Radiomics model for risk stratification of intracranial aneurysm: a high-resolution vessel wall imaging-based study
Brain imaging analysis using artificial intelligence spots dangerous aneurysms far better than current scoring systems, potentially helping doctors avoid unnecessary treatment of harmless ones.
This two-center retrospective study developed and validated a high-resolution vessel wall imaging radiomics model integrating aneurysm wall, parent artery, and anatomical location features, achieving AUC 0.888 for identifying symptomatic intracranial aneurysms versus the conventional PHASES score (AUC 0.679). The model significantly outperforms current risk stratification and could help guide clinical management decisions for the large number of incidentally discovered aneurysms.
What the study was
- Study design
- Retrospective two-center radiomics model development and validation
- Population
- Patients with intracranial aneurysms
- Sample size
- 446
- Category
- Diagnostics
- Maturity
- Exploratory
- Journal
- BMC Medical Imaging
Why it surfaced
Solid AUC improvement over existing risk tool in neuroimaging; two-center validation adds credibility but retrospective design and single organ system focus limit broader relevance. Deferred from 2026-05-13 run.
A plain-language summary of published research — not medical advice. Talk to a clinician about your own care.