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

Internal Medicine Internal Medicine

Evaluation of EfficientNet models for COVID-19 detection using lung parenchyma.

In Neural computing & applications

When the COVID-19 pandemic broke out in the beginning of 2020, it became crucial to enhance early diagnosis with efficient means to reduce dangers and future spread of the viruses as soon as possible. Finding effective treatments and lowering mortality rates is now more important than ever. Scanning with a computer tomography (CT) scanner is a helpful method for detecting COVID-19 in this regard. The present paper, as such, is an attempt to contribute to this process by generating an open-source, CT-based image dataset. This dataset contains the CT scans of lung parenchyma regions of 180 COVID-19-positive and 86 COVID-19-negative patients taken at the Bursa Yuksek Ihtisas Training and Research Hospital. The experimental studies show that the modified EfficientNet-ap-nish method uses this dataset effectively for diagnostic purposes. Firstly, a smart segmentation mechanism based on the k-means algorithm is applied to this dataset as a preprocessing stage. Then, performance pretrained models are analyzed using different CNN architectures and with our Nish activation function. The statistical rates are obtained by the various EfficientNet models and the highest detection score is obtained with the EfficientNet-B4-ap-nish version, which provides a 97.93% accuracy rate and a 97.33% F1-score. The implications of the proposed method are immense both for present-day applications and future developments.

Kurt Zuhal, Işık Şahin, Kaya Zeynep, Anagün Yıldıray, Koca Nizameddin, Çiçek Sümeyye

2023-Feb-20

COVID-19 detection, CT scan, Deep learning, EfficientNet, K-means, Lung parenchyma

Public Health Public Health

Can we predict critical care mortality with non-conventional inflammatory markers in SARS-CoV-2 infected patients?

In Clinical hemorheology and microcirculation

BACKGROUND : Severe COVID-19 disease is associated with multiple organ involvement,then failure and often fatal outcomes.In addition,inflammatory mechanisms and cytokine storms,documented in many COVID-19 patients,are responsible for the progression of the disease and high mortality rates.Inflammatory parameters,such as procalcitonin(PCT) and C-reactive protein(CRP), are widely used in clinical practice.

OBJECTIVE : To evaluate the predictive power of non-conventional inflammatory markers regarding mortality risk.

METHODS : In our prospective study 52 patients were followed for 5 days after admission to an intensive care unit immediately with severe SARS-CoV-2 infection.We compared leukocyte-,platelet antisedimentation rate (LAR, PAR),neutrophil lymphocyte ratio(NLR), CRP, PCT levels.

RESULTS : In non-surviving(NSU) patients LAR remained largely constant from D1 to D4 with a statistically significant drop(p <  0.05) only seen on D5.The NSU group showed statistically significant(p <  0.05) elevated LAR medians on D4 and D5, compared to the SU group.NLR values were continually higher in the non-survivor group.The difference between the SU and NSU groups were statistically significant on every examined day.PAR, CRP and PCT levels didn't show any significant differences between the SU and NSU groups.

CONCLUSIONS : In conclusion, this study suggests that LAR and NLR are especially worthy of further investigation as prognostic markers.LAR might be of particular relevance as it is not routinely obtained in current clinical practice.It would seem beneficial to include LAR in data sets to train prognostic artificial intelligence.

Rozanovic Martin, Domokos Kamilla, Márovics Gergő, Rohonczi Mirtill, Csontos Csaba, Bogár Lajos, Rendeki Szilárd, Kiss Tamás, Rozanovic Melánia Nacira, Loibl Csaba

2023-Feb-18

C-reactive protein, COVID-19, infection, inflammatory response, leukocyte antisedimentation rate, neutrophil-lymphocyte ratio, procalcitonin

oncology Oncology

Molecular and functional imaging in cancer-targeted therapy: current applications and future directions.

In Signal transduction and targeted therapy

Targeted anticancer drugs block cancer cell growth by interfering with specific signaling pathways vital to carcinogenesis and tumor growth rather than harming all rapidly dividing cells as in cytotoxic chemotherapy. The Response Evaluation Criteria in Solid Tumor (RECIST) system has been used to assess tumor response to therapy via changes in the size of target lesions as measured by calipers, conventional anatomically based imaging modalities such as computed tomography (CT), and magnetic resonance imaging (MRI), and other imaging methods. However, RECIST is sometimes inaccurate in assessing the efficacy of targeted therapy drugs because of the poor correlation between tumor size and treatment-induced tumor necrosis or shrinkage. This approach might also result in delayed identification of response when the therapy does confer a reduction in tumor size. Innovative molecular imaging techniques have rapidly gained importance in the dawning era of targeted therapy as they can visualize, characterize, and quantify biological processes at the cellular, subcellular, or even molecular level rather than at the anatomical level. This review summarizes different targeted cell signaling pathways, various molecular imaging techniques, and developed probes. Moreover, the application of molecular imaging for evaluating treatment response and related clinical outcome is also systematically outlined. In the future, more attention should be paid to promoting the clinical translation of molecular imaging in evaluating the sensitivity to targeted therapy with biocompatible probes. In particular, multimodal imaging technologies incorporating advanced artificial intelligence should be developed to comprehensively and accurately assess cancer-targeted therapy, in addition to RECIST-based methods.

Bai Jing-Wen, Qiu Si-Qi, Zhang Guo-Jun

2023-Feb-27

Public Health Public Health

A case-control study on predicting population risk of suicide using health administrative data: a research protocol.

In BMJ open

INTRODUCTION : Suicide has a complex aetiology and is a result of the interaction among the risk and protective factors at the individual, healthcare system and population levels. Therefore, policy and decision makers and mental health service planners can play an important role in suicide prevention. Although a number of suicide risk predictive tools have been developed, these tools were designed to be used by clinicians for assessing individual risk of suicide. There have been no risk predictive models to be used by policy and decision makers for predicting population risk of suicide at the national, provincial and regional levels. This paper aimed to describe the rationale and methodology for developing risk predictive models for population risk of suicide.

METHODS AND ANALYSIS : A case-control study design will be used to develop sex-specific risk predictive models for population risk of suicide, using statistical regression and machine learning techniques. Routinely collected health administrative data in Quebec, Canada, and community-level social deprivation and marginalisation data will be used. The developed models will be transformed into the models that can be readily used by policy and decision makers. Two rounds of qualitative interviews with end-users and other stakeholders were proposed to understand their views about the developed models and potential systematic, social and ethical issues for implementation; the first round of qualitative interviews has been completed. We included 9440 suicide cases (7234 males and 2206 females) and 661 780 controls for model development. Three hundred and forty-seven variables at individual, healthcare system and community levels have been identified and will be included in least absolute shrinkage and selection operator regression for feature selection.

ETHICS AND DISSEMINATION : This study is approved by the Health Research Ethnics Committee of Dalhousie University, Canada. This study takes an integrated knowledge translation approach, involving knowledge users from the beginning of the process.

Wang JianLi, Gholi Zadeh Kharrat Fatemeh, Pelletier Jean-François, Rochette Louis, Pelletier Eric, Lévesque Pascale, Massamba Victoria, Brousseau-Paradis Camille, Mohammed Mada, Gariépy Geneviève, Gagné Christian, Lesage Alain

2023-Feb-27

HEALTH SERVICES ADMINISTRATION & MANAGEMENT, PSYCHIATRY, PUBLIC HEALTH, Suicide & self-harm

General General

Complex hierarchical structures in single-cell genomics data unveiled by deep hyperbolic manifold learning.

In Genome research ; h5-index 99.0

With the advances in single-cell sequencing techniques, numerous analytical methods have been developed for delineating cell development. However, most are based on Euclidean space, which would distort the complex hierarchical structure of cell differentiation. Recently, methods acting on hyperbolic space have been proposed to visualize hierarchical structures in single-cell RNA-seq (scRNA-seq) data and have been proven to be superior to methods acting on Euclidean space. However, these methods have fundamental limitations and are not optimized for the highly sparse single-cell count data. To address these limitations, we propose scDHMap, a model-based deep learning approach to visualize the complex hierarchical structures of scRNA-seq data in low-dimensional hyperbolic space. The evaluations on extensive simulation and real experiments show that scDHMap outperforms existing dimensionality-reduction methods in various common analytical tasks as needed for scRNA-seq data, including revealing trajectory branches, batch correction, and denoising the count matrix with high dropout rates. In addition, we extend scDHMap to visualize single-cell ATAC-seq data.

Tian Tian, Zhong Cheng, Lin Xiang, Wei Zhi, Hakonarson Hakon

2023-Feb-27

Radiology Radiology

Image Quality Assessment of Deep Learning Image Reconstruction in Torso Computed Tomography Using Tube Current Modulation.

In Acta medica Okayama

Novel deep learning image reconstruction (DLIR) reportedly changes the image quality characteristics based on object contrast and image noise. In clinical practice, computed tomography image noise is usually controlled by tube current modulation (TCM) to accommodate changes in object size. This study aimed to evaluate the image quality characteristics of DLIR for different object sizes when the in-plane noise was controlled by TCM. Images acquisition was performed on a GE Revolution CT system to investigate the impact of the DLIR algorithm compared to the standard reconstructions of filtered-back projection (FBP) and hybrid iterative reconstruction (hybrid-IR). The image quality assessment was performed using phantom images, and an observer study was conducted using clinical cases. The image quality assessment confirmed the excellent noise- reduction performance of DLIR, despite variations due to phantom size. Similarly, in the observer study, DLIR received high evaluations regardless of the body parts imaged. We evaluated a novel DLIR algorithm by replicating clinical behaviors. Consequently, DLIR exhibited higher image quality than those of FBP and hybrid-IR in both phantom and observer studies, albeit the value depended on the reconstruction strength, and proved itself capable of providing stable image quality in clinical use.

Takeuchi Kazuhiro, Ide Yasuhiro, Mori Yuichiro, Uehara Yusuke, Sukeishi Hiroshi, Goto Sachiko

2023-Feb

computed tomography, deep learning, image reconstruction, object size, tube current modulation