Machine Learning Classifies Posterior Circulation Infarction in Vertebrobasilar Dolichoectasia

XGBoost model classifies concurrent posterior circulation infarction in vertebrobasilar dolichoectasia (AUC 0.884, 95% CI 0.842-0.925) using imaging and clinical variables. Internal validation only; external and prospective validation required. A web calculator is available.

• Retrospective, single-center study of 380 patients (40.5% with infarction); training and internal validation split (70/30).
• Random Forest AUC 0.846; basilar artery diameter and midline distance were the dominant predictive features by SHAP analysis.
• Critical limitation: no external or prospective validation; performance may not generalize beyond the single institution.

RadPigeon summaries are original and for information only. They are not clinical advice.