Table 4.

Performance achieved by the best classifiers to discriminate between benign and malignant lesions for combined CEM and DCE-MRI.

Classifier	Cross-Validation	ACC	SENS	SPEC	PPV	NPV	AUC
Performance for classifiers trained with balanced data (with ADASYN function) and a subset of 18 robust textural features (by LASSO and λminMSE).
LDA	10-fold CV	0.84	0.73	0.92	0.90	0.79	0.88
	LOOCV	0.80	0.71	0.88	0.85	0.77	0.87
SVM	10-fold CV	0.84	0.81	0.87	0.85	0.83	0.86
	LOOCV	0.83	0.79	0.87	0.84	0.82	0.86
Performance for classifiers trained with balanced data (with SASYNO function) and a subset of 3 robust textural features (by LASSO and λminMSE).
LDA	10-fold CV	0.79	0.79	0.79	0.79	0.79	0.88
	LOOCV	0.79	0.79	0.79	0.79	0.79	0.89
SVM	10-fold CV	0.80	0.79	0.79	0.79	0.79	0.86
	LOOCV	0.79	0.77	0.81	0.80	0.78	0.87