Comparison of subchondral bone density determined by quantitative computed tomography (CT) with gross and histopathologic changes have not been made in horses. The goal of this study was to determine if mean quantitative CT density and mean voxel standard deviation are associated with the presence and severity of osteochondral lesions in the palmar aspect of the distal third metacarpal bone in racing horses. Metacarpophalangeal joints from nine racehorses were imaged using CT and scored for gross damage. Four-millimeter-thick sagittal and 30° palmar dorsal plane sections were cut, decalcified and stained with hematoxylin and eosin from the distal third metacarpal bone. Microscopic osteochondral lesions and subchondral remodeling were scored on a scale of 0–3. Percent subchondral bone, expressed as the ratio of bone volume to tissue volume, was also measured. Mean quantitative CT density and mean voxel standard deviation were measured from three-dimensional models of CT images comparable with histologic sections. Mean quantitative CT density was not associated with lesion severity or number of lesions. A weak correlation between mean quantitative CT density and gross score was found, but mean quantitative CT density was not predictive for gross score. Mean voxel standard deviation was not correlated with gross or histopathologic measures, but was predictive of mild osteochondral lesions. Results support the association of subchondral remodeling with the development of palmar metacarpal lesions. However, there was not a strong correlation between mean quantitative CT density or mean voxel standard deviation and histopathologic lesions of the distal third metacarpal bone.