Risk stratification of acute symptomatic anterior circulation intracranial atherosclerotic stenosis using an ABCD3-I-informed clinical-imaging score: an exploratory study.

BACKGROUND: Acute symptomatic anterior circulation intracranial atherosclerotic stenosis (ICAS; ≥70%) carries substantial risks of early neurological deterioration (END) and disability. We developed an ABCD3-I-informed clinical-imaging score incorporating pre-baseline progressive stroke, responsible-vessel stenosis, and diffusion-weighted imaging (DWI) infarct pattern for acute risk screening. METHODS: We retrospectively enrolled 133 consecutive patients (103 men, 30 women; mean age: 64.62 ± 10.56 years) with acute symptomatic anterior circulation ICAS (≥70% stenosis; onset ≤7 days) at a single center (January 2024-June 2025). The score (0-13) categorized patients as high risk (≥8; n = 102) or non-high risk (n = 31). The primary outcome was the ordinal 90-day modified Rankin Scale (mRS) score; the secondary outcomes were mRS scores ≥3 and END. Adjusted regression and receiver operating characteristic (ROC) analyses were performed, and discrimination was compared with the original ABCD3-I. RESULTS: The high-risk group had higher admission National Institutes of Health Stroke Scale (NIHSS) scores (median 2.0 vs. 0.0; p < 0.001). The ABCD3-I-informed clinical-imaging score discriminated END (area under the curve [AUC] 0.833, 95% CI 0.760-0.906; 50.0% vs. 19.4%, p = 0.005). The unadjusted 90-day mRS score distribution was worse in the high-risk group (z = -2.405, p = 0.016), but the ABCD3-I-informed clinical-imaging score was not significant after adjustment for admission NIHSS in the proportional-odds model (adjusted common OR 1.05 per point, 95% CI 0.88-1.26; p = 0.573). The dichotomized disability rate (mRS ≥ 3: 18.6% vs. 9.7%) did not differ significantly (p = 0.285; adjusted OR 1.11, 95% CI 0.86-1.44). For 90-day disability, the AUC was 0.662 (95% CI 0.540-0.783; p = 0.017). At the prespecified ≥8-point threshold, 31 patients (23.3%) were classified as non-high-risk; the negative predictive value was 90.3% (95% CI 75.1-96.7%). Discrimination did not differ significantly from the original ABCD3-I score (ΔAUC 0.068, 95% CI - 0.063 to 0.196; DeLong p = 0.312). CONCLUSION: The ABCD3-I-informed score may identify a non-high-risk subgroup and show strong END discrimination, but its association with a 90-day outcome was modest and NIHSS-dependent. It should complement, not replace, baseline neurological assessment and should not guide treatment selection; prospective multicenter validation is required.