Stroke progression index as a dynamic metric is associated with functional outcome in progressive ischemic stroke...

Current evaluation of progressive ischemic stroke (PIS) with large vessel occlusion (LVO) relies on static clinical or imaging measures and lacks quantitative assessment of dynamic neurological deterioration. We proposed a novel metric, stroke progression index (SPI), to assess its association with clinical outcomes and benefit from endovascular treatment (EVT), and to explore its potential clinical relevance.

We retrospectively enrolled consecutive LVO-PIS patients treated at our center between January 2022 and June 2025, limited to those with large-artery atherosclerosis. SPI was derived from National Institutes of Health Stroke Scale (NIHSS) change and progression time using a mathematically adjusted ratio. The primary endpoint was poor outcome at 90 days (mRS 3-6). We analyzed the association between SPI and outcomes, evaluated the effect of EVT, and explored relationships between SPI and disease course-related factors.

Among 190 patients, 44.2% (84/190) had a poor 90-day outcome, and 63.2% (120/190) received EVT. Regression analysis identified SPI as an independent risk factor for poor outcome (OR 1.223, 95% CI 1.100-1.360; p < 0.001). Adding SPI to a baseline model with traditional clinical variables improved discrimination (AUC 0.654 to 0.757; DeLong p < 0.001). In the EVT subgroup, SPI remained an independent risk factor (OR 1.134; 95% CI 1.018-1.263; p = 0.023). In treatment-effect analyses, lower SPI was associated with greater absolute benefit from EVT. Absolute risk reduction (ARR) was approximately 30% at SPI ≈11 and 10% at SPI ≈ 20. Exploratory analyses showed overall higher SPI in early than delayed progression, with poor outcomes being more common in patients with high SPI and poor collateral status.

SPI is an independent prognostic risk factor in LVO-PIS that captures disease dynamics. The benefit of EVT varied across different SPI levels, suggesting that SPI may help inform clinical decision-making. These findings require further validation in larger, multicenter studies.