Validation and Application for the Berlin Grading System of Moyamoya Disease in Adult Patients

BACKGROUND Traditional moyamoya disease (MMD) classification relies on morphological digital subtraction angiography (DSA) assessment, which do not reflect hemodynamic status, clinical symptoms, or surgical treatment outcome. OBJECTIVE To (1) validate the new Berlin MMD preoperative symptomatology grading system and (2) determine the clinical application of the grading system in predicting radiological and clinical outcomes after surgical revascularization. METHODS Ninety-six MMD patients (192 hemispheres) with all 3 investigations (DSA, magnetic resonance imaging [MRI], Xenon-CT) performed preoperatively at our institution (2007-2013) were included. Two clinicians independently graded the imaging findings according to the proposed criteria. Patients' modified Rankin Score (mRS) scores (preoperative, postoperative, last follow-up), postoperative infarct (radiological, clinical) were collected and statistical correlations performed. RESULTS One hundred fifty-seven direct superficial temporal artery-middle cerebral artery bypasses were performed on 96 patients (66 female, mean age 41 yr, mean follow-up 4.3 yr). DSA, MRI, and cerebrovascular reserve capacity were independent factors associated hemispheric symptomatology (when analyzed individually or in the combined grading system). Mild (grade I), moderate (grade II), severe (grade III) were graded in 45, 71, and 76 hemispheres respectively; of which, clinical symptoms were found in 33% of grade I, 92% of grade II, 100% of grade III hemispheres (P < .0001). Two percent of grade I, 11% of grade II, 20% of grade III hemispheres showed postoperative radiological diffusion weighted image-positive ischemic changes or hemorrhage on MRI (P = .018). Clinical postoperative stroke was observed in 1.4% of grade II, 6.6% of grade III hemispheres (P = .077). The grading system also correlated well to dichotomized mRS postoperative outcome. CONCLUSION The Berlin MMD grading system is able to stratify preoperative hemispheric symptomatology. Furthermore, it correlated with postoperative new ischemic changes on MRI, and showed a strong trend in predicting clinical postoperative stroke.