Abstract
Endodontically treated teeth (ETT) are prone to fracture due to structural compromise and conventional finite element analysis (FEA) has limitations in accurately predicting fracture behavior. Therefore, it is of interest to evaluate an artificial intelligence (AI)-enhanced FEA model for predicting fracture patterns in ETT restored with fiberglass, carbon fiber, zirconia and cast metal posts. Hence, a total of 120 maxillary premolars were tested, with the AI model trained on 500 prior FEA simulations and validated against experimental fracture resistance outcomes. The AI-powered FEA showed superior predictive accuracy (92.3%) compared to conventional FEA (76.8%) and closely correlated with actual fracture initiation sites (r = 0.91). Integration of AI with FEA enhances fracture prediction and may guide clinicians in selecting optimal post systems for improved outcomes in ETT.