Fracture path prediction in proximal femur using strain energy-based Dijkstra's algorithm

Abstract Proximal femoral fracture is a major issue associated with high mortality rates and entails significant rehabilitation costs. Conventional fracture prediction methods, primarily based on finite element analysis (FEA), often suffer from high computational costs and limited clinical applicability. To address these limitations, this study proposes a novel fracture path prediction method using Dijkstra's algorithm based on strain energy. The proposed method extracts the proximal femur from skeletal images, performs FEA to calculate strain energy, and applies Dijkstra's algorithm to predict fracture paths by identifying regions of structural vulnerability. Compared with conventional methods, the proposed approach significantly reduces computational time while maintaining high accuracy. The predicted fracture paths demonstrate strong agreement with experimental results, validating the method's effectiveness in replicating actual fracture mechanisms. This study provides an efficient and reliable tool for understanding proximal femoral fracture mechanisms and paves the way for the development of patient-specific preventive measures.