Novel Computed Tomography-based Metric Reliably Estimates bone Strength, Offering Potentially Meaningful Enhancement in Clinical Fracture Risk Prediction

Abstract

Osteoporosis with resultant fractures is a major global health problem with huge socioeconomic implications for patients, families and healthcare services. Areal (2D bone mineral density (BMD) assessment is commonly used for predicting such fracture risk, but is unreliable, estimating only about 50% of bone strength. By contrast, computed tomography (CT) based techniques could provide improved metrics for estimating bone strength such as bone volume fraction (BVF; a 3D volumetric measure of mineralised bone), enabling cheap, safe and reliable strategies for clinical application, and to help divert resources to patients identified as most likely to benefit, meeting an unmet need. Here we describe a novel method for measuring BVF at clinical-CT like low-resolution (550µm voxel size). Femoral heads (n=8) were micro-CT scanned ex-vivo. Micro-CT data were downgraded in resolution from 30µm to 550µm voxel size and BVF calculated at high and low resolution. Experimental mechanical testing was applied to measure ex vivo bone strength of samples. BVF measures collected at high-resolution showed high correlation (correlation coefficient r 2 =0.95) with low-resolution data. Low-resolution BVF metrics showed high correlation (r 2 =0.96) with calculated sample strength. These results demonstrate that measuring BVF at low resolution is feasible, which also predicts bone strength. Measures of BVF should be useful for 214