Artery length sensitivity in patient-specific cerebral aneurysm simulations.

BACKGROUND AND PURPOSE The reconstruction of aneurysm geometry is a main factor affecting the accuracy of hemodynamics simulations in patient-specific aneurysms. We analyzed the effects of the inlet artery length on intra-aneurysmal flow estimations by using 10 ophthalmic aneurysm models. MATERIALS AND METHODS We successively truncated the inlet artery of each model, first at the cavernous segment and second at the clinoid segment. For each aneurysm, we obtained 3 models with different artery lengths: the originally segmented geometry with the longest available inlet from scans and 2 others with successively shorter artery lengths. We analyzed the velocity, wall shear stress, and the oscillatory shear index inside the aneurysm and compared the 2 truncations with the original model. RESULTS We found that eliminating 1 arterial turn resulted in root mean square errors of <18% with no visual differences for the contours of the flow parameters in 8 of the 10 models. In contrast, truncating at the second turn led to root mean square errors between 18% and 32%, with consistently large errors for wall shear stress and the oscillatory shear index in 5 of the 10 models and visual differences for the contours of the flow parameters. For 3 other models, the largest errors were between 43% and 55%, with large visual differences in the contour plots. CONCLUSIONS Excluding 2 arterial turns from the inlet artery length of the ophthalmic aneurysm resulted in large quantitative differences in the calculated velocity, wall shear stress, and oscillatory shear index distributions, which could lead to erroneous conclusions if used clinically.