Skip to main content

Table 5 Summary of the statistical results regarding “step 2 blinded validation analysis”

From: COVID-19 clinico-radiological mismatch: a proposal for a novel combined morphologic/volumetric CT severity score with blinded validation

  Mild CTSI Moderate CTSI Severe CTSI  
Mild clinical disease 86 (91%) 6 2 (with patchy crazy-paving and early fibrosis) Total (disease) = 94
Borderline severe clinical disease 1 14 (82%) 2 (with patchy crazy-paving and early fibrosis) Total (disease) = 17
Severe or critical disease 4 3 14 (67%) Total (disease) = 21
Total (CT) 91 (69%) 23 (17%) 18 (14%) Total = 132
Prevalence 71.2% 12.9% 15.9%  
Sensitivity 91.5% 82.4% 66.7% Mean = 80.2%
Specificity 86.8% 92.2% 96.4% Mean = 91.8%
PPV (positive predictive value) 94.5% 60.9% 77.8% Mean = 77.7%
NPV (negative predictive value) 80.5% 97.2% 93.9% Mean = 90.5%
IOA (inter-observer agreement) 87.5% 81.3% 100% Mean = 89.6%
AUC (area under curve) 0.9003 0.83007 0.7765 Mean = 0.8356
Confidence interval 0.832 0.748 0.684 Mean = 0.7546
Accuracy 90.1% 90.9% 91.7% Mean = 90.9%
Validity of the test Excellent Good Fair