Computed tomography as a method of estimating bone mineral content in horses

With the advancement of technology, more tools and methods are being made available for the assessment of bone density in the live animal. Computed tomography (CT) is one such tool. In a CT scan, x-rays are passed through a subject and detectors record transmitted energy. The CT computer then constructs cross-sectional images from this data to represent the internal anatomy. The first objective of this study was to investigate CT as a means of estimating bone mineral content of the equine third metacarpal bone. Our hypothesis was that the cross-sectional area of the third metacarpal bone, as measured on CT images, would correlate significantly with ash weight of that section. Our second objective was to use CT to measure and compare bone density in the dorsal, palmar, lateral and medial cortices of the equine third metacarpus. We hypothesized that CT would detect differences in density by cortex. Twelve cadaver thoracic limb specimens, from horses of varying age, breed and prior use, were scanned on a CT 9800 (GE Medical Systems). Scanned sections of the limb were then skinned, dried, ether-extracted and ashed in a muffle furnace. Bone mineral content was expressed as grams of ash/cm section of bone. The image recorded by the CT was developed and assessed using an imaging photodensitometer and a corresponding software package to estimate the area of each bone section. Correlation was determined using SAS (6.12). Differences in bone density by cortex were calculated using a simple t-test. The maximum, minimum and mean estimates of third metacarpal bone cross-sectional area as determined by CT were 519, 351 and 423 mm2, respectively, while the maximum, minimum and mean ash weights were 8.9, 4.8 and 6.4 g/cm, respectively. Estimates of bone mineral content by ash and cross-sectional area as determined by CT were significantly correlated (r=.91, P=.0001). There was no difference in CT density numbers between lateral and medial cortices, however the dorsal cortex was lower in density than the lateral (P<.05) and the palmar cortex was lower in density than all other cortices (P<.05). This data suggests that in a research setting, CT is an effective method by which to calculate cross-sectional area and, hence, estimate bone mineral content. Computed tomography was also able to detect differences in bone density by cortex.