Purpose: To prospectively evaluate, with magnetic resonance (MR) imaging, long-term outcome of the brain after endovascular versus neurosurgical treatment for aneurysmal subarachnoid hemorrhage (aSAH). Materials and Methods: Institutional review board approval and informed con sent were obtained. One hundred sixty-eight (77 men, 91 women; mean age standard deviation, 51 years 13) patients were randomly assigned to surgical versus endovascular treatment of the ruptured aneurysm with 138 (67 endovascular, 71 surgical) MR examinations I after aSAH. The presence, localization, volumes, year and cause of lesions were analyzed with chi(2), Mann-Whitney U, and Student t tests. Furthermore, correlation between MR-detectable brain parenchymal high-signal intensity (SI) lesions on T2- and intermediate-weighted A MR images and neuropsychologic outcome was evaluated by using Spearman correlation coefficient. Results: Only 44 (31.9%) of 138 patients had no lesions associated with aSNH. According to intention to treat, lesions were more frequent after surgical rather than endovascular treatment, predominating in the frontal (surgical: n = 50, [70.4%] vs endovascular: n = 34 [50.7%], P =.018) and temporal (n = 34 [47.9%] vs n = 15 [22.4%], P =.002) lobes. Only endovascular patients had subtentorial lesions (n = 4 [6.0%], P = .037). Ischemic lesions in the parental artery territory were more frequent in surgical (n = 33 [46.5%]) than in endovascular (n = 15 [22.4%], P = .003) with corresponding mean lesion volumes of 20.9 cm(3) +/- 46.5 versus 17.6 cm(3) +/- 35.8 (P = .209). Ischemic lesions in remote vascular territories were equal in frequency and size. Retraction injuries were common in the surgical (n = 40, [56.3%]) treatment group. Ischemic lesion volumes correlated with neuropsychologic test scores. Conclusion: Parenchymal high-SI lesions on T2- and intermediate-weighted MR images are rnore frequent after early surgical rather than endovascular treatment of the ruptured aneurysm, lesion volumes correlate with the neuropsychologic test performance. (C) RSNA, 2007.