Can quantitative computed tomography detect bone morphological changes associated with catastrophic proximal sesamoid bone fracture in Thoroughbred racehorses?

BACKGROUND: Fracture of the proximal sesamoid bones continues to be the most common fatal musculoskeletal injury in US racehorses. Identifying factors that influence fracture risk could lead to screening techniques to reduce catastrophic injury rates and improve animal welfare. OBJECTIVES: To identify morphological differences between proximal sesamoid bones of the contralateral limb of fracture and control horses and assess the feasibility of computed tomography (CT) to detect traits associated with proximal sesamoid bone fracture. We hypothesised that horses with proximal sesamoid bone fracture would have greater bone density. STUDY DESIGN: Cross-sectional cadaver morphological study. METHODS: Proximal sesamoid bone morphology was measured using high-resolution micro-CT images from 16 Thoroughbred racehorses (eight fracture, eight control) euthanised on New York racetracks. Nominal logistic regression models and receiver operating characteristic curves were created to assess the ability of CT-derived morphological traits to accurately classify fracture horses vs. controls. RESULTS: Bone volume fraction was greater in the fracture group (90.39 ± 1.76%) as compared to controls (87.20 ± 2.79%, P<0.0001). Bone volume fraction, bone width, trabecular thickness and degree of anisotropy were significantly different between fracture and control horses. Receiver operating characteristic curves showed that a combined model that incorporates bone volume fraction and width can identify fracture from control horses with an area under the curve of 0.938, indicating high accuracy at classifying fracture horses from controls. MAIN LIMITATIONS: The number of horses per group is small, although the total number of sesamoids imaged is reasonable (n = 62). In vivo CT at the resolution performed in this study is currently unattainable; however, density and width could be measured with quantitative CT. CONCLUSIONS: Differences in proximal sesamoid bone morphology were identified between fracture and control horses. As improved technology becomes accessible, quantitative CT could potentially be used as a clinical imaging technique to estimate proximal sesamoid bone fracture risk in Thoroughbred racehorses.

Spatial relationships between bone formation and mechanical stress within cancellous bone.

Bone adapts to mechanical stimuli. While in vivo mechanical loading has been shown to increase the density of cancellous bone, theory suggests that the relationship between tissue stress/strain and subsequent bone formation occurs at the scale of individual trabeculae. Here we examine bone formation one week following mechanical stimulus. Three bouts of cyclic loading (300 cycles/day on 3 consecutive days) were applied to caudal vertebrae of female rats (n=7). Bone formation was determined using three-dimensional images of fluorescent markers of bone formation (0.7×0.7×5.0µm(3)) and local tissue stress/strain was determined using high-resolution finite element models. Three days of mechanical stimuli resulted in an increase in mineralizing surface (loaded: 17.68±2.17%; control: 9.05±3.20%; mean±SD) and an increase in the volume of bone formed (loaded: 7.09±1.97%; control: 1.44±0.50%). The number of bone formation sites was greater in loaded animals (650.71±118.54) than pinned not loaded controls (310.71±91.55), a difference that was explained by the number of formation sites at regions with large local tissue strain energy density (SED). In addition, the probability of observing bone formation was greater at locations of the microstructure experiencing greater SED, but did not exceed 32%, consistent with prior work. Our findings demonstrate that bone formation in the week following a short term mechanical stimulus occurs near regions of bone tissue experiencing high tissue SED, although the ability of finite element models to predict the locations of bone formation remains modest and further improvements may require accounting for additional factors such as osteocyte distribution or fluid flow.