Skip to main content

Table 1 Evaluation measures for a local association of anomaly score and future lesion occurrence

From: Deep learning for predicting future lesion emergence in high-risk breast MRI screening: a feasibility study

ROC curve number

AUC

Classification threshold based on ROC curve

Sensitivity

Specificity

False positive rate

1

0.81

585.14

70%

85%

0.15

2

0.80

552.43

71%

83%

0.17

3

0.82

585.14

70%

84%

0.16

4

0.81

572.91

70%

86%

0.14

5

0.81

559.86

71%

84%

0.16

6

0.82

595.71

69%

86%

0.14

7

0.81

585.14

70%

84%

0.16

8

0.80

595.71

69%

85%

0.15

9

0.80

559.86

71%

82%

0.18

10

0.82

552.43

71%

84%

0.16

  1. ROC Receiver operating characteristic, AUC Area under the curve