Skip to main content

Table 3 The Performances of the Three Predictive Models

From: 18F-FDG-PET-based Radiomics signature predicts MGMT promoter methylation status in primary diffuse glioma

Models AUC (95%CI) ACC (95%CI) SEN (95%CI) SPE (95%CI) PPV (95%CI) NPV (95% CI)
Radiomics model
 Primary cohort 0.94 (0.93, 0.96) 91.3% (89.8, 93.3) 94.9% (93.1, 96.6) 87.5% (84.4, 90.7) 90.2% (87.8, 92.8) 93.3% (91.1, 95.6)
 Validation cohort 0.86 (0.83, 0.88) 77.8% (75.2, 80.3) 75.0% (71.5, 78.6) 81.3% (77.5, 84.9) 83.3% (80.0, 86.7) 72.2% (68.2, 76.1)
Clinical model
 Primary cohort 0.64 (0.61, 0.67) 64.8% (61.9, 67.9) 71.8% (68.1, 75.7) 56.3% (51.6, 61.1) 66.7% (62.9, 70.6) 62.1% (57.3, 67.2)
 Validation cohort 0.69 (0.66, 0.72) 66.4% (66.6, 72.3) 75.0% (71.3, 78.5) 62.5% (58.1, 67.0) 71.4% (67.9, 75.1) 66.7% (61.9, 71.1)
Fusion model
 Primary cohort 0.85 (0.83, 0.87) 64.8% (62.0, 67.7) 71.8% (68.1, 75.5) 56.3% (51.6, 60.9) 66.7% (63.0, 70.4) 62.1% (57.5, 66.7)
 Validation cohort 0.85 (0.82, 0.87) 72.7% (78.1, 82.9) 80.0% (76.6, 83.5) 62.5% (58.0, 67.0) 72.7% (69.3, 76.3) 71.4% (66.7, 76.1)
  1. Abbreviations: CI Confidence interval, AUC Area under receiver-operating characteristic curve, ACC Accuracy, SEN Sensitivity, SPE Specificity, PPV Positive predictive value, NPV Negative predictive value