Accurate and rapid diagnosis of COVID-19 suspected cases plays a crucial role in timely quarantine and medical treatment. Developing a deep learning-based model for automatic COVID-19 diagnosis on chest CT is helpful to counter the outbreak of SARS-CoV-2. A weaklysupervised deep learning framework was developed using 3D CT volumes for COVID-19 classification and lesion localization. For each patient, the lung region was segmented using a pre-trained UNet; then the segmented 3D lung region was fed into a 3D deep neural network to predict the probability of COVID-19 infectious; the COVID- 19 lesions are localized by combining the activation regions in the classification network and the unsupervised connected components. 499 CT volumes were used for training and 131 CT volumes were used for testing. Our algorithm obtained 0.959 ROC AUC and 0.976 PR AUC. When using a probability threshold of 0.5 to classify COVIDpositive and COVID-negative, the algorithm obtained an accuracy of 0.901, a positive predictive value of 0.840 and a very high negative predictive value of 0.982. The algorithm took only 1.93 seconds to process a single patient’s CT volume using a dedicated GPU. Our weakly-supervised deep learning model can accurately predict the COVID-19 infectious probability and discover lesion regions in chest CT without the need for annotating the lesions for training. The easily-trained and high-performance deep learning algorithm provides a fast way to identify COVID-19 patients, which is beneficial to control the outbreak of SARS-CoV-2. The developed deep learning software is available at https: //github:com/sydney0zq/covid-19-detection.

Accurate and rapid diagnosis of COVID-19 suspected cases plays a crucial role in timely quarantine and medical treatment. Developing a deep learning-based model for automatic COVID-19 detection on chest CT is helpful to counter the outbreak of SARS-CoV-2. A weakly-supervised deep learning-based software system was developed using 3D CT volumes to detect COVID-19. For each patient, the lung region was segmented using a pre-trained UNet; then the segmented 3D lung region was fed into a 3D deep neural network to predict the probability of COVID-19 infectious. 499 CT volumes collected from Dec. 13, 2019, to Jan. 23, 2020, were used for training and 131 CT volumes collected from Jan 24, 2020, to Feb 6, 2020, were used for testing. The deep learning algorithm obtained 0.959 ROC AUC and 0.976 PR AUC. There was an operating point with 0.907 sensitivity and 0.911 specificity in the ROC curve. When using a probability threshold of 0.5 to classify COVID-positive and COVID-negative, the algorithm obtained an accuracy of 0.901, a positive predictive value of 0.840 and a very high negative predictive value of 0.982. The algorithm took only 1.93 seconds to process a single patient’s CT volume using a dedicated GPU. Our weakly-supervised deep learning model can accurately predict the COVID-19 infectious probability in chest CT volumes without the need for annotating the lesions for training. The easily-trained and highperformance deep learning algorithm provides a fast way to identify COVID-19 patients, which is beneficial to control the outbreak of SARS-CoV-2. The developed deep learning software is available at https://github.com/sydney0zq/covid-19-detection .

Comparison of contrast-enhanced T1-weighted imaging using DANTE-SPACE, PETRA, and MPRAGE: a clinical evaluation of brain tumors at 3 Tesla Qing Fu, Qi-Guang Cheng, Xiang-Chuang Kong, Ding-Xi Liu, Yi-Hao Guo, John Grinstead, Xiao-Yong Zhang, Zi-Qiao Lei, Chuan-Sheng Zheng

Purpose: To qualitatively and quantitatively compare the image quality and diagnostic performance of turbo gradient and spin echo PROPELLER diffusion-weighted imaging (TGSE-PROPELLER-DWI) vs. readout-segmented echo-planar imaging (rs-EPI) in the evaluation of orbital tumors.Materials and Methods: A total of 43 patients with suspected orbital tumors were enrolled to perform the two DWIs with comparable spatial resolution on 3T. The overall image qualities, geometric distortions, susceptibility artifacts, and lesion conspicuities were scored by using a four-point scale (1, poor; 4, excellent). Quantitative measurements, including contrast-to-noise ratios (CNRs), apparent diffusion coefficients (ADCs), geometric distortion rates (GDRs), and lesion sizes, were calculated and compared. The two ADCs for differentiating malignant from benign orbital tumors were evaluated. Wilcoxon signed-rank test, Kappa statistic, and receiver operating characteristics (ROC) curves were used.Results: TGSE-PROPELLER-DWI performed superior in all subjective scores and quantitative GDR evaluation than rs-EPI (p < 0.001), and excellent interobserver agreement was obtained for Kappa value ranging from 0.876 to 1.000. ADC_lesion of TGSE-PROPELLER-DWI was significantly higher than those of rs-EPI (p < 0.001). Mean ADC of malignant tumors was significantly lower than that of benign tumors both in two DWIs. However, the AUC for differentiating malignant and benign tumors showed no significant difference in the two DWIs (0.860 vs. 0.854, p = 0.7448). Sensitivity and specificity could achieve 92.86% and 72.73% for TGSE-PROPELLER-DWI with a cutoff value of 1.23 × 10^–3 mm²/s, and 85.71% and 81.82% for rs-EPI with a cutoff value of 0.99 × 10^–3 mm²/s.Conclusion: Compared with rs-EPI, TGSE-PROPELLER-DWI showed minimized geometric distortion and susceptibility artifacts significantly improved the image quality for orbital tumors and achieved comparable diagnostic performance in differentiating malignant and benign orbital tumors.