Deep Learning With Chest Radiographs for Making Prognoses in Patients With COVID-19: Retrospective Cohort Study

• Published in: Journal of medical Internet research Vol. 25; no. 1; p. e42717
• Main Authors: Lee, Hyun Woo, Yang, Hyun Jun, Kim, Hyungjin, Kim, Ue-Hwan, Kim, Dong Hyun, Yoon, Soon Ho, Ham, Soo-Youn, Nam, Bo Da, Chae, Kum Ju, Lee, Dabee, Yoo, Jin Young, Bak, So Hyeon, Kim, Jin Young, Kim, Jin Hwan, Kim, Ki Beom, Jung, Jung Im, Lim, Jae-Kwang, Lee, Jong Eun, Chung, Myung Jin, Lee, Young Kyung, Kim, Young Seon, Lee, Sang Min, Kwon, Woocheol, Park, Chang Min, Kim, Yun-Hyeon, Jeong, Yeon Joo, Jin, Kwang Nam, Goo, Jin Mo
• Format: Journal Article
• Language: English
• Published: Canada Gunther Eysenbach MD MPH, Associate Professor 16.02.2023; JMIR Publications
• Subjects: Artificial Intelligence; Chest; Clinical information; Clinical outcomes; Clinical variables; Cohort analysis; COVID-19; COVID-19 - diagnostic imaging; Datasets; Deep Learning; Discrimination; Dyspnea; Extracorporeal membrane oxygenation; Hospitalized; Humans; Learning; Length of stay; Longitudinal Studies; Neural networks; Original Paper; Oxygen; Pneumonia; Prediction models; Prognosis; Psychological distress; Radiography; Respiratory Distress Syndrome; Retrospective Studies; Severe acute respiratory syndrome coronavirus 2; COVID-19; deep learning; AI model; prediction model; clinical outcome; radiography, thoracic; machine learning; prognosis; artificial intelligence; medical imaging
• ISSN: 1438-8871; 1439-4456; 1438-8871
• DOI: 10.2196/42717

An artificial intelligence (AI) model using chest radiography (CXR) may provide good performance in making prognoses for COVID-19. We aimed to develop and validate a prediction model using CXR based on an AI model and clinical variables to predict clinical outcomes in patients with COVID-19. This retrospective longitudinal study included patients hospitalized for COVID-19 at multiple COVID-19 medical centers between February 2020 and October 2020. Patients at Boramae Medical Center were randomly classified into training, validation, and internal testing sets (at a ratio of 8:1:1, respectively). An AI model using initial CXR images as input, a logistic regression model using clinical information, and a combined model using the output of the AI model (as CXR score) and clinical information were developed and trained to predict hospital length of stay (LOS) ≤2 weeks, need for oxygen supplementation, and acute respiratory distress syndrome (ARDS). The models were externally validated in the Korean Imaging Cohort of COVID-19 data set for discrimination and calibration. The AI model using CXR and the logistic regression model using clinical variables were suboptimal to predict hospital LOS ≤2 weeks or the need for oxygen supplementation but performed acceptably in the prediction of ARDS (AI model area under the curve [AUC] 0.782, 95% CI 0.720-0.845; logistic regression model AUC 0.878, 95% CI 0.838-0.919). The combined model performed better in predicting the need for oxygen supplementation (AUC 0.704, 95% CI 0.646-0.762) and ARDS (AUC 0.890, 95% CI 0.853-0.928) compared to the CXR score alone. Both the AI and combined models showed good calibration for predicting ARDS (P=.079 and P=.859). The combined prediction model, comprising the CXR score and clinical information, was externally validated as having acceptable performance in predicting severe illness and excellent performance in predicting ARDS in patients with COVID-19.