ABSTRACT
OBJECTIVES: An artificial intelligence model was adopted to identify mild COVID-19 pneumonia from computed tomography (CT) volumes, and its diagnostic performance was then evaluated. METHODS: In this retrospective multicenter study, an atrous convolution-based deep learning model was established for the computer-assisted diagnosis of mild COVID-19 pneumonia. The dataset included 2087 chest CT exams collected from four hospitals between 1 January 2019 and 31 May 2020. The true positive rate, true negative rate, receiver operating characteristic curve, area under the curve (AUC) and convolutional feature map were used to evaluate the model. RESULTS: The proposed deep learning model was trained on 1538 patients and tested on an independent testing cohort of 549 patients. The overall sensitivity was 91.5% (195/213; p < 0.001, 95% CI: 89.2-93.9%), the overall specificity was 90.5% (304/336; p < 0.001, 95% CI: 88.0-92.9%) and the general AUC value was 0.955 (p < 0.001). CONCLUSIONS: A deep learning model can accurately detect COVID-19 and serve as an important supplement to the COVID-19 reverse transcription-polymerase chain reaction (RT-PCR) test. KEY POINTS: ⢠The implementation of a deep learning model to identify mild COVID-19 pneumonia was confirmed to be effective and feasible. ⢠The strategy of using a binary code instead of the region of interest label to identify mild COVID-19 pneumonia was verified. ⢠This AI model can assist in the early screening of COVID-19 without interfering with normal clinical examinations.
Subject(s)
COVID-19 , Deep Learning , Artificial Intelligence , Humans , Retrospective Studies , SARS-CoV-2 , Tomography, X-Ray ComputedABSTRACT
CONTEXT.: Covert severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infections could be seeding new outbreaks. How to identify asymptomatic SARS-CoV-2 infections early has become a global focus. OBJECTIVE.: To explore the roles of immunoglobulin M (IgM) and immunoglobulin G (IgG) antibodies detection, nucleic acid tests, and computed tomography (CT) scanning to identify asymptomatic SARS-CoV-2 infection. DESIGN.: The clinical data of 389 individuals with close contacts, including in general characteristics, SARS-CoV-2 etiology, serum-specific IgM and IgG antibody detection and CT imaging results, were systematically analyzed. RESULTS.: The present study showed that only 89 of 389 individuals with close contacts were positive after the first nucleic acid test, while 300 individuals were still negative after 2 nucleic acid tests. Among the 300 individuals, 75 did not have pneumonia, and the other 225 individuals had pulmonary imaging changes. A total of 143 individuals were eventually diagnosed as having asymptomatic infection through IgM antibody and IgG antibody detection. The sensitivity, specificity, and false-negative rate of IgM and IgG antibody detection were approximately 97.1% (347 of 357), 95.3% (204 of 214), and 4.67% (10 of 214), respectively. It also indicated that during approximately 2 weeks, most individuals were both IgM positive and IgG positive, accounting for 68.57% (72 of 105). During approximately 3 weeks, the proportion of IgM-positive and IgG-positive individuals decreased to 8.57% (9 of 105), and the proportion of IgM-negative and IgG-positive individuals increased to 76.19% (80 of 105). CONCLUSIONS.: There are highlighted prospects of IgM/IgG antibody detection as a preferred method in identifying the individuals with asymptomatic SARS-CoV-2 infection, especially combined with nucleic acid tests and pulmonary CT scanning.