Measuring Equine Hooves in Radiographs and Computed Tomography Images Reveals Unexpected Size Differences

In a previous study on hoof biometry, we found that mathematical correction of measuring results from radiographs did not lead to complete correspondence to computed tomography (CT) results. The present study investigates this finding by comparing 13 measures of six cadaveric equine digits collected with the following workflows: radiographs with 1 and 2 m focus-object distance (FOD) (Xray 1 m/2 m), computed tomography images in planes defined based on anatomical landmarks (CTw), simulated radiographs based on the tomography dataset (virtual 120-mm slabs, Xray Sim) and measurements based on slices and projections aided by three-dimensional reconstruction models of hooves and bones based on the tomography data set (AMIRA). Furthermore, Xray 1 m/2 m values were mathematically corrected using factors calculated for each hoof (Xray 1 m/2 m corr). Results of all methods correlated, but absolute values showed differences. Xray 1 m/2 m values were systematically higher, Xray Sim and AMIRA values were lower than CTw values. Increasing FOD and mathematical correction of Xray values led to approximation to CTw values. Among measures producing unexpected results and large differences between methods were palmar process height, medial and lateral (but not dorsal) hoof wall thickness, dorsal hoof wall length, weight-bearing length and heel/toe to plumbline. Explanations might be primarily different definitions of landmarks in different methods, but also contrast settings for displaying both bone and soft tissue contours. Therefore, when comparing the measuring results collected using different methods, it is advisable to analyse relative rather than absolute values.