A Grading Scale for Pial Collaterals in Middle Cerebral Artery Total Occlusion Based on Time-of-flight MR Angiography Source Images

Abstract

PURPOSE: To verify whether a new grading based on time-of-flight magnetic resonance angiography source images (TOF-MRAsi) can reflect the abundance of pial collaterals, in patients with total occlusion of M1 segment of middle cerebral artery in the chronic stage. METHODS: In this single-center retrospective study, consecutive patients with total occlusion of M1 segment of middle cerebral artery, with both magnetic resonances angiography and digital subtraction angiography image were included. Time-of-flight magnetic resonance angiography source images were evaluated in a blinded fashion for pial collaterals (PCs) that were graded on a four-point scale. Good and poor PCs were defined as TOF-MRAsis grade <2 and ≥2, respectively. Receiver operating characteristic curve analysis was done to calculate the area under curve, sensitivity, and specificity. RESULTS: A total of 26 patients were included. The inter-reader agreement for time TOF-MRAsi and digital subtraction angiography images were 0.930 and 0.843, respectively. Compared with digital subtraction angiography grading, the area under curve of pial collateral grading based on TOF-MRAsi was 0.830 (0.636-1.000; P = 0.006). The sensitivity and specificity were 0.700 and 0.933, respectively. The modified Rankin Scale at follow-up was lower in patients with good PCs than in those with poor PCs (0[0, 1] vs. 1[1, 3], P = 0.055), although statistical significance was not reached. CONCLUSION: The grading scale based on TOF-MRAsi could be a new empirical approach for pial collateral evaluation. The clinical use of the proposed approach for identifying patients with total occlusion of middle cerebral artery with a high risk of poor outcome requires evaluation in further studies.