Supporting clinical decisions through artificial intelligencebased biomarker analysis and finite material modeling for non-invasive maxillofacial bone regeneration using BTCP/PLLA scaffolds

Abstract

Maxillofacial bone regeneration remains a critical challenge in reconstructive surgery, where the integration of bioresorbable scaffolds such as B-TCP/PLLA (beta-tricalcium phosphate/poly-L-lactic acid) has shown promising osteoconductive and osteoinductive properties. In this study, we developed a comprehensive clinical decision support framework that combines finite element modeling (FEM) and machine learning-driven biomarker analysis to evaluate bone healing in patients treated with B-TCP/PLLA scaffolds. Quantitative biomarkers, including alkaline phosphatase (ALP), vascular endothelial growth factor (VEGF), and runt-related transcription factor 2 (Runx2), were collected at multiple postoperative intervals and processed using supervised learning algorithms such as Random Forest and Support Vector Machine (SVM). Our FEM simulations provided insight into stress distribution and scaffold integration, supporting the prediction of ossification patterns. Results show a strong correlation between computational stress models and biomarker expression levels, with the predictive model achieving an AUC of 0.91, precision of 0.88, and recall of 0.85. These findings underscore the potential of combining biomechanical modeling with molecular profiling to enhance postoperative monitoring and early prediction of bone regeneration outcomes. The proposed system offers a novel, data-driven approach for personalized treatment planning and real-time clinical support, contributing to the development of next-generation AIassisted healthcare tools in maxillofacial surgery.