OBJECTIVE The authors compared the image quality and diagnostic sensitivity and specificity of 7.0-T and 3.0-T MRI and time-of-flight (TOF) MR angiography (MRA) in patients with moyamoya disease (MMD). METHODS MR images of 15 patients with ischemic-type MMD (8 males, 7 females; age 13-48 years) and 13 healthy controls (7 males, 6 females; age 19-28 years) who underwent both 7.0-T and 3.0-T MRI and MRA were studied retrospectively. The main intracranial arteries were assessed by using the modified Houkin's grading system (MRA score). Moyamoya vessels (MMVs) were evaluated by 2 grading systems: the MMV quality score and the MMV area score. Two diagnostic criteria for MMD were used: the T2 criteria, which used flow voids in the basal ganglion on T2-weighted images, and the TOF criteria, which used the high-intensity areas in the basal ganglion on source images from TOF MRA. All data were evaluated by 2 independent readers who were blinded to the strength field and presence or absence of MMD. Using conventional angiography as the gold standard, the sensitivity and specificity of 7.0-T and 3.0-T MRI/MRA in the diagnosis of MMD were calculated. The differences between 7.0-T and 3.0-T MRI and MRA were statistically compared. RESULTS No significant differences were observed between 7.0-T and 3.0-T MRA in MRA score (p = 0.317) or MRA grade (p = 0.317). There was a strong correlation between the Suzuki's stage and MRA grade in both 3.0-T (r(s) = 0.930; p < 0.001) and 7.0-T (r(s) = 0.966; p < 0.001) MRA. However, MMVs were visualized significantly better on 7.0-T than on 3.0-T MRA, suggested by both the MMV quality score (p = 0.001) and the MMV area score (p = 0.001). The correlation between the Suzuki's stage and the MMV area score was moderate in 3.0-T MRA (r(s) = 0.738; p = 0.002) and strong in 7.0-T MRA (r(s) = 0.908; p < 0.001). Moreover, 7.0-T MR images showed a greater capacity for detecting flow voids in the basal ganglion on both T2-weighted MR images (p < 0.001) and TOF source images (p < 0.001); 7.0-T MRA also revealed the subbranches of superficial temporal arteries much better. Receiver operating characteristic curve analysis showed that, according to the T2 criteria, 7.0-T MRI/MRA was more sensitive (sensitivity 1.000; specificity 0.933) than 3.0-T MRI/MRA (sensitivity 0.692; specificity 0.933) in diagnosing MMD; based on the TOF criteria, 7.0-T MRI/MRA was more sensitive (1.000 vs 0.733, respectively) and more specific (1.000 vs 0.923, respectively) than 3.0-T MRI/MRA. CONCLUSIONS Compared with 3.0-T MRI/MRA, 7.0-T MRI/MRA detected and delineated MMVs more clearly and provided higher diagnostic sensitivity and specificity, although it did not show significant improvement in depicting main intracranial arteries. The authors speculate that 7.0-T MRI/MRA is a promising technique in the diagnosis of MMD because it is noninvasive compared with conventional angiography and it is more sensitive than 3.0-T MRI/MRA.