Early immunohistochemical detection of pulmonary micrometastases in dogs with osteosarcoma

BACKGROUND: Despite decades of research, the early phases of metastatic development are still not fully understood. Canine osteosarcoma (OS) is a highly aggressive cancer, with a high metastatic rate (> 90%), despite a low overt metastatic prevalence at initial diagnosis (< 15%). Canine OS is generally regarded as a good clinically relevant model for human OS. The aim of this hypothesis-generating study was to evaluate a method to detect pulmonary micrometastases and study their prevalence in dogs with OS without macroscopic metastases. We prospectively enrolled dogs with OS that received no cancer-specific treatment (n = 12) and control dogs without cancer (n = 2). Dogs were necropsied and sampled immediately after euthanasia. The OS dogs were classified as having macroscopic metastases (n = 2) or not (n = 10). We immunohistochemically stained one tissue sample from each of the seven lung lobes from each dog with a monoclonal antibody (TP-3) to identify micrometastases (defined as clusters of 5–50 tumour cells), microscopic metastases (> 50 tumour cells) and TP-3 positive single cells (< 5 tumour cells). RESULTS: We showed that pulmonary micrometastases easily overseen on routine histology could be detected with TP-3. Pulmonary micrometastases and microscopic metastases were present in two dogs with OS without macroscopic metastases (20%). Micrometastases were visualised in three (43%) and four (57%) of seven samples from these two dogs, with a mean of 0.6 and 1.7 micrometastases per sample. Microscopic metastases were present in one (14%) and four (57%) of seven samples from the same two dogs, with a mean of 0.14 and 1.0 microscopic metastases per sample. There were four (57%) and two (29%) samples with neither microscopic metastases nor micrometastases for each of these two dogs. The prevalence of pulmonary micrometastases (20%) was significantly lower than expected (> 90%) based on commonly expected metastatic rates after amputation (P < 0.0001). There was no statistically significant difference in the number of TP-3 positive single cells in between groups (P = 0.85). CONCLUSIONS: Pulmonary micrometastases could be detected with TP-3 immunohistochemistry in a subset of dogs with OS before macroscopic metastases had developed. We propose that dogs with spontaneous OS represent clinically relevant models to study early micrometastatic disease.