Table 2. Classification scores for the discovery cohort of the classification tree model.

	Sensitivity	Specificity	PPV	Balanced accuracy	AUROC
INIU	80.39% ± 6.4%	82.86% ± 1.23%	54.41% ± 2.57%	81.4 ± 3.75%	0.839 ± 0.195
AMD	71.28% ± 2.45%	87.48% ± 1.28%	65.57% ± 2.5%	79.4 ± 1.9%	0.863 ± 0.104
RRD	76.9% ± 0%	100% ± 0%	100% ± 0%	88.45 ± 0%	0.887 ± 0.197
PVRL	62.93% ± 20.11%	97.93% ± 1.04%	84.73% ± 7.18%	80.22 ± 10.61%	0.878 ± 0.291
Model	75.15% ± 1.67%	93.09% ± 0.45%	75.13% ± 11.5%	84.12% ± 1.06%	0.87 ± 0.123

INIU, idiopathic non-infectious uveitis; AMD, age-related macular degeneration; RRD, rhegmatogenous retinal detachment; PVRL, primary vitreoretinal lymphoma. PPV, positive predictive value; balanced accuracy, mathematical mean of sensitivity and specificity; AUROC, area under receiver operator characteristics. Values expressed as the mean ± standard deviation of 100 cross-validation runs.