Prediction of COVID-19 prognosis by heterogeneity analysis based on chest CT scans (preprint)

ABSTRACT

This study aimed to estimate clinical outcome in individual COVID-19 patient by using histogram heterogeneity analysis based on CT opacities. 57 COVID-19 cases’ medical records were retrospectively reviewed from a designated hospital in Wuhan, China. Two characteristic lung abnormity opacities, ground-glass opacity (GGO) and consolidation opacity (CLO) were drawn on CT images to identify the heterogeneity by using quantitative histogram analysis. The parameters (mean, mode, kurtosis, skewness) derived from histograms evaluate the accuracy of clinical classification and outcome prediction. Nomograms were built to predict the risk of death and median length of hospital stays (LOS), respectively. The most highly frequency of lung abnormalities was GGO mixed with CLO in both survival population (26 in 42, 61.9%) and died population (10 in 15, 66.7%). The best performance heterogeneity parameters to discriminate severe type from mild/moderate counterparts were as following GGO_skewness specificity=66.67%, sensitivity=78.12%, AUC=0.706; CLO_mean specificity=70.00%, sensitivity=76.92%, AUC=0.746. Nomogram based on histogram parameters can predict the individual risk of death and the prolonged median LOS of COVID-19 patients. C-indexes were 0.763 and 0.888 for risk of death and prolonged median LOS, respectively. The histogram analysis method based on GGO and CLO has the ability for individual risk prediction in COVID-19 patients.