Predicting Distal Radius Failure Load during a Fall using Mechanical Testing and Peripheral Quantitative Computed Tomography

Distal radius fractures are the most common form of osteoporotic fracture in women and play an important role in predicting other osteoporotic fractures. Colles’ fracture, a type of DRF, result from a fall from standing height or less. Peripheral quantitative computed tomography (pQCT) imaging is commonly used to estimate distal radius strength (resistance to failure) via bone strength indices such as BSIc (related to compressive axial loading resistance). BSIc has been validated in experimental compressive testing. However, during a fall, the distal radius is subjected to a combination of dorsal-directed forces (which result in bending) and axial compression. The primary objective of this study was to validate new pQCT-based bone strength indices combining resistance to bending and compression using optimized and clinically-applied image resolutions. The secondary objective was to validate these new indices against reported bone strength indices and bone properties for predicting the failure load in a mechanical testing scenario representing a fall on the extended hand. Fourteen cadaveric forearms, with the hand intact, were scanned using pQCT at 4% of the length of the radius away from the distal end. Bone was defined as pixels with density > 100 mg/cm3 and cortical bone as pixels with density > 480 mg/cm3 using BoneJ, a tool designed to be used with ImageJ, an open source image analysis tool. This thresholding provided the basis for various measures which have been used in existing literature to predict failure load. Novel bone strength indices were calculated using composite beam theory based on the density of each pixel using total bone area, total volumetric bone mineral density and a density weighted modulus. Each of the novel measures examined the point of maximum stress in a single direction; this combined the uniform axial load applied over the cross-section and the bending resulting from an off-axis load, like that experienced during a fall. After scanning, potted samples were placed in a material testing system (MTS Bionix) with 15° of dorsal inclination and 3-6° of radial inclination, corresponding with the hand positon during a fall. Testing was performed at 3mm/s (180 mm/min) until fracture occurred and ultimate failure load was recorded. Linear regression models were used to assess imaged-based bone strength indices and bone properties predicting variance (coefficient of determination, R2) in the experimentally derived failure load. A new bone strength index BSIM , bone strength index in medial direction - which considered axial loading and bending stresses at the farther medial point on the radius, explained up to 90% of variance in the experimental failure load. The highest coefficient of determination from metrics used previously in the literature was total bone mineral content (R2 = 0.88). Two other novel bone strength indices, BSIV (farthest point in the volar direction) and BSID (farthest point in the dorsal direction) predicted 88% of variance. Additionally, BSIL (farthest point in the lateral direction) explained 86% of variance. This validates the use of these new measures as predictors of failure load in the distal radius during a fall. This work also found the existing measure of bone strength index in compression, BSIc, predicted up to 83% of variance in the experimental failure load, which validates its use on the radius instead of remaining as a tibia specific tool