LAUSR

To the Editor: I read with great interest the article by Kim et al published in a recent issue of Stroke. The authors aimed to elucidate whether magnetic resonance imaging (MRI)–based selection for endovascular therapy is safe and effective within and after a 6-hour time window compared with conventional computed tomography angiography (CTA)–based selection in patients with anterior circulation ischemic stroke presenting with large vessel occlusion. They analyzed 1265 patients from a prospective, nationwide, multicenter stroke registry. Multivariable analysis showed that imaging modality was not significantly associated with 3-month functional outcomes (all P>0.1) or mortality (P=0.051); however, the MRI group was less likely to develop symptomatic intracranial hemorrhage than the CTA group (odds ratio, 0.34; 95% CI, 0.17–0.77; P=0.01). In this study, the time from arrival to decision imaging and time from arrival to puncture were delayed by ≈28 minutes and 21 minutes, respectively, in the MRI group compared with the CTA group. Although MRI-based selection led to substantially delay, the MRI group was less likely to develop symptomatic intracranial hemorrhage (P=0.01) and was associated with a trend of lower mortality (P=0.051) than the CTA group. These suggest that MRI-based selection may be associated with a significant better 3-month functional outcome if the delayed time was decreased substantially. There was higher-level evidence to support this speculation. In a recently published individual patient-level meta-analysis, Campbell et al assessed endovascular thrombectomy predominantly performed with stent retrievers versus medical therapy in patients with anterior circulation ischemic stroke by analyzing randomized controlled trials published between January 2010 and May 2017. Imaging was done within 6 hours of stroke onset in 887 (99%) of 900 patients with CT perfusion (CTP) or diffusion MRI. When comparing imaging modalities, diffusion MRI was independently associated with 3-month functional independence (adjusted odds ratio, 2.13; 95% CI, 1.39–3.33; P=0.0007) than CTP. The time from arrival to arterial access was delayed by only ≈4 minutes in the diffusion MRI group compared with the CTP group. These results suggested that diffusion MRI-based selection was associated with better functional outcome compared with CTA/CTP-based selection under the similar arrival to puncture time. However, the flaws of observational study and post hoc analysis cannot make the difference a solid conclusion. But if this imaging modality difference was real, the difference may occur because of underestimation of infarct volume by CTP and overestimation by diffusion MRI. On the contrary, the absence of interaction between imaging modality and treatment effect (P=0.86) suggests that the prognostic influence per milliliter increase in ischemic core was similar between CTP group and diffusion MRI group. Therefore, although with the possible difference, CTP and diffusion MRI can all be used to inform endovascular thrombectomy treatment decisions. In conclusion, under the current evidence, CTA/CTP and diffusion MRI can all be used to inform endovascular thrombectomy treatment decisions by combining with other factors. The possible better 3-month functional outcome based on MRI approach may be caused by more selective with MRI and overestimation of infarct volume by MRI. The selection of imaging modality should be mainly decided according to the workflow features in each center.