Skip to main content

Table 1 The predicting performance for CT signs model, radiomics model and combined model

From: CT-based radiomics combined with signs: a valuable tool to help radiologist discriminate COVID-19 and influenza pneumonia

Model CT signs Radiomics CT signs + radiomics
Threshold 0.431 0.107 0.458
AUC (95% CI) 0.906 (0.86–0.953) 0.888 (0.8335–0.9417) 0.959 (0.930–0.987)
Sensitivity (95% CI) 0.865 (0.7774 to 0.9227) 0.865 (0.7774 to 0.9227) 0.899 (0.8168 to 0.9479)
Specificity (95% CI) 0.815 (0.7029 to 0.8927) 0.784 (0.6691 to 0.8683) 0.907 (0.8096 to 0.9603)
Positive prediction (95% CI) 0.865 (0.7774 to 0.9227) 0.846 (0.7569 to 0.9074) 0.930 (0.8532 to 0.9705)
Negative prediction (95% CI) 0.815 (0.7029 to 0.8927) 0.809 (0.6944 to 0.8891) 0.868 (0.7650 to 0.9310)
Accuracy (95% CI) 0.844 (0.7780 to 0.8936) 0.831 (0.7636 to 0.8826) 0.903 (0.8445 to 0.9411)
  1. AUC, area under curve