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

Table 3 ROC metrics for predicting PCR using post-contrast image, subtraction image, and subtraction image with dilation at tp1 + tp2 using Ensemble RUSBoosted Tree classifier

Method	Sens.	Spec.	PPV	NPV	Accuracy	AUC	P-value
Post-contrast image	89.18	53.84	84.61	63.64	80	0.68 (0.48, 0.87)	0.212
Subtraction image	89.18	61.54	86.84	66.66	82	0.83 (0.70, 0.97)	0.128
Subtraction + 5 pixel dilation	94.59	92.31	97.22	85.71	94	0.98 (0.94, 1.0)	0.00029

ISSN: 1475-925X