Skip to main content

Table 6 ROC metrics for predicting PCR based on molecular subtypes, MRI features at pre- and during NAC using Ensemble RUSBoosted Tree classifier based on single view multiview techniques with SMOTE

From: Machine learning classification of texture features of MRI breast tumor and peri-tumor of combined pre- and early treatment predicts pathologic complete response

Time point

Features type

Sens.

Spec.

PPV

NPV

Accuracy

AUC

(View 1)

Molecular subtypes

65.0

69.0

85.70

40.90

66.0

0.69 (0.54, 0.90)

Tp1 (view 2)

MRI texture only

68.0

100

100

52.0

76.0

0.86 (0.77, 0.96)

Tp2 (view 3)

70.0

62.0

83.90

42.10

68.0

0.76 (0.58, 0.94)

Tp1 + Tp2 (View 4)

73.0

92.0

96.40

54.50

78.0

0.88 (0.78, 0.97)

Multiview

Molecular subtype and MRI texture

84.0

100

100

68.40

88.0

0.98 (0.94, 1.00)

  1. The data were tumor contours based on post-contrast-enhanced MRI with morphological dilation. The numbers in parenthesis show the 95% confidence intervals