Centralized Triage of Suspected Coronary Artery Disease Using Coronary Computed Tomographic Angiography to Optimize the Diagnostic Yield of Invasive Angiography.

Background: Coronary computed tomographic angiography (CCTA) is preferable to invasive coronary angiography (ICA) for coronary artery disease (CAD) diagnosis in elective patients without known CAD. Methods: We conducted a nonrandomized interventional study involving 2 tertiary care centres in Ontario. From July 2018 to February 2020, outpatients referred for elective ICA were identified through a centralized triage process and were recommended to undergo CCTA first instead of ICA. Patients with borderline or obstructive CAD on CCTA were recommended to undergo subsequent ICA. Intervention acceptability, fidelity, and effectiveness were assessed. Results: A total of 226 patients were screened, with 186 confirmed to be eligible, of whom 166 had patient and physician approval to proceed with CCTA (89% acceptability). Among consenting patients, 156 (94%) underwent CCTA first; 43 (28%) had borderline/obstructive CAD on CCTA, and only 1 with normal/nonobstructive CAD on CCTA was referred for subsequent ICA against protocol (99% fidelity). Overall, 119 of 156 CCTA-first patients did not have ICA within the following 90 days (i.e., 76% potentially avoided ICA, due to the intervention). Among the 36 who underwent ICA post-CCTA per protocol, 24 had obstructive CAD (66.7% diagnostic yield). If all patients who were referred for and underwent ICA at either centre between July 2016 and February 2020 (n = 694 pre-implementation; n = 333 post-implementation) had had CCTA first, an additional 42 patients per 100 would have had an obstructive CAD finding on their ICA (95% confidence interval = 26-59). Conclusion: A centralized triage process, in which elective outpatients referred for ICA are instead referred for CCTA first, appears to be acceptable and effective in diagnosing obstructive CAD and improving efficiencies in our healthcare system.

Contexte: La coronarographie par tomodensitométrie (coro-TDM) est préférable à la coronarographie invasive chez les patients sans coronaropathie connue chez qui le diagnostic d'une coronaropathie n'est pas urgent. Méthodologie: Nous avons réalisé une étude interventionnelle non randomisée dans deux centres de soins tertiaires en Ontario. Les patients ambulatoires pour qui une coronarographie invasive non urgente a été demandée entre juillet 2018 et février 2020 ont été recensés par un processus centralisé de triage et se sont fait recommander de subir d'abord une coro-TDM. Les patients qui présentaient une co-ronaropathie obstructive ou dont les résultats se trouvaient tout juste à la limite de ce diagnostic lors de la coro-TDM se faisaient recommander une coronarographie invasive subséquente. L'acceptabilité de l'intervention, sa fidélité et son efficacité ont été évaluées. Résultats: Au total, 226 patients ont été sélectionnés et 186 ont été jugés admissibles. Parmi ces derniers, 166 ont accepté de subir la coro-TDM recommandée par le médecin (acceptabilité de 89 %). Parmi les patients ayant donné leur consentement, 156 (94 %) se sont d'abord soumis à une coro-TDM, et 43 (28 %) présentaient une coronaropathie obstructive ou des résultats limites selon cet examen. Seulement un patient ne présentant pas de coronaropathie ou présentant une coronaropathie non obstructive à la coro-TDM a été orienté vers une coronarographie invasive subséquente, contrairement au protocole (fidélité de 99 %). En tout, 119 des 156 patients s'étant d'abord soumis à une coro-TDM n'ont pas eu à subir une coronarographie invasive dans les 90 jours suivants (76 % d'entre eux ont potentiellement évité une coronarographie invasive grâce à cette première intervention). Parmi les 36 patients qui ont subi une coronarographie invasive après la coro-TDM, comme le recommandait le protocole, 24 présentaient une coronaropathie obstructive (rendement diagnostique de 66,7 %). Si tous les patients qui ont été orientés vers une coronarographie invasive et qui se sont soumis à cet examen dans l'un ou l'autre des centres entre juillet 2016 et février 2020 (n = 694 avant l'instauration; n = 333 après l'instauration) avaient d'abord passé une coro-TDM, une coronaropathie obstructive aurait été décelée lors de la coronarographie invasive chez 42 patients de plus par tranche de 100 patients (intervalle de confiance à 95 % : 26 à 59). Conclusion: Le recours à un processus de triage centralisé permettant de faire d'abord passer une coro-TDM aux patients ambulatoires dans une situation non urgente qui doivent subir une coronarographie invasive semble être un moyen acceptable et efficace de diagnostiquer la coronaropathie obstructive et d'améliorer l'efficacité dans notre système de santé.

COVID-Net CXR-2: An Enhanced Deep Convolutional Neural Network Design for Detection of COVID-19 Cases From Chest X-ray Images.

As the COVID-19 pandemic devastates globally, the use of chest X-ray (CXR) imaging as a complimentary screening strategy to RT-PCR testing continues to grow given its routine clinical use for respiratory complaint. As part of the COVID-Net open source initiative, we introduce COVID-Net CXR-2, an enhanced deep convolutional neural network design for COVID-19 detection from CXR images built using a greater quantity and diversity of patients than the original COVID-Net. We also introduce a new benchmark dataset composed of 19,203 CXR images from a multinational cohort of 16,656 patients from at least 51 countries, making it the largest, most diverse COVID-19 CXR dataset in open access form. The COVID-Net CXR-2 network achieves sensitivity and positive predictive value of 95.5 and 97.0%, respectively, and was audited in a transparent and responsible manner. Explainability-driven performance validation was used during auditing to gain deeper insights in its decision-making behavior and to ensure clinically relevant factors are leveraged for improving trust in its usage. Radiologist validation was also conducted, where select cases were reviewed and reported on by two board-certified radiologists with over 10 and 19 years of experience, respectively, and showed that the critical factors leveraged by COVID-Net CXR-2 are consistent with radiologist interpretations.

TB-Net: A Tailored, Self-Attention Deep Convolutional Neural Network Design for Detection of Tuberculosis Cases From Chest X-Ray Images.

Tuberculosis (TB) remains a global health problem, and is the leading cause of death from an infectious disease. A crucial step in the treatment of tuberculosis is screening high risk populations and the early detection of the disease, with chest x-ray (CXR) imaging being the most widely-used imaging modality. As such, there has been significant recent interest in artificial intelligence-based TB screening solutions for use in resource-limited scenarios where there is a lack of trained healthcare workers with expertise in CXR interpretation. Motivated by this pressing need and the recent recommendation by the World Health Organization (WHO) for the use of computer-aided diagnosis of TB in place of a human reader, we introduce TB-Net, a self-attention deep convolutional neural network tailored for TB case screening. We used CXR data from a multi-national patient cohort to train and test our models. A machine-driven design exploration approach leveraging generative synthesis was used to build a highly customized deep neural network architecture with attention condensers. We conducted an explainability-driven performance validation process to validate TB-Net's decision-making behavior. Experiments on CXR data from a multi-national patient cohort showed that the proposed TB-Net is able to achieve accuracy/sensitivity/specificity of 99.86/100.0/99.71%. Radiologist validation was conducted on select cases by two board-certified radiologists with over 10 and 19 years of experience, respectively, and showed consistency between radiologist interpretation and critical factors leveraged by TB-Net for TB case detection for the case where radiologists identified anomalies. The proposed TB-Net not only achieves high tuberculosis case detection performance in terms of sensitivity and specificity, but also leverages clinically relevant critical factors in its decision making process. While not a production-ready solution, we hope that the open-source release of TB-Net as part of the COVID-Net initiative will support researchers, clinicians, and citizen data scientists in advancing this field in the fight against this global public health crisis.

Radiogenomic Models Using Machine Learning Techniques to Predict EGFR Mutations in Non-Small Cell Lung Cancer.

BACKGROUND: The purpose of this study was to build radiogenomics models from texture signatures derived from computed tomography (CT) and 18F-FDG PET-CT (FDG PET-CT) images of non-small cell lung cancer (NSCLC) with and without epidermal growth factor receptor (EGFR) mutations. METHODS: Fifty patients diagnosed with NSCLC between 2011 and 2015 and with known EGFR mutation status were retrospectively identified. Texture features extracted from pretreatment CT and FDG PET-CT images by manual contouring of the primary tumor were used to develop multivariate logistic regression (LR) models to predict EGFR mutations in exon 19 and exon 20. RESULTS: An LR model evaluating FDG PET-texture features was able to differentiate EGFR mutant from wild type with an area under the curve (AUC), sensitivity, specificity, and accuracy of 0.87, 0.76, 0.66, and 0.71, respectively. The model derived from CT texture features had an AUC, sensitivity, specificity, and accuracy of 0.83, 0.84, 0.73, and 0.78, respectively. FDG PET-texture features that could discriminate between mutations in EGFR exon 19 and 21 demonstrated AUC, sensitivity, specificity, and accuracy of 0.86, 0.84, 0.73, and 0.78, respectively. Based on CT texture features, the AUC, sensitivity, specificity, and accuracy were 0.75, 0.81, 0.69, and 0.75, respectively. CONCLUSION: Non-small cell lung cancer texture analysis using FGD-PET and CT images can identify tumors with mutations in EGFR. Imaging signatures could be valuable for pretreatment assessment and prognosis in precision therapy.

MEDUSA: Multi-Scale Encoder-Decoder Self-Attention Deep Neural Network Architecture for Medical Image Analysis.

Medical image analysis continues to hold interesting challenges given the subtle characteristics of certain diseases and the significant overlap in appearance between diseases. In this study, we explore the concept of self-attention for tackling such subtleties in and between diseases. To this end, we introduce, a multi-scale encoder-decoder self-attention (MEDUSA) mechanism tailored for medical image analysis. While self-attention deep convolutional neural network architectures in existing literature center around the notion of multiple isolated lightweight attention mechanisms with limited individual capacities being incorporated at different points in the network architecture, MEDUSA takes a significant departure from this notion by possessing a single, unified self-attention mechanism with significantly higher capacity with multiple attention heads feeding into different scales in the network architecture. To the best of the authors' knowledge, this is the first "single body, multi-scale heads" realization of self-attention and enables explicit global context among selective attention at different levels of representational abstractions while still enabling differing local attention context at individual levels of abstractions. With MEDUSA, we obtain state-of-the-art performance on multiple challenging medical image analysis benchmarks including COVIDx, Radiological Society of North America (RSNA) RICORD, and RSNA Pneumonia Challenge when compared to previous work. Our MEDUSA model is publicly available.

COVID-Net CXR-S: Deep Convolutional Neural Network for Severity Assessment of COVID-19 Cases from Chest X-ray Images.

The world is still struggling in controlling and containing the spread of the COVID-19 pandemic caused by the SARS-CoV-2 virus. The medical conditions associated with SARS-CoV-2 infections have resulted in a surge in the number of patients at clinics and hospitals, leading to a significantly increased strain on healthcare resources. As such, an important part of managing and handling patients with SARS-CoV-2 infections within the clinical workflow is severity assessment, which is often conducted with the use of chest X-ray (CXR) images. In this work, we introduce COVID-Net CXR-S, a convolutional neural network for predicting the airspace severity of a SARS-CoV-2 positive patient based on a CXR image of the patient's chest. More specifically, we leveraged transfer learning to transfer representational knowledge gained from over 16,000 CXR images from a multinational cohort of over 15,000 SARS-CoV-2 positive and negative patient cases into a custom network architecture for severity assessment. Experimental results using the RSNA RICORD dataset showed that the proposed COVID-Net CXR-S has potential to be a powerful tool for computer-aided severity assessment of CXR images of COVID-19 positive patients. Furthermore, radiologist validation on select cases by two board-certified radiologists with over 10 and 19 years of experience, respectively, showed consistency between radiologist interpretation and critical factors leveraged by COVID-Net CXR-S for severity assessment. While not a production-ready solution, the ultimate goal for the open source release of COVID-Net CXR-S is to act as a catalyst for clinical scientists, machine learning researchers, as well as citizen scientists to develop innovative new clinical decision support solutions for helping clinicians around the world manage the continuing pandemic.

COVID-Net CT-2: Enhanced Deep Neural Networks for Detection of COVID-19 From Chest CT Images Through Bigger, More Diverse Learning.

The COVID-19 pandemic continues to rage on, with multiple waves causing substantial harm to health and economies around the world. Motivated by the use of computed tomography (CT) imaging at clinical institutes around the world as an effective complementary screening method to RT-PCR testing, we introduced COVID-Net CT, a deep neural network tailored for detection of COVID-19 cases from chest CT images, along with a large curated benchmark dataset comprising 1,489 patient cases as part of the open-source COVID-Net initiative. However, one potential limiting factor is restricted data quantity and diversity given the single nation patient cohort used in the study. To address this limitation, in this study we introduce enhanced deep neural networks for COVID-19 detection from chest CT images which are trained using a large, diverse, multinational patient cohort. We accomplish this through the introduction of two new CT benchmark datasets, the largest of which comprises a multinational cohort of 4,501 patients from at least 16 countries. To the best of our knowledge, this represents the largest, most diverse multinational cohort for COVID-19 CT images in open-access form. Additionally, we introduce a novel lightweight neural network architecture called COVID-Net CT S, which is significantly smaller and faster than the previously introduced COVID-Net CT architecture. We leverage explainability to investigate the decision-making behavior of the trained models and ensure that decisions are based on relevant indicators, with the results for select cases reviewed and reported on by two board-certified radiologists with over 10 and 30 years of experience, respectively. The best-performing deep neural network in this study achieved accuracy, COVID-19 sensitivity, positive predictive value, specificity, and negative predictive value of 99.0%/99.1%/98.0%/99.4%/99.7%, respectively. Moreover, explainability-driven performance validation shows consistency with radiologist interpretation by leveraging correct, clinically relevant critical factors. The results are promising and suggest the strong potential of deep neural networks as an effective tool for computer-aided COVID-19 assessment. While not a production-ready solution, we hope the open-source, open-access release of COVID-Net CT-2 and the associated benchmark datasets will continue to enable researchers, clinicians, and citizen data scientists alike to build upon them.

The challenge of deciding between home-discharge versus hospitalization in COVID-19 patients: The role of initial imaging and clinicolaboratory data.

BACKGROUND/OBJECTIVE: It is important to predict the COVID-19 patient's prognosis, particularly in countries with lack or deficiency of medical resource for patient's triage management. Currently, WHO guideline suggests using chest imaging in addition to clinicolaboratory evaluation to decide on triage between home-discharge versus hospitalization. We designed our study to validate this recommendation to guide clinicians. This study providing some suggestions to guide clinicians for better decision making in 2020. METHODS: In this retrospective study, patients with RT-PCR confirmed COVID-19 (N = 213) were divided in different clinical and management scenarios: home-discharge, ward hospitalization and ICU admission. We reviewed the patient's initial chest CT if available. We evaluated quantitative and qualitative characteristics of CT as well as relevant available clinicolaboratory data. Chi-square, One-Way ANOVA and Paired t-test were used for analysis. RESULTS: The finding showed that most patients with mixed patterns, pleural effusion, 5 lobes involved, total score ≥10, SpO2% ≤ 90, ESR (mm/h) ≥ 60 and WBC (103/µL) ≥ 8000 were hospitalized. Most patients with Ground-glass opacities only, ≤3 lobes involvement, peripheral distribution, SpO2% ≥ 95, ESR (mm/h) < 30 and WBC(103/µL) < 6000 were home-discharged. CONCLUSIONS: This study suggests the use of initial chest CT (qualitative and quantitative evaluation) in addition to initial clinicolaboratory data could be a useful supplementary method for clinical management and it is an excellent decision making tool (home-discharge versus ICU/Ward admission) for clinicians.

Risk Factors for Poor Outcome in Patients with Severe Viral Pneumonia on Chest CT during the COVID-19 Outbreak: a Perspective from Iran.

We investigated significant predictors of poor in-hospital outcomes for patients admitted with viral pneumonia during the COVID-19 outbreak in Tehran, Iran. Between February 22 and March 22, 2020, patients who were admitted to three university hospitals during the COVID-19 outbreak in Tehran, Iran were included. Demographic, clinical, laboratory, and chest CT scan findings were gathered. Two radiologists evaluated the distribution and CT features of the lesions and also scored the extent of lung involvement as the sum of three zones in each lung. Of 228 included patients, 45 patients (19.7%) required ICU admission and 34 patients (14.9%) died. According to regression analysis, older age (OR = 1.06; P < 0.001), blood oxygen saturation (SpO2) < 88% (OR = 2.88; P = 0.03), and higher chest CT total score (OR = 1.10; P = 0.03) were significant predictors for in-hospital death. The same three variables were also recognized as significant predictors for invasive respiratory support: SpO2 < 88% (OR = 3.97, P = 0.002), older age (OR = 1.05, P < 0.001), and higher CT total score (OR = 1.13, P = 0.008). Potential predictors of invasive respiratory support and in-hospital death in patients with viral pneumonia were older age, SpO2 < 88%, and higher chest CT score.

Novel coronavirus disease 2019: predicting prognosis with a computed tomography-based disease severity score and clinical laboratory data.

INTRODUCTION: Currently, there are known contributing factors but no comprehensive methods for predicting the mortality risk or intensive care unit (ICU) admission in patients with novel coronavirus disease 2019 (COVID­19). OBJECTIVES: The aim of this study was to explore risk factors for mortality and ICU admission in patients with COVID­19, using computed tomography (CT) combined with clinical laboratory data. PATIENTS AND METHODS: Patients with polymerase chain reaction-confirmed COVID­19 (n = 63) from university hospitals in Tehran, Iran, were included. All patients underwent CT examination. Subsequently, a total CT score and the number of involved lung lobes were calculated and compared against collected laboratory and clinical characteristics. Univariable and multivariable proportional hazard analyses were used to determine the association among CT, laboratory and clinical data, ICU admission, and in­hospital death. RESULTS: By univariable analysis, in­hospital mortality was higher in patients with lower oxygen saturation on admission (below 88%), higher CT scores, and a higher number of lung lobes (more than 4) involved with a diffuse parenchymal pattern. By multivariable analysis, in­hospital mortality was higher in those with oxygen saturation below 88% on admission and a higher number of lung lobes involved with a diffuse parenchymal pattern. The risk of ICU admission was higher in patients with comorbidities (hypertension and ischemic heart disease), arterial oxygen saturation below 88%, and pericardial effusion. CONCLUSIONS: We can identify factors affecting in­hospital death and ICU admission in COVID-19. This can help clinicians to determine which patients are likely to require ICU admission and to inform strategic healthcare planning in critical conditions such as the COVID­19 pandemic.

Predicting EGFR mutation status in lung cancer:Proposal for a scoring model using imaging and demographic characteristics.

OBJECTIVE: To determine if a combination of CT and demographic features can predict EGFR mutation status in bronchogenic carcinoma. METHODS: We reviewed demographic and CT features for patients with molecular profiling for resected non-small cell lung carcinoma. Using multivariate logistic regression, we identified features predictive of EGFR mutation. Prognostic factors identified from the logistic regression model were then used to build a more practical scoring system. RESULTS: A scoring system awarding 5 points for no or minimal smoking history, 3 points for tumours with ground glass component, 3 points for airbronchograms, 2 points for absence of preoperative evidence of nodal enlargement or metastases and 1 point for doubling time of more than a year, resulted in an AUROC of 0.861. A total score of at least 8 yielded a specificity of 95 %. On multivariate analysis sex was not found to be predictor of EGFR status. CONCLUSIONS: A weighted scoring system combining imaging and demographic data holds promise as a predictor of EGFR status. Further studies are necessary to determine reproducibility in other patient groups. A predictive score may help determine which patients would benefit from molecular profiling and may help inform treatment decisions when molecular profiling is not possible. KEY POINTS: • EGFR mutation-targeted chemotherapy for bronchogenic carcinoma has a high success rate. • Mutation testing is not possible in all patients. • EGFR associations include subsolid density, slow tumour growth and minimal/no smoking history. • Demographic or imaging features alone are weak predictors of EGFR status. • A scoring system, using imaging and demographic features, is more predictive.

Serotonin and histamine mediate gastroprotective effect of fluoxetine against experimentally-induced ulcers in rats.

Research in the treatment of gastric ulcer has involved the investigation of new alternatives, such as anti-depressant drugs. The present study was designed to investigate the gastroprotective effects of fluoxetine against indomethacin and alcohol induced gastric ulcers in rats and the potential mechanisms of that effect. Fluoxetine (20 mg/kg) was administered IP for 14 days. For comparative purposes, other rats were treated with ranitidine (30 mg/kg). Thereafter, after 24 h of fasting, INDO (100 mg/kg) or absolute alcohol (5 ml/kg) was administered to all rats (saline was administered to naïve controls) and rats in each group were sacrificed 5 h (for INDO rats) or 1 h (for alcohol rats) later. Macroscopic examination revealed that both fluoxetine and ranitidine decreased ulcer scores in variable ratios, which was supported by microscopic histopathological examination. Biochemical analysis of fluoxetine- or ranitidine-pre-treated host tissues demonstrated reductions in tumor necrosis factor (TNF)-α and myeloperoxidase (MPO) levels and concomitant increases in gastric pH, nitric oxide (NO) and reduced glutathione (GSH) contents. Fluoxetine, more than ranitidine, also resulted in serotonin and histamine levels nearest to control values. Moreover, immuno-histochemical analysis showed that fluoxetine markedly enhanced expression of cyclo-oxygenases COX-1 and COX-2 in both models; in comparison, ranitidine did not affect COX-1 expression in either ulcer model but caused moderate increases in COX-2 expression in INDO-induced hosts and high expression in alcohol-induced hosts. The results here indicated fluoxetine exhibited better gastroprotective effects than ranitidine and this could be due to anti-secretory, anti-oxidant, anti-inflammatory and anti-histaminic effects of the drug, as well as a stabilization of gastric serotonin levels.

Assessment of the reliability of reproducing two-dimensional resistivity models using an image processing technique.

This study attempts to combine the results of geophysical images obtained from three commonly used electrode configurations using an image processing technique in order to assess their capabilities to reproduce two-dimensional (2-D) resistivity models. All the inverse resistivity models were processed using the PCI Geomatica software package commonly used for remote sensing data sets. Preprocessing of the 2-D inverse models was carried out to facilitate further processing and statistical analyses. Four Raster layers were created, three of these layers were used for the input images and the fourth layer was used as the output of the combined images. The data sets were merged using basic statistical approach. Interpreted results show that all images resolved and reconstructed the essential features of the models. An assessment of the accuracy of the images for the four geologic models was performed using four criteria: the mean absolute error and mean percentage absolute error, resistivity values of the reconstructed blocks and their displacements from the true models. Generally, the blocks of the images of maximum approach give the least estimated errors. Also, the displacement of the reconstructed blocks from the true blocks is the least and the reconstructed resistivities of the blocks are closer to the true blocks than any other combined used. Thus, it is corroborated that when inverse resistivity models are combined, most reliable and detailed information about the geologic models is obtained than using individual data sets.