Table 3.

Predictive model performance—Performance metrics for each of the predictive models trained. For accuracy, sensitivity and specificity, values presented are percentages. * indicates the model(s) with the best performance for each given metric. For all metrics, except AIC, higher values indicate better performance. AUC = Area under the Receiver-Operator Characteristic curve, AIC = Akaike Information Criterion.

	Accuracy	F1-Score	AUC	Sensitivity	Specificity	AIC
Basic Model	63.16	0.49	0.45	0.00	98.82 *	176.94
Clinical model	78.95 *	0.78 *	0.79	62.50	88.24	146.96
Imaging model	72.18	0.71	0.74	47.92	85.88	148.43
Robotic model	77.44	0.77	0.84	68.75*	82.35	120.07 *
Augmented model	76.69	0.77	0.86*	64.58	83.53	126.07