Table 2:

Results from the Spearman rank correlation analysis for collateral score as a function of perfusion parameters (ρ [95% CI] and P value) and from ROC analysis for the identification of good collateral status^{^a}

Perfusion Parameter	ρ (95% CI)	P Value	AUC (95% CI)	Cutoff Value	Sensitivity	Specificity	Accuracy
Tmax >6 sec (mL)	−0.50 (−0.61 to −0.39)	<.001	0.75 (0.68–0.81)^{^b}	124 mL	56%	82%	69%
Tmax >8 sec (mL)	−0.54 (−0.64 to −0.43)	<.001	0.77 (0.71–0.83)^{^b}	74 mL	68%	76%	72%
Tmax >10 sec (mL)	−0.50 (−0.60 to−0.39)	<.001	0.77 (0.71–0.83)^{^b}	53 mL	64%	80%	72%
rCBF < 30% (mL)	−0.61 (−0.71 to −0.52)	<.001	0.81 (0.75–0.87)	14 mL	72%	82%	77%
rCBF < 34% (mL)	−0.64 (−0.73 to −0–55)	<.001	0.83 (0.77–0.88)	25 mL	67%	87%	77%
rCBF < 38% (mL)	−0.66 (−0.74 to −0.57)	<.001	0.83 (0.78–0.89)	27 mL	75%	80%	77%
CBV-index	+0.51 (0.40–0.63)	<.001	0.76 (0.69–0.81)^{^b}	0.8	60%	78%	69%
HIR	−0.49 (−0.60−0.37)	<.001	0.73 (0.66–0.79)^{^b}	0.4	75%	62%	65%

↵a Overall, the volume with an rCBF < 38% performed best. Optimal cutoff values to identify good collateral supply were estimated according to the Youden index.
↵b Significant differences in AUC compared with the AUC for the volume with rCBF < 38%.