Table 2.

Model performance of Jin’s model.

Variables		Accuracy	Sensitivity	Specificity	PPV	NPV	PLR	NLR	F1-score
Montreal classification
E2 vs. E1	Internal validation	0.65	0.58	0.81	0.86	0.48	2.98	0.52	0.83
	External validation	0.68	0.56	0.85	0.85	0.57	3.77	0.52	0.85
E3 vs. E1	Internal validation	0.79	0.83	0.64	0.89	0.51	2.30	0.27	0.74
	External validation	0.74	0.76	0.69	0.89	0.46	2.43	0.35	0.77
Severity
Mayo score*	Internal validation	0.70	0.55	0.77	0.55	0.78	2.48	0.58	0.65
	External validation	0.71	0.57	0.79	0.57	0.78	2.63	0.55	0.66
TWS*	Internal validation	0.65	0.46	0.74	0.48	0.73	1.80	0.73	0.58
	External validation	0.73	0.59	0.79	0.59	0.79	2.87	0.52	0.68
MES	Internal validation	0.61	0.66	0.52	0.76	0.39	1.37	0.66	0.62
	External validation	0.67	0.69	0.50	0.94	0.13	1.38	0.62	0.65
DUBLIN score	Internal validation	0.65	0.71	0.55	0.73	0.53	1.58	0.52	0.63
	External validation	0.68	0.67	0.76	0.95	0.25	2.80	0.43	0.84

E2 vs. E1 presents distinguish left-sided from proctitis, E3 vs. E1 presents distinguish extensive from proctitis.

DUBLIN degree of ulcerative colitis burden of luminal inflammation, MES Mayo endoscopic score, NLR negative likelihood ratio, NPV negative predictive value, PLR positive likelihood ratio, PPV positive predictive value, TWS Truelove and Witts score.

*The multi-categorical models were used micro-average.