A Novel Approach by Integrating CT-Based Imaging Data and Machine Learning to Predict Patient-Specific Young's Modulus Values.

Abstract

Finite element analysis (FEA) stands as a cornerstone in preclinical investigations for implant therapy, particularly in orthopaedics and biomechanics. Accurate modelling of bone properties is crucial for meaningful FEA outcomes, considering the complex nature of bone tissue. This study proposes a novel approach by integrating CT-based imaging data and machine learning to predict patient-specific Young's modulus values. A back propagation neural network (BPNN), incorporating texture properties extracted from CT images, demonstrates robustness in predicting Young's modulus. Validation against three-point bending experiments on rabbit femur bones shows promising results, with stress values within 13% of those from FEA. The proposed methodology holds the potential for enhancing preclinical evaluations of implant therapy and fostering the development of patient-specific implants for improved clinical outcomes.