Optimal Surgical Timing and Outcome Prediction in Hemorrhagic Moyamoya Disease: A Retrospective Cohort Study

To develop and validate a prognostic nomogram incorporating optimal surgical timing thresholds for predicting long-term outcomes in adult hemorrhagic moyamoya disease (MMD) patients undergoing encephaloduroarteriosynangiosis (EDAS).

We conducted a retrospective cohort study of 256 consecutive adults with hemorrhagic MMD treated with EDAS (2013-2017) at a tertiary neurosurgical center. Using least absolute shrinkage and selection operator (LASSO) regression with 10-fold cross-validation, we identified key predictors for nomogram development. The model's performance was rigorously evaluated through discrimination (C-index, AUC-ROC), calibration, and decision curve analysis (DCA) in both training (70%) and validation (30%) cohorts.

Our hemorrhage-specific nomogram incorporated five independent predictors: (1) age at initial hemorrhage, (2) surgical timing (optimal 3-6 month window post-ictus), (3) preoperative functional status (modified Rankin Scale), (4) perioperative complications, and (5) recurrent hemorrhage. The model demonstrated excellent discrimination (C-index: 0.889 training, 0.819 validation) and calibration. DCA confirmed superior clinical utility across 15%-80% probability thresholds compared to conventional approaches. Restricted cubic spline analysis validated the non-linear relationship between surgical timing and outcomes.

This study provides the first validated clinical decision tool for adult hemorrhagic MMD, establishing 3-6 months post-hemorrhage as the optimal surgical window for EDAS. The nomogram's robust performance metrics and RCS-validated temporal risk stratification offer neurosurgeons an evidence-based approach to optimize revascularization timing and improve functional outcomes.