Abstract

Introduction: Indices of collateral flow deficit derived from MR perfusion imaging that are predictive of MCA-M1 recanalization after intravenous thrombolysis have been recently reported. Our objective was to test the performance of such MRI-derived collateral flow indices for prediction of recanalization after endovascular thrombectomy. Methods: Fifty-seven patients with MCA-M1 occlusion evaluated with multimodal MRI prior to thrombectomy were included. Bayesian processing allowed quantification of collateral perfusion indices like the volume of tissue with severely prolonged arterial-tissue delay. Baseline DWI lesion volume was also measured. Correlations with angiographic collateral flow grading and post-thrombectomy recanalization were assessed. Results: VolATD6 27 ml and DWI lesion volume >15 ml significantly discriminates recanalizers versus nonrecanalizers, p = 0.024). When both criteria are positive, 76.2 % of the patients treated with the MERCI retriever do not fully recanalize. In multivariate analysis, the aforementioned combined criterion and the angiographic collateral grade are the only independent predictors of recanalization with the MERCI retriever. Conclusion: Bayesian arterial-tissue delay maps and DWI maps provide a non-invasive assessment of the degree of collateral flow and a combined index that is predictive of MCA-M1 recanalization after endovascular thrombectomy. Further studies are needed to evaluate the accuracy of this index in patients treated with novel stent retriever devices. © 2013 The Author.