Osteosarcoma (OSA) is the most common primary bone tumor in dogs and cats, and accurate diagnosis of OSA has therapeutic and prognostic implications. Immunohistochemistry targeting osterix (Osx), a transcription factor expressed in osteoblasts, was applied to 80 canine samples (40 OSAs and 40 controls (8 chondrosarcomas, 8 fibrosarcomas, 8 hemangiosarcomas, 8 histiocytic sarcomas, and 8 malignant melanomas)) and 40 feline samples (20 OSAs and 20 controls (5 fibrosarcomas, 5 hemangiosarcomas, 5 chondrosarcomas, 3 malignant melanomas, and 2 histiocytic sarcomas)), diagnosed histologically. All 120 samples were assessed blinded using a previously established semiquantitative scoring and classification system. In dogs, 37/40 OSAs (true positives, sensitivity 92.5%) and 2/40 controls (false positives, specificity 95.0%) were classified as OSA. In cats, 14/20 OSAs (true positives, sensitivity 70.0%) and 5/20 controls (false positives, specificity 75.0%) were classified as OSA. Tumors most often "misclassified" as OSAs were chondrosarcomas in both dogs (2/8) and cats (4/5). In 4 cases for which paired decalcified and nondecalcified samples of the same OSA were available, decalcification of OSAs did not affect classification. Quiescent canine, feline, and murine osteoblasts and canine reactive, metaplastic, and non-OSA neoplastic osseous conditions were also tested and demonstrated Osx immunoreactivity; avian and piscine osteoblasts did not. In conclusion, Osx is a highly sensitive and specific marker of canine OSA and a moderately sensitive and specific marker of feline OSA. The diagnostic algorithm applied in this study is a useful adjunctive tool in the diagnosis of canine OSA.
Ings et al. (Tue,) studied this question.