Reperfusion failure after successful thrombectomy of large vessel occlusion stroke: clinical and imaging evidence

Introduction: Reperfusion failure (RF) describes a condition in which patients suffering a large vessel occlusion (LVO) stroke present insufficient tissue reperfusion and recovery despite optimal mechanical thrombectomy (MT) results. Approximately 50% of patients suffering from LVO are affected. Our current understanding of the underlying pathomechanisms is limited and mostly based on rodent models. The goal of this study was to further characterize RF by applying advanced multimodal hemodynamic imaging in stroke patients. Methods: Patients from the IMPreST study with LVO stroke and successful recanalization corresponding to thrombolysis in cerebral ischemia grade (TICI) 2b-3 were included. Follow-ups with blood oxygenation-level dependent cerebrovascular reactivity (BOLD-CVR) and non-invasive optimal vessel analysis (NOVA) imaging were performed (<72 h, 7 days and 90 days). Demographic and clinical data (NIHSS and mRS) were collected. Results: Of the 49 patients included in IMPreST, 18 patients met the inclusion criteria. Based on the perfusion weighted imaging (PWI) of the affected area compared to the contralateral side after MT, patients were stratified into three groups: hypoperfusion (n = 3), normalization (n = 8), and hyperperfusion (n = 7). The hyperperfusion group tended to show poorest clinical outcome (mRS 3 months: 2.5 Q1–Q3 2.0–3.0 vs. normalization: 1 Q1–Q3 0.75–3.0, p = 0.169) and had significantly lower BOLD-CVR values at visit one and two compared to hypoperfusion and normalization groups, indicating impaired cerebrovascular reactivity (visit1 hyperperfusion group −0.01 Q1–Q3–0.02 – 0.07, normalization group 0.12 0.09, 0.19, hypoperfusion group, 0.09 0.09, 0.11 p = 0.049, visit2 hyperperfusion group 0.07 Q1–Q3 0.03–0.10, normalization group 0.17 0.16, 0.18, hypoperfusion group 0.10 0.09, 0.11, p = 0.014). Discussion: We found three patterns of reperfusion after successful MT of LVO stroke: normalization, hypo- and hyperperfusion of the ischemic area on days at < 72 h after stroke. There was substantial inhomogeneity in perfusion and clinical outcomes between the three groups. Next to poorest clinical outcome, the hyperperfusion-group showed poorest cerebrovascular reserve, reflecting findings of RF in rodent models. Thus, we suggest that RF includes both hypo- as well as hyperperfusion. Early detection using advanced imaging would allow a better identification of patients at risk for poor clinical outcome. Clinical trial registration: http://clinicaltrials.gov, Identifier (NCT04035746).