J Neurosurg 143:607–614, 2025
This study evaluated predictive models for brain aneurysm rupture risk using clinical, morphological, and advanced MRI radiomics data. Models combining these factors, especially radiomics, significantly improved accuracy in identifying symptomatic aneurysms, supporting a comprehensive, personalized approach for aneurysm risk stratification and treatment decisions.
• The study evaluated predictive models for assessing the risk of brain aneurysm rupture using high-resolution MRI, clinical, morphological, and radiomic data.
• 129 intracranial aneurysms were analyzed; 26% were symptomatic (ruptured or with warning symptoms).
• The PHASES score alone had low predictive ability (AUC 0.61–0.62) for symptomatic aneurysms.
• Adding morphological metrics (especially size ratio) and smoking status improved prediction (AUC up to 0.79).
• Incorporating aneurysm wall enhancement (AWE) metrics further increased model performance (AUC 0.82).
• The best model included age and radiomics features, achieving the highest predictive accuracy (AUC 0.87, 88% sensitivity).
• Younger age, current smoking, larger size ratio, and higher wall enhancement were associated with symptomatic aneurysms.
• A comprehensive approach using clinical, morphological, and advanced imaging/radiomics data improves aneurysm risk stratification.












You must be logged in to post a comment.