Work

Analysis of a clot-based combined radiomics model for predicting embolic etiology in acute ischemic stroke patients

Year: 2025

Type: article

Abstract: Background and objectives: Different etiologies of acute ischemic stroke (AIS) are considered important factors affecting neurological function. The accurate etiological classification of AIS prior to surgery is crucial. To investigate and explore the predictive value of a clot-based combined radiomics model for identifying the etiological subtypes of acute ischemic stroke. Materials and methods: A total of 263 patients with acute ischemic stroke caused by anterior circulation large artery occlusion were retrospectively enrolled. These were grouped into training (180), testing (45), and external validation cohorts (38). NCCT and CTA scans were adopted to segment region of interest (ROI) of clots. Feature selection was conducted to establish Clinical model, radiomics models (NCCT, CTA, and NCCT&CTA), and combined model. Results: The AUCs of the clinical model and radiomics models (NCCT, CTA and NCCT&CTA) in the testing cohort were 0.8288(95 % CI: 0.7174-0.9403), 0.8133(95 % CI:0.6853-0.9414), 0.8075(95 % CI:0.6844-0.9307) and 0.8535(95 % CI:0.6774-1), respectively. The combined model achieved a greater AUC than the other four models in the testing cohort (0.9077 [95 % CI: 0.821-0.9944]). Clinical decision curve analysis (DCA) demonstrated that the radiomics (NCCT&CTA) model and combined model show better net benefits within a relatively wide range of threshold probabilities. Conclusion: The combined radiomics model achieved good predictive efficacy for distinguishing the etiological subtypes of acute ischemic stroke and can provide valuable information for the precise selection of recanalization strategies in clinical practice.

Source: European Journal of Radiology Open

DOI: 10.1016/j.ejro.2025.100709

Authors: Ru Ding, Kun An, Fen Wang, Zirui Cao, Yan Zeng, Lili Guo

Institutions: Huaian First People’s Hospital, Nanjing Medical University, United Imaging Intelligence (China)

Language: English

Topic: Radiomics and Machine Learning in Medical Imaging

Subfield: Radiology, Nuclear Medicine and Imaging

Field: Medicine

Domain: Health Sciences

SDG: Good health and well-being

Open Access status: gold

Estimated APC paid: $2,610