Table 1.

Feeding artery detection by both observers using 3DCT and DSA images compared to Ground truth

		Ground truth
		Positive (n = 173)	Negative (n = 12)
3D (observer-1/ Observer-2)	Positive	159/157	4/3
	Negative	14/16	8/9
DSA (observer-1/ Observer-2)	Positive	139 /136	5/5
	Negative	34/37	7/7
	3D (observer-1/observer-2)	DSA (observer-1/observer-2)
Sensitivity	91.9% / 90.8%	80.3% / 78.6%
Specificity	66.7% / 75%	58.3% / 58.3%
PPV	97.5% / 98.1%	96.5% / 96.5%
NPV	36.4% / 36%	17.1% / 15.9%
Accuracy	90.3% / 89.7%	78.9% / 77.3%