Identification of vulnerable intracranial atherosclerotic plaques via the 3D-HRMR-VWI radiomics model

Motivation: Intracranial artery atherosclerosis (ICAS) is a major stroke cause. Enhancing diagnostic efficiency for plaque vulnerability may aid in risk stratification and targeted treatment. Goal(s): To evaluate the capability of a 3D-HRMR VWI radiomics model in diagnosing vulnerable plaques within intracranial arteries. Approach: Using non-contrast T1WI and enhanced CE-T1WI sequences, 303 stroke patients' plaques were categorized as vulnerable or stable. Radiomic features were extracted and used to construct three diagnostic models, with their performance evaluated through AUC analysis. Results: The combined model demonstrated high diagnostic performance with AUCs of 0.878 in the training set and 0.889 in the test set. Impact: Integrating radiomics into intracranial plaque risk assessment enables quantitative scoring of plaque vulnerability, enhancing diagnostic performance and supporting clinical risk stratification of intracranial atherosclerotic plaques.