Digital image analysis and machine learning-assisted prediction of neoadjuvant chemotherapy response in triple-negative breast cancer

Table 3 NAC prediction performance with Galway cohort

	Accuracy	AUC	Sensitivity	Specificity	Precision	F1
1NN	0.684 [0.645, 0.717]	0.671 [0.524, 0.824]	0.585 [0.522, 0.648]	0.789 [0.765, 0.814]	0.75 [0.712, 0.788]	0.653 [0.619, 0.696]
Linear SVM	0.608 [0.572, 0.643]	0.617 [0.574, 0.660]	0.488 [0.428, 0.548]	0.737 [0.707, 0.766]	0.667 [0.643, 0.690]	0.563 [0.542, 0.585]
RBF SVM	0.810 [0.783, 0.837]	0.832 [0.792, 0.873]	0.805 [0.749, 0.860]	0.816 [0.779, 0.853]	0.825 [0.787, 0.863]	0.815 [0.786, 0.844]
EnsembleTree RUSBoost	0.671 [0.643, 0.699]	0.691 [0.654, 0.727]	0.683 [0.625, 0.741]	0.658 [0.622, 0.694]	0.683 [0.645, 0.721]	0.683 [0.652, 0.714]

The average and 95% confidence interval of accuracy, AUC, sensitivity, specificity, precision, and F1-Score are presented for each model by a leave-one-out cross-validation strategy

ISSN: 1465-542X