Adaptive Osteotomy Plan Generation and Optimization for Maxillectomy

Maxillectomy requires a balance between oncologically safe resection margins and preservation of the bony framework. These scaffolds are essential for maintaining orbital and oral functions. However, traditional manual planning depends heavily on clinical expertise. This often leads to long preparation times and suboptimal trade-offs between tumor resection and facial preservation. To address these challenges, we propose a morphology-adaptive automated osteotomy planning framework that integrates anatomical priors with global optimization for personalized surgical planning. The framework first defines resection boundaries using a morphology-adaptive initialization mechanism informed by tumor convex-hull volume and compactness. It then incorporates an anatomy-aware radial field to enforce dual geometric constraints and protect critical dentoalveolar structures. A global multi-objective optimization model is subsequently applied to determine the optimal osteotomy plane parameters, which are further refined through an iterative clipping mechanism guided by the maximum boundary violation depth to ensure clinical safety and boundary integrity. Validation on real-world clinical datasets shows that our method achieves high consistency with expert manual planning. Furthermore, it significantly enhances planning efficiency by reducing the time required for surgical preparation compared to manual methods. These findings demonstrate the potential of the framework to enable rapid, standardized preoperative planning for complex maxillofacial osteotomies.