Skip to main content

Table 2 Prediction performance of three imaging radiomics models in the training and validation sets

From: Development and validation of ultrasound-based radiomics model to predict germline BRCA mutations in patients with breast cancer

Model

Accuracy (%)

Sensitivity (%)

Specificity (%)

AUC (95% CI)

Training set

 Intratumor

66.4

77.4

62.9

0.754 (0.695, 0.812)

 Combined 3 mm

67.2

77.4

64.0

0.783 (0.727, 0.839)

 Combined 5 mm

66.7

77.4

63.3

0.779 (0.723, 0.835)

Validation set

 Intratumor

57.0

72.2

52.2

0.718 (0.619, 0.817)

 Combined 3 mm

65.1

80.6

60.2

0.782 (0.702, 0.862)

 Combined 5 mm

67.1

75.0

64.6

0.745 (0.650, 0.840)

  1. AUC Area under the receiver operating characteristic curve, CI Confidence interval, combined 3 mm, 3 mm intratumor and peritumor region; combined 5 mm, 5 mm intratumor and peritumor region
Back to article page

Cancer Imaging

ISSN: 1470-7330

Contact us