Mechanical Thrombectomy for M2 versus M1 Occlusions: A Propensity-Matched Analysis With Severity-Stratified Assessment of Early Neurological Improvement.

BACKGROUND AND PURPOSE: Recent randomized controlled trials of endovascular therapy (EVT) for medium-vessel occlusions yielded neutral results, partly attributed to enrollment of patients with mild deficits and exclusive reliance on the 90-day mRS. Using propensity score matching (PSM), we compared mechanical thrombectomy (MT) outcomes between M1 and M2 occlusions.

METHODS: This retrospective cohort study used data from the Top-Gear imaging repository. Patients with isolated M1 or M2 occlusions who underwent EVT were included. PSM was performed on age, sex, pre-stroke mRS, admission NIHSS, IV tPA administration, and number of recanalization passes (>3). Primary outcome was 90-day mRS ≤2. Secondary outcomes included ENI and 90-day mRS distribution.

RESULTS: Of 419 screened patients, 171 met eligibility criteria (M1: n=98, M2: n=73). After PSM (n=52 per group), M2 occlusions were associated with significantly higher rates of good outcome (51.9% vs. 30.8%, p=0.04) and lower mortality (19.3% vs. 38.5%, p=0.05). In multivariable logistic regression, M2 occlusion was independently associated with favorable outcome (OR 2.32, 95% CI 1.01-5.3, p=0.04). Stratified analysis by baseline NIHSS after PSM showed that ENI was uncommon in mild strokes (NIHSS ≤5) but occurred in 75% and 65.6% of patients with moderate and severe M2 strokes, respectively. Among M2 patients with ENI, 64.5% achieved mRS ≤2 at 90 days, compared to 33.3% of M1 patients with ENI.

CONCLUSIONS: The M2 occlusion location was independently associated with favorable outcome following MT, with a particularly pronounced benefit among patients with moderate-to-severe baseline deficits. ENI was associated with good functional outcome in M2 occlusions, not in M1 occlusions. These findings suggest that the benefit of MT in M2 occlusions may be concentrated in patients with NIHSS ≥6, a population potentially underrepresented in recent RCTs.

© 2026 by American Journal of Neuroradiology.

PMID: 42276759