Disseminated tumour cells in bone marrow of patients with uveal melanoma

Abstract. Purpose: Approximately 50% of patients with uveal melanomas develop metastases. Thus, it is important to improve our understanding of how melanoma metastases develop. Methods: As part of a uveal melanoma micrometastasis study, we compared the detection rates of immunomagnetically selected (IMS) tumour cells in bone marrow (BM) with positively stained tumour cells using immunocytochemistry (ICC). Bone marrow mononuclear cells were isolated. Immunocytochemistry cytospin preparations were immunocytochemically stained in parallel with two different melanoma antibodies, 9.2.27 and HMB45. Using IMS, melanoma cells were selected from BM mononuclear cell fractions using immunomagnetic beads coated with the 9.2.27 antibody and identified by light microscopy. Results: In cytospin preparations from 226 patients, melanoma cells were detected in 24 (10.6%), 10 with 9.2.27 and 17 with the HMB45 antibody. In three cases, we found positive cells with both antibodies. Six of the 226 (2.6%) patients that stained positively with ICC died with metastatic disease, all also positive with IMS. Sixty-six (29.2%) patients had positive BM samples with IMS at the first examination. Immunomagnetic selection (IMS) was positive in 36.8% of the 57 patients who later developed clinical metastases. Twenty-one IMS-positive patients and 31 IMS-negative patients died of metastases, in total 52 of 226 patients (23.0%). The mortality rate of melanoma metastasis was 24% (6/24) after at least 4 1/2 years in ICC-positive patients compared to 38.5% (20/52) in IMS-positive patients. Conclusion: The presence of melanoma cells in BM of patients with uveal melanoma is documented in our study with IMS and ICC. Immunomagnetically selected is more sensitive than ICC in detecting tumour cells in BM. However, statistically, we did not find any prognostic impact of the presence of melanoma cells in BM.