ABSTRACT NUMBER: ESOC2026A1037 STROKE NEUROIMAGING UTILIZATION AND THE EFFICIENCY OF ACUTE STROKE TREATMENT: AN ANALYSIS OF 20,789 CONSECUTIVE ISCHEMIC STROKE CASES FROM THE CRCS-K IMAGE REPOSITORY

Abstract Background and aims Neuroimaging is essential for acute stroke triage and decision-making; however, significant heterogeneity exists in imaging modalities and sequences due to varying socioeconomic and clinical environments. Real-world utilization patterns remain underreported. Methods We established the CRCS-K Imaging Repository, a dedicated imaging database built upon the infrastructure of the prospective multicenter CRCS-K registry from 18 centers. Through this repository, we comprehensively collected all acute neuroimaging data for consecutive ischemic stroke patients since June 2022, encompassing scans from arrival to discharge, including those from referring hospitals (Figure 1). Anonymized imaging data were linked to clinical records and validated by a central imaging lab. We analyzed door-to-treatment times and outcomes in patients receiving reperfusion therapy, comparing CT-based protocols (NCCT±CTA±CTP) versus MR-based protocols (±PWI). Results Between June 2022 and May 2025, 214,589 imaging sequences were collected from 20,789 patients. Intravenous thrombolysis (IVT) was administered to 2,258 (11%) patients, and endovascular treatment (EVT) to 2,173 (11%). In the IVT group, pre-treatment imaging consisted of CT (68.7%), MR (18.8%), and pre-arrival imaging (12.5%). In the EVT group, the modalities used were CT (39.6%), MR (51.5%), and pre-arrival imaging (8.9%). Door-to-treatment times increased stepwise with the addition of imaging sequences (Figure 2). However, comparative analysis showed no significant differences in clinical outcomes between imaging modalities (Table). Conclusions Initial neuroimaging practices vary significantly across clinical settings. While the use of advanced neuroimaging extends door-to-treatment times, it does not appear to provide additional clinical benefits for the general stroke population compared to standard protocols. Conflict of interest All authors have nothing to disclose. Figure 1 - belongs to Results Figure 2 - belongs to Results Table 1 - belongs to Results