Table 4. Performance metric comparison of our combined model and reported performance metrics of the model proposed by [21] using the same selection criteria.

*Question marks indicate values not clearly reported.

	Threshold	False Positive Rate	Sensitivity	Specificity
Yim et al.[21]	High sensitivity	0.434	> = 0.8?*	0.55
	High specificity	0.096	0.34	> = 0.9?*
Combined Model	High sensitivity	0.203 (0.129, 0.3)	0.704 (0.444, 0.948)	0.797 (0.7, 0.871)
	High specificity	0.135 (0.078, 0.198)	0.63 (0.367, 0.875)	0.865 (0.802, 0.922)