Determination of patient prognosis based on unruptured basilar artery aneurysm morphology using a novel dynamic prediction model

Introduction: The study aimed to develop a novel predictive system to assess the clinical survival of patients with basilar artery aneurysms (BAAs) following endovascular embolization by analyzing aneurysm and peripheral vessel morphologies. Methods: Fifty-eight BAA patients who underwent endovascular embolization were retrospectively studied, with a median follow-up of 17.5 months. A total of nine deaths (15.5%) occurred during the follow-up period. Due to the limited number of outcome events, we adopted a conservative modeling strategy that involved using the least absolute shrinkage and selection operator (LASSO)-penalized Cox regression for variable selection, followed by multivariable Cox regression to construct the prognostic model. An online dynamic nomogram was developed to predict survival. Model discrimination was evaluated using the concordance index (C-index) and time-dependent receiver operating characteristic (ROC) curves. Internal validation was performed using bootstrap resampling with calibration curves, and clinical utility was assessed using decision curve analysis (DCA). Results: = 0.003). The model demonstrated acceptable discrimination, with a C-index of 0.805. Calibration curves showed good agreement between predicted and observed survival rates. The DCA demonstrated superior net benefit across a broad range of threshold probabilities (0-100%), with a net reduction of more than 80 unsuccessful procedures per 100 patients at a threshold of 85%. Conclusion: This morphology-based, online, dynamic nomogram serves as a practical tool for predicting survival in BAA patients after endovascular embolization. However, given the limited sample size and the number of events, the findings should be considered exploratory, and external validation in larger cohorts is warranted.