Skip to main content

Table 2 The fivefold cross validation results of the proposed network

From: Automatic identification of suspicious bone metastatic lesions in bone scintigraphy using convolutional neural network

  Fold 1 Fold 2 Fold 3 Fold 4 Fold 5
Number of patients (benign/malignant) 669 (376/293) 669 (376/293) 671 (376/295) 669 (376/293) 674 (378/296)
Number of lesions (benign/malignant) 2986 (1505/1481) 2952 (1389/1563) 3084 (1468/1616) 2897 (1475/1422) 3053 (1465/1588)
Sensitivity 83.19 83.11 79.89 76.79 83.5
Specificity 78.07 82.58 82.7 82.78 79.59
Accuracy 80.61 82.86 81.23 79.84 81.62
PPV 78.87 84.3 83.56 81.13 81.6
NPV 82.51 81.29 78.88 78.72 81.65