Automated Quantification of CT Patterns Associated with COVID-19 from Chest CT.

PURPOSE: To present a method that automatically segments and quantifies abnormal CT patterns commonly present in coronavirus disease 2019 (COVID-19), namely ground glass opacities and consolidations. MATERIALS AND METHODS: In this retrospective study, the proposed method takes as input a non-contrasted chest CT and segments the lesions, lungs, and lobes in three dimensions, based on a dataset of 9749 chest CT volumes. The method outputs two combined measures of the severity of lung and lobe involvement, quantifying both the extent of COVID-19 abnormalities and presence of high opacities, based on deep learning and deep reinforcement learning. The first measure of (PO, PHO) is global, while the second of (LSS, LHOS) is lobe-wise. Evaluation of the algorithm is reported on CTs of 200 participants (100 COVID-19 confirmed patients and 100 healthy controls) from institutions from Canada, Europe and the United States collected between 2002-Present (April 2020). Ground truth is established by manual annotations of lesions, lungs, and lobes. Correlation and regression analyses were performed to compare the prediction to the ground truth. RESULTS: Pearson correlation coefficient between method prediction and ground truth for COVID-19 cases was calculated as 0.92 for PO (P < .001), 0.97 for PHO (P < .001), 0.91 for LSS (P < .001), 0.90 for LHOS (P < .001). 98 of 100 healthy controls had a predicted PO of less than 1%, 2 had between 1-2%. Automated processing time to compute the severity scores was 10 seconds per case compared to 30 minutes required for manual annotations. CONCLUSION: A new method segments regions of CT abnormalities associated with COVID-19 and computes (PO, PHO), as well as (LSS, LHOS) severity scores.

A Soft Labeling Approach to Develop Automated Algorithms that Incorporate Uncertainty in Pulmonary Opacification on Chest CT using COVID-19 Pneumonia.

RATIONALE AND OBJECTIVES: Hard data labels for automated algorithm training are binary and cannot incorporate uncertainty between labels. We proposed and evaluated a soft labeling methodology to quantify opacification and percent well-aerated lung (%WAL) on chest CT, that considers uncertainty in segmenting pulmonary opacifications and reduces labeling burden. MATERIALS AND METHODS: We retrospectively sourced 760 COVID-19 chest CT scans from five international centers between January and June 2020. We created pixel-wise labels for >27,000 axial slices that classify three pulmonary opacification patterns: pure ground-glass, crazy-paving, consolidation. We also quantified %WAL as the total area of lung without opacifications. Inter-user hard label variability was quantified using Shannon entropy (range=0-1.39, low-high entropy/variability). We incorporated a soft labeling and modeling cycle following an initial model with hard labels and compared performance using point-wise accuracy and intersection-over-union of opacity labels with ground-truth, and correlation with ground-truth %WAL. RESULTS: Hard labels annotated by 12 radiologists demonstrated large inter-user variability (3.37% of pixels achieved complete agreement). Our soft labeling approach increased point-wise accuracy from 60.0% to 84.3% (p=0.01) compared to hard labeling at predicting opacification type and area involvement. The soft label model accurately predicted %WAL (R=0.900) compared to the hard label model (R=0.856), but the improvement was not statistically significant (p=0.349). CONCLUSION: Our soft labeling approach increased accuracy for automated quantification and classification of pulmonary opacification on chest CT. Although we developed the model on COVID-19, our intent is broad application for pulmonary opacification contexts and to provide a foundation for future development using soft labeling methods.

The Effect of Mask Wearing on the Accuracy of Radiology Reports in an Academic Hospital Setting.

PURPOSE: In response to the pandemic, some public health agencies recommend the wearing of surgical masks in indoor spaces including radiology common reporting rooms. We aim to demonstrate whether mask wearing may lead to increased errors incidence in radiology reports. MATERIALS AND METHODS: Our prospective studywas conveyed in 2 parts. Firstly, the participants were surveyed if they believed that mask affected dictation. Then participants performed a dictation: they read artificial radiology reports using a commercial voice recognition (VR) system. They performed this task 5 times, each time donning a different mask in random order: a surgical mask, surgical visor, N-95, combination of 2 surgical masks and no mask. Error rates were compared with the Friedman test followed by pairwise Wilcoxon with bootstrapping. Multivariate Poisson regression was performed to test for interaction effects between potential predictors. RESULTS: 52 members of an academic radiology department participatedin the study (January - March 2021) . 65.4% of survey participants did not think or were not sure whether mask wearing could affect dictation process. Treating the no-mask condition as baseline, our study found that mean error rates significantly increased up to 2 times the baseline rate when a surgical mask, surgical visor, N-95 or a combination of 2 masks was donned (p < 0.0001). No significant differences in error rates were found between the different mask types (p > 0.05). Error rates were higher for participants with shorter VR training time (p < 0.0001) or who were non-native English speakers (p < 0.0001). There were no interaction effects between mask type, VR training time or English nativity, suggesting these variables to be independent predictors for error rate. Academic rank did not significantly affect the error rate. CONCLUSION: radiologists underestimate the influence of masks on dictation accuracy. mask wearing may lead to significant increase in dictational errors.

Machine learning automatically detects COVID-19 using chest CTs in a large multicenter cohort.

OBJECTIVES: To investigate machine learning classifiers and interpretable models using chest CT for detection of COVID-19 and differentiation from other pneumonias, interstitial lung disease (ILD) and normal CTs. METHODS: Our retrospective multi-institutional study obtained 2446 chest CTs from 16 institutions (including 1161 COVID-19 patients). Training/validation/testing cohorts included 1011/50/100 COVID-19, 388/16/33 ILD, 189/16/33 other pneumonias, and 559/17/34 normal (no pathologies) CTs. A metric-based approach for the classification of COVID-19 used interpretable features, relying on logistic regression and random forests. A deep learning-based classifier differentiated COVID-19 via 3D features extracted directly from CT attenuation and probability distribution of airspace opacities. RESULTS: Most discriminative features of COVID-19 are the percentage of airspace opacity and peripheral and basal predominant opacities, concordant with the typical characterization of COVID-19 in the literature. Unsupervised hierarchical clustering compares feature distribution across COVID-19 and control cohorts. The metrics-based classifier achieved AUC = 0.83, sensitivity = 0.74, and specificity = 0.79 versus respectively 0.93, 0.90, and 0.83 for the DL-based classifier. Most of ambiguity comes from non-COVID-19 pneumonia with manifestations that overlap with COVID-19, as well as mild COVID-19 cases. Non-COVID-19 classification performance is 91% for ILD, 64% for other pneumonias, and 94% for no pathologies, which demonstrates the robustness of our method against different compositions of control groups. CONCLUSIONS: Our new method accurately discriminates COVID-19 from other types of pneumonia, ILD, and CTs with no pathologies, using quantitative imaging features derived from chest CT, while balancing interpretability of results and classification performance and, therefore, may be useful to facilitate diagnosis of COVID-19. KEY POINTS: • Unsupervised clustering reveals the key tomographic features including percent airspace opacity and peripheral and basal opacities most typical of COVID-19 relative to control groups. • COVID-19-positive CTs were compared with COVID-19-negative chest CTs (including a balanced distribution of non-COVID-19 pneumonia, ILD, and no pathologies). Classification accuracies for COVID-19, pneumonia, ILD, and CT scans with no pathologies are respectively 90%, 64%, 91%, and 94%. • Our deep learning (DL)-based classification method demonstrates an AUC of 0.93 (sensitivity 90%, specificity 83%). Machine learning methods applied to quantitative chest CT metrics can therefore improve diagnostic accuracy in suspected COVID-19, particularly in resource-constrained environments.

RSNA Expert Consensus Statement on Reporting Chest CT Findings Related to COVID-19: Interobserver Agreement Between Chest Radiologists.

PURPOSE: To assess the interobserver variability between chest radiologists in the interpretation of the Radiological Society of North America (RSNA) expert consensus statement reporting guidelines in patients with suspected coronavirus disease 2019 (COVID-19) pneumonia in a setting with limited reverse transcription polymerase chain reaction testing availability. METHODS: Chest computed tomography (CT) studies in 303 consecutive patients with suspected COVID-19 were reviewed by 3 fellowship-trained chest radiologists. Cases were assigned an impression of typical, indeterminate, atypical, or negative for COVID-19 pneumonia according to the RSNA expert consensus statement reporting guidelines, and interobserver analysis was performed. Objective CT features associated with COVID-19 pneumonia and distribution of findings were recorded. RESULTS: The Fleiss kappa for all observers was almost perfect for typical (0.815), atypical (0.806), and negative (0.962) COVID-19 appearances (P < .0001) and substantial (0.636) for indeterminate COVID-19 appearance (P < .0001). Using Cramer V analysis, there were very strong correlations between all radiologists' interpretations, statistically significant for all (typical, indeterminate, atypical, and negative) COVID-19 appearances (P < .001). Objective CT imaging findings were recorded in similar percentages of typical cases by all observers. CONCLUSION: The RSNA expert consensus statement on reporting chest CT findings related to COVID-19 demonstrates substantial to almost perfect interobserver agreement among chest radiologists in a relatively large cohort of patients with clinically suspected COVID-19. It therefore serves as a reliable reference framework for radiologists to accurately communicate their level of suspicion based on the presence of evidence-based objective findings.

The Effectiveness of Image Augmentation in Deep Learning Networks for Detecting COVID-19: A Geometric Transformation Perspective.

Chest X-ray imaging technology used for the early detection and screening of COVID-19 pneumonia is both accessible worldwide and affordable compared to other non-invasive technologies. Additionally, deep learning methods have recently shown remarkable results in detecting COVID-19 on chest X-rays, making it a promising screening technology for COVID-19. Deep learning relies on a large amount of data to avoid overfitting. While overfitting can result in perfect modeling on the original training dataset, on a new testing dataset it can fail to achieve high accuracy. In the image processing field, an image augmentation step (i.e., adding more training data) is often used to reduce overfitting on the training dataset, and improve prediction accuracy on the testing dataset. In this paper, we examined the impact of geometric augmentations as implemented in several recent publications for detecting COVID-19. We compared the performance of 17 deep learning algorithms with and without different geometric augmentations. We empirically examined the influence of augmentation with respect to detection accuracy, dataset diversity, augmentation methodology, and network size. Contrary to expectation, our results show that the removal of recently used geometrical augmentation steps actually improved the Matthews correlation coefficient (MCC) of 17 models. The MCC without augmentation (MCC = 0.51) outperformed four recent geometrical augmentations (MCC = 0.47 for Data Augmentation 1, MCC = 0.44 for Data Augmentation 2, MCC = 0.48 for Data Augmentation 3, and MCC = 0.49 for Data Augmentation 4). When we retrained a recently published deep learning without augmentation on the same dataset, the detection accuracy significantly increased, with a χ McNema r ' s statistic 2 = 163 . 2 and a p-value of 2.23 × 10-37. This is an interesting finding that may improve current deep learning algorithms using geometrical augmentations for detecting COVID-19. We also provide clinical perspectives on geometric augmentation to consider regarding the development of a robust COVID-19 X-ray-based detector.

Radiological Society of North America (RSNA) Expert Consensus Statement Related to Chest CT Findings in COVID-19 Versus CO-RADS: Comparison of Reporting System Performance Among Chest Radiologists and End-User Preference.

PURPOSE: The RSNA expert consensus statement and CO-RADS reporting system assist radiologists in describing lung imaging findings in a standardized manner in patients under investigation for COVID-19 pneumonia and provide clarity in communication with other healthcare providers. We aim to compare diagnostic performance and inter-/intra-observer among chest radiologists in the interpretation of RSNA and CO-RADS reporting systems and assess clinician preference. METHODS: Chest CT scans of 279 patients with suspected COVID-19 who underwent RT-PCR testing were retrospectively and independently examined by 3 chest radiologists who assigned interpretation according to the RSNA and CO-RADS reporting systems. Inter-/intra-observer analysis was performed. Diagnostic accuracy of both reporting systems was calculated. 60 clinicians participated in a survey to assess end-user preference of the reporting systems. RESULTS: Both systems demonstrated almost perfect inter-observer agreement (Fleiss kappa 0.871, P < 0.0001 for RSNA; 0.876, P < 0.0001 for CO-RADS impressions). Intra-observer agreement between the 2 scoring systems using the equivalent categories was almost perfect (Fleiss kappa 0.90-0.92, P < 0.001). Positive predictive values were high, 0.798-0.818 for RSNA and 0.891-0.903 CO-RADS. Negative predictive value were similar, 0.573-0.585 for RSNA and 0.573-0.58 for CO-RADS. Specificity differed between the 2 systems, 68-73% for CO-RADS and 52-58% for RSNA with superior specificity of CO-RADS. Of 60 survey participants, the majority preferred the RSNA reporting system rather than CO-RADS for all options provided (66.7-76.7%; P < 0.05). CONCLUSIONS: RSNA and CO-RADS reporting systems are consistent and reproducible with near perfect inter-/intra-observer agreement and excellent positive predictive value. End-users preferred the reporting language in the RSNA system.

Quantification of Tomographic Patterns associated with COVID-19 from Chest CT.

PURPOSE: To present a method that automatically segments and quantifies abnormal CT patterns commonly present in coronavirus disease 2019 (COVID-19), namely ground glass opacities and consolidations. MATERIALS AND METHODS: In this retrospective study, the proposed method takes as input a non-contrasted chest CT and segments the lesions, lungs, and lobes in three dimensions, based on a dataset of 9749 chest CT volumes. The method outputs two combined measures of the severity of lung and lobe involvement, quantifying both the extent of COVID-19 abnormalities and presence of high opacities, based on deep learning and deep reinforcement learning. The first measure of (PO, PHO) is global, while the second of (LSS, LHOS) is lobewise. Evaluation of the algorithm is reported on CTs of 200 participants (100 COVID-19 confirmed patients and 100 healthy controls) from institutions from Canada, Europe and the United States collected between 2002-Present (April, 2020). Ground truth is established by manual annotations of lesions, lungs, and lobes. Correlation and regression analyses were performed to compare the prediction to the ground truth. RESULTS: Pearson correlation coefficient between method prediction and ground truth for COVID-19 cases was calculated as 0.92 for PO (P < .001), 0.97 for PHO(P < .001), 0.91 for LSS (P < .001), 0.90 for LHOS (P < .001). 98 of 100 healthy controls had a predicted PO of less than 1%, 2 had between 1-2%. Automated processing time to compute the severity scores was 10 seconds per case compared to 30 minutes required for manual annotations. CONCLUSION: A new method segments regions of CT abnormalities associated with COVID-19 and computes (PO, PHO), as well as (LSS, LHOS) severity scores.

The Performance of Deep Neural Networks in Differentiating Chest X-Rays of COVID-19 Patients From Other Bacterial and Viral Pneumonias.

Chest radiography is a critical tool in the early detection, management planning, and follow-up evaluation of COVID-19 pneumonia; however, in smaller clinics around the world, there is a shortage of radiologists to analyze large number of examinations especially performed during a pandemic. Limited availability of high-resolution computed tomography and real-time polymerase chain reaction in developing countries and regions of high patient turnover also emphasizes the importance of chest radiography as both a screening and diagnostic tool. In this paper, we compare the performance of 17 available deep learning algorithms to help identify imaging features of COVID19 pneumonia. We utilize an existing diagnostic technology (chest radiography) and preexisting neural networks (DarkNet-19) to detect imaging features of COVID-19 pneumonia. Our approach eliminates the extra time and resources needed to develop new technology and associated algorithms, thus aiding the front-line healthcare workers in the race against the COVID-19 pandemic. Our results show that DarkNet-19 is the optimal pre-trained neural network for the detection of radiographic features of COVID-19 pneumonia, scoring an overall accuracy of 94.28% over 5,854 X-ray images. We also present a custom visualization of the results that can be used to highlight important visual biomarkers of the disease and disease progression.

Emergency Radiology During the COVID-19 Pandemic: The Canadian Association of Radiologists Recommendations for Practice.

Coronavirus Disease 2019 (COVID-19) is the disease caused by the novel coronavirus officially named the Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2), declared as a pandemic by the World Health Organization on March 11, 2020. The COVID-19 pandemic presents an unprecedented challenge to emergency radiology practice. The continuity of an effective emergency imaging service for both COVID-19 and non-COVID-19 patients is essential, while adhering to best infection control practices. Under the direction of the Board of the Canadian Association of Radiologists, this general guidance document has been synthesized by collaborative consensus of a group of emergency radiologists. These recommendations aim to assist radiologists involved in emergency diagnostic imaging to help mitigate the spread of COVID-19 and continue to add value to patient care in the emergency setting.

Interventional radiology and COVID-19: evidence-based measures to limit transmission.

As we face an explosion of COVID-19 cases and deal with an unprecedented set of circumstances all over the world, healthcare personnel are at the forefront, dealing with this emerging scenario. Certain subspecialties like interventional radiology entails a greater risk of acquiring and transmitting infection due to the close patient contact and invasive patient care the service provides. This makes it imperative to develop and set guidelines in place to limit transmission and utilize resources in an optimal fashion. A multi-tiered approach needs to be devised and monitored at the administrative level, taking into account the various staff and patient contact points. Based on these factors, work site and health force rearrangements need to be in place while enforcing segregation and disinfection parameters. We are putting forth an all-encompassing review of infection control measures that cover the dynamics of patient care and staff protocols that such a situation demands of an interventional department.

The Role of Emergency Radiology in COVID-19: From Preparedness to Diagnosis [Formula: see text].

Emergency trauma radiology, although a relatively new subspecialty of radiology, plays a critical role in both the diagnosis/triage of acutely ill patients, but even more important in providing leadership and taking the lead in the preparedness of imaging departments in dealing with novel highly infectious communicable diseases and mass casualties. This has become even more apparent in dealing with COVID-19, the disease caused by the novel coronavirus SARS-CoV-2, first emerged in late 2019. We review the symptoms, epidemiology, and testing for this disease. We discuss characteristic imaging findings of COVID-19 in relation to other modern coronavirus diseases including SARS and MERS. We discuss roles that community radiology clinics, outpatient radiology departments, and emergency radiology departments can play in the diagnosis of this disease. We review practical methods to reduce spread of infections within radiology departments.