Skip to main content

Table 6 The ROC analysis of the different models

From: A novel radiomics based on multi-parametric magnetic resonance imaging for predicting Ki-67 expression in rectal cancer: a multicenter study

Cohort

Model

Accuracy

AUC(95%CI)

Sensitivity

Specificity

PPV

NPV

Training

Radscore

0.81

0.83 (0.747–0.903)

0.83

0.76

0.90

0.62

Clinical

0.74

0.75 (0.646–0.853)

0.76

0.68

0.87

0.51

Combined

0.84

0.86 (0.794–0.928)

0.84

0.73

0.90

0.61

In-valid

Radscore

0.83

0.85 (0.751–0.942)

0.85

0.79

0.93

0.61

Clinical

0.80

0.81 (0.702–0.924)

0.80

0.79

0.93

0.55

Combined

0.88

0.90 (0.826–0.976)

0.87

0.93

0.98

0.68

Ex-valid

Radscore

0.78

0.80 (0.673–0.919)

0.78

0.79

0.92

0.52

Clinical

0.72

0.72 (0.533–0.911)

0.74

0.64

0.87

0.43

Combined

0.85

0.86 (0.749–0.971)

0.85

0.86

0.95

0.63