High-resolution magnetic resonance angiography for predicting short-term prognosis in young patients with atherosclerotic stroke: a mediation analysis based on middle cerebral artery plaque characteristics

Abstract Objective This study primarily evaluated the predictive value of middle cerebral artery (MCA) plaque characteristics, assessed by high-resolution magnetic resonance vessel wall imaging (HR-VWI), for 90-day prognosis in young patients with atherosclerotic ischemic stroke. As supportive analyses, we further explored whether low-density lipoprotein cholesterol (LDL-C) and homocysteine (HCY) influence prognosis through mediating pathways involving these plaque characteristics. Methods A total of 526 young patients (aged 18–50 years) with acute ischemic stroke in the MCA territory, admitted to Cangzhou Central Hospital between January 2023 and December 2025, were retrospectively selected. Patients were stratified into a favorable prognosis group [90-day modified Rankin Scale(mRS) score 0–1; n = 380] and an unfavorable prognosis group (mRS score 2–6; n = 146). Baseline clinical data, laboratory indices, and quantitative HR-VWI plaque parameters were collected. Core predictive variables were identified using LASSO regression combined with the Boruta algorithm, followed by multivariable logistic regression to assess independent associations. Dose–response relationships were analyzed using restricted cubic splines (RCS), threshold effects were determined via piecewise regression, and mediation analyses were performed using the bootstrap method. Results Multivariable logistic regression identified that MCA plaque length (adjusted OR = 1.26; 95% CI 1.17–1.37; P < 0.001), MCA stenosis rate (adjusted OR = 1.02; 95% CI 1.01–1.03; P = 0.001), MCA wall area (adjusted OR = 1.20; 95% CI 1.10–1.32; P < 0.001), LDL-C (adjusted OR = 1.30; 95% CI 1.11–1.54; P = 0.002), and HCY (adjusted OR = 1.05; 95% CI 1.01–1.09; P = 0.011) were significantly associated with an increased risk of poor 90-day prognosis. Conversely, MCA lumen area showed a significant negative association (adjusted OR = 0.92; 95% CI 0.84–1.00; P = 0.045). ROC curve analysis showed that individual parameters had limited predictive performance (AUC range: 0.559–0.648). RCS analysis revealed significant dose–response relationships for all aforementioned indicators (overall P < 0.05), with plaque length, stenosis rate, wall area, and lumen area exhibiting nonlinear patterns (nonlinear P < 0.05), whereas LDL-C and HCY showed linear relationships. Threshold effect analysis demonstrated significant inflection points for all HR-VWI parameters (likelihood ratio tests, all P < 0.01), identified at 7.19 mm (plaque length), 5.99 mm 2 (wall area), 63.90% (stenosis rate), and 8.11 mm 2 (lumen area). Mediation analysis showed that LDL-C exerted an indirect effect on prognosis via MCA plaque length (indirect effect: 0.012; 95% CI 0.007–0.019), accounting for 36.4% of the total effect. HCY showed an indirect effect of 0.003 (95% CI 0.0009–0.005) through MCA wall area (mediation proportion: 40.5%). In an exploratory analysis, MCA plaque length, lumen area, degree of stenosis, wall area, LDL-C, and Hcy were included in a combined model, and after bootstrap internal validation with 1000 resamples, the AUC of combined model was 0.724 (95% CI 0.674–0.773). Conclusion HR-VWI-derived parameters—specifically MCA plaque length, wall area, stenosis rate, and lumen area—are independent predictors of 90-day prognosis in young patients with atherosclerotic ischemic stroke and exhibit distinct threshold effects, supporting their potential role as imaging-based risk stratification tools. In supportive analyses, LDL-C and HCY were associated with prognosis partially through plaque length and wall area, respectively; however, these mediation pathways require confirmation in larger prospective studies with adequate statistical power.