Clinical Research Collaboration for Stroke in Korea Imaging Repository: A Prospective Multicenter Neuroimaging Repository
Authors:
- Beom Joon KimBKBeom Joon Kim
- Wi-Sun RyuWRWi-Sun Ryu
- Myung Jae LeeMLMyung Jae Lee
- Kyusik KangKKKyusik Kang
- Jae Guk KimJKJae Guk Kim
- Soo Joo LeeSLSoo Joo Lee
- Jae-Kwan ChaJCJae-Kwan Cha
- Tai Hwan ParkTPTai Hwan Park
- Jeong-Yoon LeeJLJeong-Yoon Lee
- Kyungbok LeeKLKyungbok Lee
- Doo Hyuk KwonDKDoo Hyuk Kwon
- Jun LeeJLJun Lee
Preprint
Abstract
Background: Prospective stroke registries have advanced our understanding of cerebrovascular disease, yet most reduce neuroimaging to categorical variables, forfeiting the multidimensional information inherent in clinical imaging. We describe the CRCS-K Imaging Repository, a prospective multicenter platform that systematically collects all stroke neuroimaging and integrates artificial intelligence (AI)-based automated quantification with clinical and outcome data through a dedicated research platform, AISCAN. Methods: Building upon the Clinical Research Collaboration for Stroke in Korea (CRCS-K), a nationwide prospective registry, all neuroimaging (computed tomography [CT], magnetic resonance [MR], and angiography) performed during index hospitalization of consecutive acute ischemic stroke patients was collected from 18 comprehensive stroke centers. Imaging underwent centralized quality verification, sequence classification, and AI-based quantification. As a proof-of-concept application, we examined the association between pre-treatment imaging modality, treatment workflow efficiency, and functional outcomes in patients receiving intravenous thrombolysis (IVT) or endovascular treatment (EVT). Results: From June 2022 through May 2025, 225,159 imaging sequences were collected from 20,792 patients. AI-based quantification modules converted these into standardized numeric features encompassing ischemic lesion volumes, perfusion parameters, white matter hyperintensity burden, and cerebral microbleed counts. Substantial inter-hospital variation in imaging modality selection was observed, with MR-first workflows ranging from 1.0% to 56.7% across centers. In the proof-of-concept analysis, each additional imaging sequence was associated with prolonged door-to-treatment times for both IVT and EVT. Propensity score overlap-weighted analyses suggested numerically more favorable functional outcomes with CT-based imaging among EVT-treated patients, whereas differences among IVT-treated patients were smaller and less consistent. Conclusions: The CRCS-K Imaging Repository demonstrates the feasibility of large-scale, prospective neuroimaging collection integrated with AI-based quantification and clinical data. The infrastructure enables clinically consequential questions that conventional registries cannot address.Keywords
This publication has 38 references indexed in Scilit:
- RAPID CTA versus JLK LVO for large vessel occlusion detection: a pragmatic comparison of performance and common pitfallsNeuroradiology, 2026
- Validation of a novel automated Alberta Stroke Program Early Computed Tomography Score (ASPECTS) software: reliability, magnetic resonance imaging (MRI) correlation, and early ischaemic core discriminationClinical Radiology, 2025
- Segmentation of Leukoaraiosis on Noncontrast Head CT Using CT‐MRI Paired Data Without Human AnnotationBrain and Behavior, 2025
- Deep learning-assisted detection of intracranial hemorrhage: validation and impact on reader performanceNeuroradiology, 2025
- An international study presenting a federated learning AI platform for pediatric brain tumorsNature Communications, 2024
- Integrating the BIDS Neuroimaging Data Format and Workflow Optimization for Large-Scale Medical Image AnalysisJournal of Digital Imaging, 2022
- Federated learning for predicting clinical outcomes in patients with COVID-19Nature Medicine, 2021
- Time to Treatment With Endovascular Thrombectomy and Outcomes From Ischemic Stroke: A Meta-analysisJAMA, 2016
- Endovascular thrombectomy after large-vessel ischaemic stroke: a meta-analysis of individual patient data from five randomised trialsThe Lancet, 2016
- Moving towards best practice when using inverse probability of treatment weighting (IPTW) using the propensity score to estimate causal treatment effects in observational studiesStatistics in Medicine, 2015