvg did have improved generalizability compared to personalized FL algorithms. On average, FedBN had the best rank performance on internal and external validation.

CONCLUSION : FedAvg can significantly improve the generalization of the model compared to other personalization FL algorithms; however, at the cost of poor internal validity. Personalized FL may offer an opportunity to develop both internal and externally validated algorithms.

Peng Le, Luo Gaoxiang, Walker Andrew, Zaiman Zachary, Jones Emma K, Gupta Hemant, Kersten Kristopher, Burns John L, Harle Christopher A, Magoc Tanja, Shickel Benjamin, Steenburg Scott D, Loftus Tyler, Melton Genevieve B, Gichoya Judy Wawira, Sun Ju, Tignanelli Christopher J

2022-Oct-10

Artificial Intelligence, COVID-19, Computer Vision, Federated Learning

Radiology Radiology

Evaluation of Federated Learning Variations for COVID-19 diagnosis using Chest Radiographs from 42 US and European hospitals.

In Journal of the American Medical Informatics Association : JAMIA

OBJECTIVE : Federated learning (FL) allows multiple distributed data holders to collaboratively learn a shared model without data sharing. However, individual health system data are heterogeneous. "Personalized" FL variations have been developed to counter data heterogeneity, but few have been evaluated using real-world healthcare data. The purpose of this study is to investigate the performance of a single-site versus a 3-client federated model using a previously described COVID-19 diagnostic model. Additionally, to investigate the effect of system heterogeneity, we evaluate the performance of 4 FL variations.

MATERIALS AND METHODS : We leverage a FL healthcare collaborative including data from 5 international healthcare systems (US and Europe) encompassing 42 hospitals. We implemented a COVID-19 computer vision diagnosis system using the FedAvg algorithm implemented on Clara Train SDK 4.0. To study the effect of data heterogeneity, training data was pooled from 3 systems locally and federation was simulated. We compared a centralized/pooled model, versus FedAvg, and 3 personalized FL variations (FedProx, FedBN, FedAMP).

RESULTS : We observed comparable model performance with respect to internal validation (local model: AUROC 0.94 vs FedAvg: 0.95, p = 0.5) and improved model generalizability with the FedAvg model (p < 0.05). When investigating the effects of model heterogeneity, we observed poor performance with FedAvg on internal validation as compared to personalized FL algorithms. FedAvg did have improved generalizability compared to personalized FL algorithms. On average, FedBN had the best rank performance on internal and external validation.

CONCLUSION : FedAvg can significantly improve the generalization of the model compared to other personalization FL algorithms; however, at the cost of poor internal validity. Personalized FL may offer an opportunity to develop both internal and externally validated algorithms.

Peng Le, Luo Gaoxiang, Walker Andrew, Zaiman Zachary, Jones Emma K, Gupta Hemant, Kersten Kristopher, Burns John L, Harle Christopher A, Magoc Tanja, Shickel Benjamin, Steenburg Scott D, Loftus Tyler, Melton Genevieve B, Gichoya Judy Wawira, Sun Ju, Tignanelli Christopher J

2022-Oct-10

Artificial Intelligence, COVID-19, Computer Vision, Federated Learning

General General

DLoopCaller: A deep learning approach for predicting genome-wide chromatin loops by integrating accessible chromatin landscapes.

In PLoS computational biology

In recent years, major advances have been made in various chromosome conformation capture technologies to further satisfy the needs of researchers for high-quality, high-resolution contact interactions. Discriminating the loops from genome-wide contact interactions is crucial for dissecting three-dimensional(3D) genome structure and function. Here, we present a deep learning method to predict genome-wide chromatin loops, called DLoopCaller, by combining accessible chromatin landscapes and raw Hi-C contact maps. Some available orthogonal data ChIA-PET/HiChIP and Capture Hi-C were used to generate positive samples with a wider contact matrix which provides the possibility to find more potential genome-wide chromatin loops. The experimental results demonstrate that DLoopCaller effectively improves the accuracy of predicting genome-wide chromatin loops compared to the state-of-the-art method Peakachu. Moreover, compared to two of most popular loop callers, such as HiCCUPS and Fit-Hi-C, DLoopCaller identifies some unique interactions. We conclude that a combination of chromatin landscapes on the one-dimensional genome contributes to understanding the 3D genome organization, and the identified chromatin loops reveal cell-type specificity and transcription factor motif co-enrichment across different cell lines and species.

Wang Siguo, Zhang Qinhu, He Ying, Cui Zhen, Guo Zhenghao, Han Kyungsook, Huang De-Shuang

2022-Oct-07

Pathology Pathology

Prediction of Epstein-Barr Virus Status in Gastric Cancer Biopsy Specimens Using a Deep Learning Algorithm.

In JAMA network open

Importance : Epstein-Barr virus (EBV)-associated gastric cancer (EBV-GC) is 1 of 4 molecular subtypes of GC and is confirmed by an expensive molecular test, EBV-encoded small RNA in situ hybridization. EBV-GC has 2 histologic characteristics, lymphoid stroma and lace-like tumor pattern, but projecting EBV-GC at biopsy is difficult even for experienced pathologists.

Objective : To develop and validate a deep learning algorithm to predict EBV status from pathology images of GC biopsy.

Design, Setting, and Participants : This diagnostic study developed a deep learning classifier to predict EBV-GC using image patches of tissue microarray (TMA) and whole slide images (WSIs) of GC and applied it to GC biopsy specimens from GCs diagnosed at Kangbuk Samsung Hospital between 2011 and 2020. For a quantitative evaluation and EBV-GC prediction on biopsy specimens, the area of each class and the fraction in total tissue or tumor area were calculated. Data were analyzed from March 5, 2021, to February 10, 2022.

Main Outcomes and Measures : Evaluation metrics of predictive model performance were assessed on accuracy, recall, precision, F1 score, area under the receiver operating characteristic curve (AUC), and κ coefficient.

Results : This study included 137 184 image patches from 16 TMAs (708 tissue cores), 24 WSIs, and 286 biopsy images of GC. The classifier was able to classify EBV-GC image patches from TMAs and WSIs with 94.70% accuracy, 0.936 recall, 0.938 precision, 0.937 F1 score, and 0.909 κ coefficient. The classifier was used for predicting and measuring the area and fraction of EBV-GC on biopsy tissue specimens. A 10% cutoff value for the predicted fraction of EBV-GC to tissue (EBV-GC/tissue area) produced the best prediction results in EBV-GC biopsy specimens and showed the highest AUC value (0.8723; 95% CI, 0.7560-0.9501). That cutoff also obtained high sensitivity (0.895) and moderate specificity (0.745) compared with experienced pathologist sensitivity (0.842) and specificity (0.854) when using the presence of lymphoid stroma and a lace-like pattern as diagnostic criteria. On prediction maps, EBV-GCs with lace-like pattern and lymphoid stroma showed the same prediction results as EBV-GC, but cases lacking these histologic features revealed heterogeneous prediction results of EBV-GC and non-EBV-GC areas.

Conclusions and Relevance : This study showed the feasibility of EBV-GC prediction using a deep learning algorithm, even in biopsy samples. Use of such an image-based classifier before a confirmatory molecular test will reduce costs and tissue waste.

Vuong Trinh Thi Le, Song Boram, Kwak Jin T, Kim Kyungeun

2022-Oct-03

General General

Hospital trajectories and early predictors of clinical outcomes differ between SARS-CoV-2 and influenza pneumonia.

In EBioMedicine

BACKGROUND : A comparison of pneumonias due to SARS-CoV-2 and influenza, in terms of clinical course and predictors of outcomes, might inform prognosis and resource management. We aimed to compare clinical course and outcome predictors in SARS-CoV-2 and influenza pneumonia using multi-state modelling and supervised machine learning on clinical data among hospitalised patients.

METHODS : This multicenter retrospective cohort study of patients hospitalised with SARS-CoV-2 (March-December 2020) or influenza (Jan 2015-March 2020) pneumonia had the composite of hospital mortality and hospice discharge as the primary outcome. Multi-state models compared differences in oxygenation/ventilatory utilisation between pneumonias longitudinally throughout hospitalisation. Differences in predictors of outcome were modelled using supervised machine learning classifiers.

FINDINGS : Among 2,529 hospitalisations with SARS-CoV-2 and 2,256 with influenza pneumonia, the primary outcome occurred in 21% and 9%, respectively. Multi-state models differentiated oxygen requirement progression between viruses, with SARS-CoV-2 manifesting rapidly-escalating early hypoxemia. Highly contributory classifier variables for the primary outcome differed substantially between viruses.

INTERPRETATION : SARS-CoV-2 and influenza pneumonia differ in presentation, hospital course, and outcome predictors. These pathogen-specific differential responses in viral pneumonias suggest distinct management approaches should be investigated.

FUNDING : This project was supported by NIH/NCATS UL1 TR002345, NIH/NCATS KL2 TR002346 (PGL), the Doris Duke Charitable Foundation grant 2015215 (PGL), NIH/NHLBI R35 HL140026 (CSC), and a Big Ideas Award from the BJC HealthCare and Washington University School of Medicine Healthcare Innovation Lab and NIH/NIGMS R35 GM142992 (PS).

Lyons Patrick G, Bhavani Sivasubramanium V, Mody Aaloke, Bewley Alice, Dittman Katherine, Doyle Aisling, Windham Samuel L, Patel Tej M, Raju Bharat Neelam, Keller Matthew, Churpek Matthew M, Calfee Carolyn S, Michelson Andrew P, Kannampallil Thomas, Geng Elvin H, Sinha Pratik

2022-Oct-03

Hospital outcomes, Influenza, SARS-CoV-2, Statistical modelling, Viral pneumonia

General General

Hospital trajectories and early predictors of clinical outcomes differ between SARS-CoV-2 and influenza pneumonia.

In EBioMedicine

BACKGROUND : A comparison of pneumonias due to SARS-CoV-2 and influenza, in terms of clinical course and predictors of outcomes, might inform prognosis and resource management. We aimed to compare clinical course and outcome predictors in SARS-CoV-2 and influenza pneumonia using multi-state modelling and supervised machine learning on clinical data among hospitalised patients.

METHODS : This multicenter retrospective cohort study of patients hospitalised with SARS-CoV-2 (March-December 2020) or influenza (Jan 2015-March 2020) pneumonia had the composite of hospital mortality and hospice discharge as the primary outcome. Multi-state models compared differences in oxygenation/ventilatory utilisation between pneumonias longitudinally throughout hospitalisation. Differences in predictors of outcome were modelled using supervised machine learning classifiers.

FINDINGS : Among 2,529 hospitalisations with SARS-CoV-2 and 2,256 with influenza pneumonia, the primary outcome occurred in 21% and 9%, respectively. Multi-state models differentiated oxygen requirement progression between viruses, with SARS-CoV-2 manifesting rapidly-escalating early hypoxemia. Highly contributory classifier variables for the primary outcome differed substantially between viruses.

INTERPRETATION : SARS-CoV-2 and influenza pneumonia differ in presentation, hospital course, and outcome predictors. These pathogen-specific differential responses in viral pneumonias suggest distinct management approaches should be investigated.

FUNDING : This project was supported by NIH/NCATS UL1 TR002345, NIH/NCATS KL2 TR002346 (PGL), the Doris Duke Charitable Foundation grant 2015215 (PGL), NIH/NHLBI R35 HL140026 (CSC), and a Big Ideas Award from the BJC HealthCare and Washington University School of Medicine Healthcare Innovation Lab and NIH/NIGMS R35 GM142992 (PS).

Lyons Patrick G, Bhavani Sivasubramanium V, Mody Aaloke, Bewley Alice, Dittman Katherine, Doyle Aisling, Windham Samuel L, Patel Tej M, Raju Bharat Neelam, Keller Matthew, Churpek Matthew M, Calfee Carolyn S, Michelson Andrew P, Kannampallil Thomas, Geng Elvin H, Sinha Pratik

2022-Oct-03

Hospital outcomes, Influenza, SARS-CoV-2, Statistical modelling, Viral pneumonia

oncology Oncology

Comparison of Natural Language Processing of Clinical Notes With a Validated Risk-Stratification Tool to Predict Severe Maternal Morbidity.

In JAMA network open

Importance : Risk-stratification tools are routinely used in obstetrics to assist care teams in assessing and communicating risk associated with delivery. Electronic health record data and machine learning methods may offer a novel opportunity to improve and automate risk assessment.

Objective : To compare the predictive performance of natural language processing (NLP) of clinician documentation with that of a previously validated tool to identify individuals at high risk for maternal morbidity.

Design, Setting, and Participants : This retrospective diagnostic study was conducted at Brigham and Women's Hospital and Massachusetts General Hospital, Boston, Massachusetts, and included individuals admitted for delivery at the former institution from July 1, 2016, to February 29, 2020. A subset of these encounters (admissions from February to December 2018) was part of a previous prospective validation study of the Obstetric Comorbidity Index (OB-CMI), a comorbidity-weighted score to stratify risk of severe maternal morbidity (SMM).

Exposures : Natural language processing of clinician documentation and OB-CMI scores.

Main Outcomes and Measures : Natural language processing of clinician-authored admission notes was used to predict SMM in individuals delivering at the same institution but not included in the prospective OB-CMI study. The NLP model was then compared with the OB-CMI in the subset with a known OB-CMI score. Model discrimination between the 2 approaches was compared using the DeLong test. Sensitivity and positive predictive value for the identification of individuals at highest risk were prioritized as the characteristics of interest.

Results : This study included 19 794 individuals; 4034 (20.4%) were included in the original prospective validation study of the OB-CMI (testing set), and the remaining 15 760 (79.6%) composed the training set. Mean (SD) age was 32.3 (5.2) years in the testing cohort and 32.2 (5.2) years in the training cohort. A total of 115 individuals in the testing cohort (2.9%) and 468 in the training cohort (3.0%) experienced SMM. The NLP model was built from a pruned vocabulary of 2783 unique words that occurred within the 15 760 admission notes from individuals in the training set. The area under the receiver operating characteristic curve of the NLP-based model for the prediction of SMM was 0.76 (95% CI, 0.72-0.81) and was comparable with that of the OB-CMI model (0.74; 95% CI, 0.70-0.79) in the testing set (P = .53). Sensitivity (NLP, 28.7%; OB-CMI, 24.4%) and positive predictive value (NLP, 19.4%; OB-CMI, 17.6%) were comparable between the NLP and OB-CMI high-risk designations for the prediction of SMM.

Conclusions and Relevance : In this study, the NLP method and a validated risk-stratification tool had a similar ability to identify patients at high risk of SMM. Future prospective research is needed to validate the NLP approach in clinical practice and determine whether it could augment or replace tools requiring manual user input.

Clapp Mark A, Kim Ellen, James Kaitlyn E, Perlis Roy H, Kaimal Anjali J, McCoy Thomas H, Easter Sarah Rae

2022-Oct-03

Public Health Public Health

Baseline host determinants of robust human HIV-1 vaccine-induced immune responses: A meta-analysis of 26 vaccine regimens.

In EBioMedicine

BACKGROUND : The identification of baseline host determinants that associate with robust HIV-1 vaccine-induced immune responses could aid HIV-1 vaccine development. We aimed to assess both the collective and relative performance of baseline characteristics in classifying individual participants in nine different Phase 1-2 HIV-1 vaccine clinical trials (26 vaccine regimens, conducted in Africa and in the Americas) as High HIV-1 vaccine responders.

METHODS : This was a meta-analysis of individual participant data, with studies chosen based on participant-level (vs. study-level summary) data availability within the HIV-1 Vaccine Trials Network. We assessed the performance of 25 baseline characteristics (demographics, safety haematological measurements, vital signs, assay background measurements) and estimated the relative importance of each characteristic in classifying 831 participants as High (defined as within the top 25th percentile among positive responders or above the assay upper limit of quantification) versus Non-High responders. Immune response outcomes included HIV-1-specific serum IgG binding antibodies and Env-specific CD4+ T-cell responses assessed two weeks post-last dose, all measured at central HVTN laboratories. Three variable importance approaches based on SuperLearner ensemble machine learning were considered.

FINDINGS : Overall, 30.1%, 50.5%, 36.2%, and 13.9% of participants were categorized as High responders for gp120 IgG, gp140 IgG, gp41 IgG, and Env-specific CD4+ T-cell vaccine-induced responses, respectively. When including all baseline characteristics, moderate performance was achieved for the classification of High responder status for the binding antibody responses, with cross-validated areas under the ROC curve (CV-AUC) of 0.72 (95% CI: 0.68, 0.76) for gp120 IgG, 0.73 (0.69, 0.76) for gp140 IgG, and 0.67 (95% CI: 0.63, 0.72) for gp41 IgG. In contrast, the collection of all baseline characteristics yielded little improvement over chance for predicting High Env-specific CD4+ T-cell responses [CV-AUC: 0.53 (0.48, 0.58)]. While estimated variable importance patterns differed across the three approaches, female sex assigned at birth, lower height, and higher total white blood cell count emerged as significant predictors of High responder status across multiple immune response outcomes using Approach 1. Of these three baseline variables, total white blood cell count ranked highly across all three approaches for predicting vaccine-induced gp41 and gp140 High responder status.

INTERPRETATION : The identified features should be studied further in pursuit of intervention strategies to improve vaccine responses and may be adjusted for in analyses of immune response data to enhance statistical power.

FUNDING : National Institute of Allergy and Infectious Diseases (UM1AI068635 to YH, UM1AI068614 to GDT, UM1AI068618 to MJM, and UM1 AI069511 to MCK), the Duke CFAR P30 AI064518 to GDT, and National Institute of Dental and Craniofacial Research (R01DE027245 to JJK). This work was also supported by the Bill and Melinda Gates Foundation. The content is solely the responsibility of the authors and does not necessarily represent the official views of any of the funding sources.

Huang Yunda, Zhang Yuanyuan, Seaton Kelly E, De Rosa Stephen, Heptinstall Jack, Carpp Lindsay N, Randhawa April Kaur, McKinnon Lyle R, McLaren Paul, Viegas Edna, Gray Glenda E, Churchyard Gavin, Buchbinder Susan P, Edupuganti Srilatha, Bekker Linda-Gail, Keefer Michael C, Hosseinipour Mina C, Goepfert Paul A, Cohen Kristen W, Williamson Brian D, McElrath M Juliana, Tomaras Georgia D, Thakar Juilee, Kobie James J

2022-Sep-27

Antibody, Baseline characteristics, CD4+ T cell, SuperLearner, Vaccine response heterogeneity, Variable importance measurements

Public Health Public Health

Baseline host determinants of robust human HIV-1 vaccine-induced immune responses: A meta-analysis of 26 vaccine regimens.

In EBioMedicine

BACKGROUND : The identification of baseline host determinants that associate with robust HIV-1 vaccine-induced immune responses could aid HIV-1 vaccine development. We aimed to assess both the collective and relative performance of baseline characteristics in classifying individual participants in nine different Phase 1-2 HIV-1 vaccine clinical trials (26 vaccine regimens, conducted in Africa and in the Americas) as High HIV-1 vaccine responders.

METHODS : This was a meta-analysis of individual participant data, with studies chosen based on participant-level (vs. study-level summary) data availability within the HIV-1 Vaccine Trials Network. We assessed the performance of 25 baseline characteristics (demographics, safety haematological measurements, vital signs, assay background measurements) and estimated the relative importance of each characteristic in classifying 831 participants as High (defined as within the top 25th percentile among positive responders or above the assay upper limit of quantification) versus Non-High responders. Immune response outcomes included HIV-1-specific serum IgG binding antibodies and Env-specific CD4+ T-cell responses assessed two weeks post-last dose, all measured at central HVTN laboratories. Three variable importance approaches based on SuperLearner ensemble machine learning were considered.

FINDINGS : Overall, 30.1%, 50.5%, 36.2%, and 13.9% of participants were categorized as High responders for gp120 IgG, gp140 IgG, gp41 IgG, and Env-specific CD4+ T-cell vaccine-induced responses, respectively. When including all baseline characteristics, moderate performance was achieved for the classification of High responder status for the binding antibody responses, with cross-validated areas under the ROC curve (CV-AUC) of 0.72 (95% CI: 0.68, 0.76) for gp120 IgG, 0.73 (0.69, 0.76) for gp140 IgG, and 0.67 (95% CI: 0.63, 0.72) for gp41 IgG. In contrast, the collection of all baseline characteristics yielded little improvement over chance for predicting High Env-specific CD4+ T-cell responses [CV-AUC: 0.53 (0.48, 0.58)]. While estimated variable importance patterns differed across the three approaches, female sex assigned at birth, lower height, and higher total white blood cell count emerged as significant predictors of High responder status across multiple immune response outcomes using Approach 1. Of these three baseline variables, total white blood cell count ranked highly across all three approaches for predicting vaccine-induced gp41 and gp140 High responder status.

INTERPRETATION : The identified features should be studied further in pursuit of intervention strategies to improve vaccine responses and may be adjusted for in analyses of immune response data to enhance statistical power.

FUNDING : National Institute of Allergy and Infectious Diseases (UM1AI068635 to YH, UM1AI068614 to GDT, UM1AI068618 to MJM, and UM1 AI069511 to MCK), the Duke CFAR P30 AI064518 to GDT, and National Institute of Dental and Craniofacial Research (R01DE027245 to JJK). This work was also supported by the Bill and Melinda Gates Foundation. The content is solely the responsibility of the authors and does not necessarily represent the official views of any of the funding sources.

Huang Yunda, Zhang Yuanyuan, Seaton Kelly E, De Rosa Stephen, Heptinstall Jack, Carpp Lindsay N, Randhawa April Kaur, McKinnon Lyle R, McLaren Paul, Viegas Edna, Gray Glenda E, Churchyard Gavin, Buchbinder Susan P, Edupuganti Srilatha, Bekker Linda-Gail, Keefer Michael C, Hosseinipour Mina C, Goepfert Paul A, Cohen Kristen W, Williamson Brian D, McElrath M Juliana, Tomaras Georgia D, Thakar Juilee, Kobie James J

2022-Sep-27

Antibody, Baseline characteristics, CD4+ T cell, SuperLearner, Vaccine response heterogeneity, Variable importance measurements