Apparent Diffusion Coefficient Fraction Predicts Endovascular Thrombectomy Benefit in Large Ischemic Core

BACKGROUND: Endovascular thrombectomy (EVT) is recommended for large ischemic core stroke, but which patients within this population derive meaningful benefit remains uncertain. We investigated whether the fraction of apparent diffusion coefficient (ADC)–defined infarct volume within the ischemic lesion predicts EVT outcomes in large ischemic core stroke. METHODS: This retrospective study included patients with diffusion-weighted imaging volume ≥50 cm 3 and large vessel occlusion within 24-hour of onset (2016–2024; EVT:130, no-EVT:192). The ADC fraction was defined as the volume of ADC <540 × 10 −6 mm 2 /s divided by initial diffusion-weighted imaging volume. Overlap weighting (OW) with doubly robust estimation corrected for indication bias. Primary outcomes were modified Rankin Scale shift and favorable outcome (modified Rankin Scale score 0–3) at 3-month. The relationship between ADC fraction and EVT treatment effect was examined using overlap-weighted restricted cubic spline interaction analysis, and treatment effects were subsequently quantified above and below the crossover point. RESULTS: A total of 322 patients were included (age, 73.8 years, 51.9% men). Before OW, patients with EVT were younger with lower National Institutes of Health Stroke Scale score and smaller diffusion-weighted imaging volumes (EVT: 93.8 cm 3 , no-EVT: 129.9 cm 3 ). After OW, all covariates were well-balanced. The doubly robust–OW analysis demonstrated attenuation of EVT effect on ordinal modified Rankin Scale score shift (odds ratio, 1.37 [95% CI, 0.85–2.21]) and favorable outcome (odds ratio, 1.73 [95% CI, 0.91–3.28]). ADC fraction significantly modified the EVT treatment effect ( P interaction =0.009), with a crossover at 48%, identified as a data-derived exploratory threshold. Below this threshold, EVT was associated with a favorable ordinal modified Rankin Scale score shift (odds ratio, 1.85 [95% CI, 1.05–3.26]), whereas above it the direction of the association was reversed (odds ratio, 0.42 [95% CI, 0.19–0.95]). CONCLUSIONS: In large ischemic core stroke, the association between EVT and functional outcome varied according to infarct tissue composition. The ADC fraction distinguished patients in whom EVT was associated with more favorable outcomes from those in whom this association was not observed, supporting its potential as an imaging biomarker to inform individualized treatment decisions.