Micro-computed tomography prediction of biomechanical strength in murine structural bone grafts

Correlating massive bone graft strength to parameters derived from non-invasive imaging is important for pre-clinical and clinical evaluation of therapeutic adjuvants designed to improve graft repair. Towards that end, univariate and multivariate regression between measures of graft and callus geometry from micro-CT imaging and torsional strength and rigidity were investigated in a mouse femoral graft model. Four millimeter mid-diaphyseal defects were grafted with live autografts or processed allografts and allowed to heal for 6, 9, 12, or 18 weeks. We observed that allograft remodeling and incorporation into the host remained severely impaired compared to autografts mainly due to the extent of callus formation around the graft, the rate and extent of the graft resorption, and the degree of union between the graft and host bone as judged by post-mechanical testing analysis of the mode of failure. The autografts displayed greater ultimate torque and torsional rigidity compared to the allografts over time. However the biomechanical properties of allografts were equivalent to autografts by 9 weeks but significantly decreased at 12 and 18 weeks. Multivariate regression analysis demonstrated significant statistical correlations between combinations of the micro-CT parameters (graft and callus volume and cross-sectional polar moment of inertia) with the measured ultimate torque and torsional rigidity (adjusted R2=44% and 50%, respectively). The statistical correlations approach used in this mouse study could be useful in guiding future development of non-invasive predictors of the biomechanical properties of allografts using clinical CT.