Table 2. Accuracy, AUROC, AUPRC, sensitivity, specificity, PPVs, NPVs, and F1 scores for the ensemble model in the internal validation dataset, the external validation dataset and the cardiac subgroups of the internal validation and the external validation datasets.

Population	Dataset	Accuracy (95% CI)	AUROC (95% CI)	AUPRC (95% CI)	Sensitivity (95% CI)	Specificity (95% CI)	PPV (95% CI)	NPV (95% CI)	F1 score
Original analysis	Internal validation	0.9620 (0.9352–0.9889)	0.9800 (0.9612–0.9988)	0.9950 (0.9860–1.0000)	0.9594 (0.9245–0.9943)	0.9714 (0.9162–1.0000)	0.9916 (0.9752–1.0000)	0.8718 (0.7669–0.9767)	0.9752
	External validation	0.8509 (0.7825–0.9192)	0.9301 (0.8845–0.9756)	0.9476 (0.9087–0.9867)	0.9595 (0.9145–1.0000)	0.6500 (0.5022–0.7978)	0.8353 (0.7564–0.9142)	0.8966 (0.7857–1.0000)	0.8931
Cardiac subgroup analysis	Internal validation	0.9661 (0.9191–1.0000)	0.9954 (0.9781–1.0000)	0.9959 (0.9797–1.0000)	1.0000 (1.0000–1.0000)	0.9286 (0.8332–1.0000)	0.9394 (0.8580–1.0000)	1.0000 (1.0000–1.0000)	0.9688
	External validation	0.6600 (0.5021–0.8179)	0.8917 (0.7906–0.9928)	0.8968 (0.7980–0.9956)	0.9000 (0.7685–1.0000)	0.5000 (0.3211–0.6789)	0.5455 (0.3756–0.7153)	0.8824 (0.7292–1.0000)	0.6792

The cutoff probability scores for unfavorable neurological outcomes were set at 0.605 and 0.525 for the ensemble models in the original and the cardiac subgroup analyses, respectively.

AUROC = area under the receiver operating characteristic curve, AUPRC = area under the precision-recall curve, PPV = positive predictive value, NPV = negative predictive value, CI = confidence interval.