Table 2.

Individual subset predictive value: ROC analysis and cut-off for patients’ stratification as high-risk of progression.

		Normalised Naive	Normalised Treg	IRC	Risk-Score
At-risk participants n = 158	AUC (range)	0.313 (0.212–0.413)	0.330 (0.227–0.432)	0.313 (0.212–0.413)	3 subsets
	P	0.001	0.003	0.001	02-Mar
	Cut-off	−10%	−3%	4%	93%
	Specificity	80%	82%	80%	36.50%
	Sensitivity	41%	35%	41%	4.95
	OR	1.95	2.3	1.95	74%
	PPV	61%	64%	61%	71.50%
	NPV	63%	63%	63%
Evolving IA n = 294	AUC (range)	0.363 (0.281–0.446) 0.001	0.389 (0.307–0.471)	NA	2 subsets
	P	−10%	0.01		Absent/Present
	Cut-off	92%	−3%		80%
	Specificity	51%	83%		69%
	Sensitivity	6.4	21%		3.38
	OR	84%	1.5		83%
	PPV	69%	80%		64%
	NPV		29.5
MTX-induced Remission n = 120	AUC (range)	0.822 (0.710–0.933)	NA	NA	Naive subset only
	P	<0.0001
	Cut-off	2.50%
	Specificity	81%
	Sensitivity	78%
	OR	15.4
	PPV	78%
	NPV	81%
Remission/ Flare n = 145	AUC (range)	0.271 (0.151–0.389) <0.0001	NA	NA	Naive subset only
	P	−2.50%
	Cut-off	77%
	Specificity	62%
	Sensitivity	2.7
	OR	43%
	PPV	88%
	NPV

AUC: area under the curve (95% CI), OR: odd ratio, PPV: positive predictive value, NPV: negative predictive value, NA not applicable.