Abstract DP260: Preprocedural Prediction of ICAD in Acute LVO: A Two-Center Head-to-Head Analysis

Background and Purpose: In acute large-vessel occlusion (LVO), intracranial atherosclerosis–related occlusion (ICAD) constitutes a substantial proportion of cases and carries a high risk of early reocclusion after thrombectomy, underscoring the need for preprocedural etiologic differentiation. Several preprocedural scores have been proposed to distinguish ICAD from embolic mechanisms (e.g., ISAT, ABC2D, ATHE, ICAS-LVO, REMIT, Score-ICAD), but their relative performance has not been evaluated head-to-head within a single cohort. We therefore aimed to compare the discriminative ability of six published scores for identifying ICAD versus embolic etiology. Methods: We utilized a prospectively maintained stroke registry from two centers and identified 1,288 consecutive stroke admissions between June 2021 and March 2025. Patients who underwent mechanical thrombectomy with complete clinical and imaging data were included, whereas those with missing clinical data were excluded due to the inability to calculate scores. The etiology of large vessel occlusion (intracranial atherosclerotic disease [ICAD] vs. embolic) was determined based on intraoperative findings. Each score was calculated, and its discriminative ability for etiology was assessed using the area under the receiver operating characteristic curve (AUC). Results: Ninety-one patients met the inclusion criteria (ICAD = 18, embolic = 73). AUCs (95 % CI) for stroke etiology were: ISAT 0.870 (0.664-1.000, P = 0.064), REMIT 0.793 (0.676–0.911, P < 0.001), Score-ICAD 0.707 (0.582-0.833, P = 0.013), ABC2D 0.627(0.504-0.751, P = 0.095), ATHE 0.584(0.435-0.734, P = 0.311) and ICAS-LVO 0.465(0.301-0.630, P = 0.650). Conclusion: In this two-center registry, REMIT provided the most robust and statistically significant preprocedural discrimination between ICAD and embolic LVO, with Score-ICAD showing modest performance. ISAT had the highest point estimate but did not reach statistical significance, while ABC2D, ATHE, and ICAS-LVO showed limited or no discrimination. REMIT may aid strategy selection, but larger prospective validation is warranted.