A pharmacogenetic signature of high response to Copaxone in late-phase clinical-trial cohorts of multiple sclerosis

Table 5 Model performance summary on all the cohorts

Cohort	Number of patients			Follow-up duration (years)	Specificity	Sensitivity	AUC
Cohort	Total	Sig+	Sig−	Follow-up duration (years)	Specificity	Sensitivity	AUC
Discovery
GALA DB	639	323	316	1	66%	54%	0.65
FORTE DB	532	268	264	1	71%	54%	0.68
GALA DB + FORTE DB	1171	591	580	1	68%	54%	0.66
Independent assessment
GALA OL	333	190	143	~3	47%	58%	0.54
GA-9001 DB	38	21	17	~3	41%	52%	0.45
GA-9001 OL	74	35	39	~20	48%	45%	0.49
GA-9003 DB	40	21	19	0.75	67%	61%	0.65
GA-9003 OL	84	41	43	0.75	67%	54%	0.59
PreCISe DB	132	69	63	3	48%	52%	0.49
PreCISe OL	240	129	111	5	49%	54%	0.50

Sig + (signature-positive) and Sig − (signature-negative) indicate patients classified as relapse-free and relapsing, respectively, after applying the “top-left” threshold on the predicted probabilities from the four-SNP logistic regression model (see “Methods”). AUC is a threshold-independent metric that computes the overall performance of the model at all possible thresholds on the predicted probabilities. All performance metrics are rounded to two decimal places.
DB double-blind phase, OL open-label phase

ISSN: 1756-994X