Computed tomographic versus radiographic assessment of the visibility and features of subchondral cystic lesions in equine limbs

Subchondral cystic lesions (SCLs) are common in horses and often cause lameness and they are commonly diagnosed by radiography. This study compares radiographic and computed tomographic (CT) features of subchondral cystic lesions in horses. It was hypothesized, that CT has a superior diagnostic value for the assessment of the features of SCLs compared to radiography. Horses diagnosed with one or two SCLs in CT and a radiographic study of the affected bone were included. SCLs were evaluated on radiographs and CT for location, size, shape, homogeneity, continuity of the subchondral bone, peripheral sclerosis, presence of a fissure, osteoarthritis (OA) and periostitis. Vascular channels were only assessed with CT. Subsequently, radiographic findings were compared to CT. Forty-two SCLs in a total of 32 horses were included in the study. Radiographic visibility of SCLs diagnosed with CT was 79% (95% CI: 0.64 to 0.88). In joints with two SCLs, the second cystic lesion was more often missed radiographically (p = 0.014). A discontinuous subchondral bone was evident in CT in 40 of 42 SCLs whereas it was suspected in only 23 of 33 radiographically diagnosed SCLs. With CT, fissures were more commonly diagnosed and in 16 SCLs, vascular channels were identified. SCLs with fissures showed more periostitis with CT (p = 0.01). Radiographically missed SCLs presented less peripheral sclerosis in CT (p = 0.023). Intermodality agreement was borderline good (k = 0.60-0.66) for location (proximal, distal; lateral, medial, midsagittal), OA and periostitis, borderline moderate for presence of a fissure (k = 0.42), fair for homogeneity (k = 0.35), and only poor for location (dorsal, central, palmar/plantar), shape, peripheral sclerosis and continuity of the subchondral bone (k = 0.06-0.18). This study highlights the advantages of CT for the characterization of SCLs in horses. Vascular channels appear to be a frequent finding.