Abstract A086: Classification of territorial reperfusion pattern of the middle cerebral artery after mechanical thrombectomy: a latent class analysis

Background: The Thrombolysis in Cerebral Infarction scale is the most widely used scale to evaluate reperfusion after mechanical thrombectomy (MT), defined according to the percentage of reperfusion in downstream arterial branches. A limitation of this volume-based score is that it does not account for the varying eloquences of different arterial territories. Understanding territorial reperfusion patterns may advance the knowledge of incomplete reperfusion and more granular outcome prediction after MT. We therefore aimed to classify the territorial reperfusion patterns of cortical branches of the middle cerebral artery (MCA) after MT and to evaluate their impact on clinical outcomes. Methods: This single-center observational study included patients who underwent MT for internal carotid artery (ICA) or MCA M1 occlusion between January 2014 and May 2025. Patients with concomitant anterior cerebral artery occlusion were excluded. Reperfusion status of the twelve MCA branches was evaluated on final angiography and subsequently subjected to latent class analysis (LCA) to identify distinct reperfusion patterns. The primary outcome was achievement of modified Rankin Scale 0–2 or returning to baseline at 90 days. Multivariable logistic regression was performed to compare outcomes across classes. Results: A total of 473 patients (median age 78 [IQR 70–85] years; 222 female [47%]) were analyzed. The LCA model identified five classes of reperfusion patterns (Figure 1); class 1: good reperfusion in all territory (n=326), class 2: poor reperfusion in the parietal and the occipital lobe (n=41), class 3: poor reperfusion in the frontal lobe (n=52), class 4: reperfusion limited to the temporal lobe (n=13), class 5: poor reperfusion in all territory (n=45). In multivariable analysis, the likelihood of achieving the primary outcome was significantly higher in class 1 (61.0%, adjusted odds ratio [aOR] 15.1, 95% confidence interval [CI] 5.9–46.4), class 2 (56.1%, aOR 9.9, 95% CI 3.2–35.5), class 3 (53.8%, aOR 9.9, 95% CI 3.3–33.9), but not in class 4 (30.8%, aOR 4.4 95% CI 0.8–22.0) compared with class 5 (14.6%) (Figure 2,3). Conclusion: LCA identified five territorial reperfusion patterns of MCA cortical branches after MT. These patterns may provide insights into the characterization of incomplete reperfusion, but further studies are needed for validation.