JMIR Publications

ABSTRACT

Background:

An artificial intelligence (AI) model using chest radiography (CXR) may provide good performance in predicting the prognosis of COVID-19.

Objective:

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.

Methods:

This retrospective longitudinal study included patients hospitalized at multiple medical centers for COVID-19 between February 2020 and October 2020. The patients in Boramae Medical Center were randomly classified into training, validation, and internal testing sets (8:1:1). An AI model using initial CXR as input, a logistic regression model using clinical information, and a combined model using the output of the AI model (CXR score) and clinical information were developed and trained to predict the hospital length of stay (LOS) ≤2 weeks, need for oxygen supplementation, and risk of acute respiratory distress syndrome (ARDS). The models were externally validated in the KICC-19 cohort for discrimination and calibration.

Results:

The AI model using CXR and the logistic regression model using clinical variables were suboptimal to predict hospital LOS ≤2 weeks or need for oxygen supplementation but showed acceptable performance to predict ARDS (AI model area under the curve [AUC] = 0.782; 95% confidence interval [CI] = 0.720-0.845; logistic regression model AUC = 0.878; 95% CI = 0.838-0.919). The combined model showed a better performance in predicting oxygen supplementation (AUC, 0.704; 95% CI, 0.646-0.762) and ARDS (AUC, 0.890; 95% CI, 0.853-0.928) compared to CXR score alone. The AI and combined models showed good calibration for predicting ARDS (p  > 0.05).

Conclusions:

The prediction model combined with the CXR score and clinical information was externally validated as having acceptable performance in predicting severe illness and excellent performance in predicting ARDS in COVID-19 patients.