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

Pathology Pathology

Clinically Relevant Latent Space Embedding of Cancer Histopathology Slides through Variational Autoencoder Based Image Compression

ArXiv Preprint

In this paper, we introduce a Variational Autoencoder (VAE) based training approach that can compress and decompress cancer pathology slides at a compression ratio of 1:512, which is better than the previously reported state of the art (SOTA) in the literature, while still maintaining accuracy in clinical validation tasks. The compression approach was tested on more common computer vision datasets such as CIFAR10, and we explore which image characteristics enable this compression ratio on cancer imaging data but not generic images. We generate and visualize embeddings from the compressed latent space and demonstrate how they are useful for clinical interpretation of data, and how in the future such latent embeddings can be used to accelerate search of clinical imaging data.

Mohammad Sadegh Nasr, Amir Hajighasemi, Paul Koomey, Parisa Boodaghi Malidarreh, Michael Robben, Jillur Rahman Saurav, Helen H. Shang, Manfred Huber, Jacob M. Luber

2023-03-23

General General

SARS-CoV-2 related adaptation mechanisms of rehabilitation clinics affecting patient-centred care: Qualitative study of online patient reports.

In JMIR rehabilitation and assistive technologies

BACKGROUND : The SARS-CoV-2 pandemic impacted the access to inpatient rehabilitation services. At the current state of research, it is unclear to what extent the adaptation of rehabilitation services to infection-protective standards affected patient-centred care in Germany.

OBJECTIVE : This study aimed to explore which aspects of patient-centred care were relevant for patients in inpatient rehabilitation clinics under early-phase pandemic conditions.

METHODS : A deductive-inductive framework analysis of online patient reports posted on a leading German hospital rating website was conducted (www.klinikbewertungen.de). The selected hospital rating website is a third party, patient-centred commercial platform which operates independently of governmental entities. Following a theoretical sampling approach, online reports of rehabilitation stays in two federal states of Germany (Brandenburg, Saarland) uploaded between March 2020 and September 2021 were included. Independently of medical specialty groups, all reports were included. Keywords addressing framework domains were analysed descriptively.

RESULTS : In total, 649 online reports reflecting inpatient rehabilitation services of 31 clinics (Brandenburg N = 23; Saarland N = 8) were analysed. Keywords addressing the care environment were most frequently reported (59.9%) followed by staff prerequisites (33.0%), patient-centred processes (4.5%) and expected outcomes (2.6%). Qualitative in depth-analysis revealed SARS-CoV-2 related reports to be associated with domains of patient-centred processes and staff prerequisites. Discontinuous communication of infection protection standards was perceived to threaten patient autonomy. This was amplified by a tangible gratification crisis of medical staff. Established and emotional supportive relationships to clinicians and peer-groups offered the potential to mitigate adverse effects of infection protection standards.

CONCLUSIONS : Patients predominantly reported feedback associated with the care environment. SARS-CoV-2 related reports were strongly affected by increased staff workloads as well as patient-centred processes addressing discontinuous communication and organizationally demanding implementation of infection protection standards which were perceived to threaten patient autonomy. Peer-relationships formed during inpatient rehabilitation had the potential to mitigate these mechanisms.

CLINICALTRIAL : Not applicable.

Kühn Lukas, Lindert Lara, Kuper Paulina, Choi Kyung-Eun Anna

2023-Mar-05

Public Health Public Health

Predictors of Cyberchondria during the COVID-19 pandemic: A cross-sectional study using supervised machine learning.

In JMIR formative research

BACKGROUND : Cyberchondria is characterized by repeated and compulsive online searches for health information, resulting in increased health anxiety and distress. It has been conceptualized as a multi-dimensional construct fueled by both anxiety and compulsivity-related factors and described as a "transdiagnostic compulsive behavioral syndrome" which is associated with health anxiety, problematic internet use and obsessive-compulsive symptoms. Cyberchondria is not included in the ICD-11 or the DSM-5, and its defining features, etiological mechanisms and assessment continue to be debated.

OBJECTIVE : This study aimed to investigate changes in the severity of cyberchondria during the pandemic and identify predictors of cyberchondria at this time.

METHODS : Data collection started on May 4, 2020 and ended on June 10, 2020, which corresponds to the first wave of the COVID-19 pandemic in Europe. At the time the present study took place, French-speaking countries in Europe (France, Switzerland, Belgium and Luxembourg) all implemented lockdown or semi-lockdown measures. The survey consisted of a questionnaire collecting demographic information (sex, age, education level and country of residence) and information on socioeconomic circumstances during the first lockdown (e.g., economic situation, housing and employment status), and was followed by several instruments assessing various psychological and health-related constructs. Inclusion criteria for the study were being at least 18 years of age and having a good understanding of French. Self-report data were collected from 725 participants aged 18 to 77 years (mean 33.29, SD 12.88 years), with females constituting the majority (416/725, 57.4%).

RESULTS : The results show that the COVID-19 pandemic affected various facets of cyberchondria: cyberchondria-related distress and interference with functioning increased (distress z=-3.651, P<.001; compulsion z=-5.697, P<.001), whereas the reassurance facet of cyberchondria decreased (z=-6.680, P<.001). Also, COVID-19-related fears and health anxiety emerged as the strongest predictors of cyberchondria-related distress and interference with functioning during the pandemic.

CONCLUSIONS : These findings provide evidence about the impact of the COVID-19 pandemic on cyberchondria and identify factors that should be considered in efforts to prevent and manage cyberchondria at times of public health crises. Also, they are consistent with the theoretical model of cyberchondria during the COVID-19 pandemic proposed by Starcevic and his colleagues in 2020. In addition, the findings have implications for the conceptualization and future assessment of cyberchondria.

Infanti Alexandre, Starcevic Vladan, Schimmenti Adriano, Khazaal Yasser, Karila Laurent, Giardina Alessandro, Flayelle Maèva, Hedayatzadeh Razavi Seyedeh Boshra, Baggio Stéphanie, Vögele Claus, Billieux Joël

2023-Mar-09

General General

The hair cell analysis toolbox is a precise and fully automated pipeline for whole cochlea hair cell quantification.

In PLoS biology

Our sense of hearing is mediated by sensory hair cells, precisely arranged and highly specialized cells subdivided into outer hair cells (OHCs) and inner hair cells (IHCs). Light microscopy tools allow for imaging of auditory hair cells along the full length of the cochlea, often yielding more data than feasible to manually analyze. Currently, there are no widely applicable tools for fast, unsupervised, unbiased, and comprehensive image analysis of auditory hair cells that work well either with imaging datasets containing an entire cochlea or smaller sampled regions. Here, we present a highly accurate machine learning-based hair cell analysis toolbox (HCAT) for the comprehensive analysis of whole cochleae (or smaller regions of interest) across light microscopy imaging modalities and species. The HCAT is a software that automates common image analysis tasks such as counting hair cells, classifying them by subtype (IHCs versus OHCs), determining their best frequency based on their location along the cochlea, and generating cochleograms. These automated tools remove a considerable barrier in cochlear image analysis, allowing for faster, unbiased, and more comprehensive data analysis practices. Furthermore, HCAT can serve as a template for deep learning-based detection tasks in other types of biological tissue: With some training data, HCAT's core codebase can be trained to develop a custom deep learning detection model for any object on an image.

Buswinka Christopher J, Osgood Richard T, Simikyan Rubina G, Rosenberg David B, Indzhykulian Artur A

2023-Mar-22

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

General General

Classification of truck-involved crash severity: Dealing with missing, imbalanced, and high dimensional safety data.

In PloS one ; h5-index 176.0

While the cost of road traffic fatalities in the U.S. surpasses $240 billion a year, the availability of high-resolution datasets allows meticulous investigation of the contributing factors to crash severity. In this paper, the dataset for Trucks Involved in Fatal Accidents in 2010 (TIFA 2010) is utilized to classify the truck-involved crash severity where there exist different issues including missing values, imbalanced classes, and high dimensionality. First, a decision tree-based algorithm, the Synthetic Minority Oversampling Technique (SMOTE), and the Random Forest (RF) feature importance approach are employed for missing value imputation, minority class oversampling, and dimensionality reduction, respectively. Afterward, a variety of classification algorithms, including RF, K-Nearest Neighbors (KNN), Multi-Layer Perceptron (MLP), Gradient-Boosted Decision Trees (GBDT), and Support Vector Machine (SVM) are developed to reveal the influence of the introduced data preprocessing framework on the output quality of ML classifiers. The results show that the GBDT model outperforms all the other competing algorithms for the non-preprocessed crash data based on the G-mean performance measure, but the RF makes the most accurate prediction for the treated dataset. This finding indicates that after the feature selection is conducted to alleviate the computational cost of the machine learning algorithms, bagging (bootstrap aggregating) of decision trees in RF leads to a better model rather than boosting them via GBDT. Besides, the adopted feature importance approach decreases the overall accuracy by only up to 5% in most of the estimated models. Moreover, the worst class recall value of the RF algorithm without prior oversampling is only 34.4% compared to the corresponding value of 90.3% in the up-sampled model which validates the proposed multi-step preprocessing scheme. This study also identifies the temporal and spatial (roadway) attributes, as well as crash characteristics, and Emergency Medical Service (EMS) as the most critical factors in truck crash severity.

Mohammadpour Seyed Iman, Khedmati Majid, Zada Mohammad Javad Hassan

2023