Malignant melanomas may be difficult to differentiate from benign nevi on the basis of histology. Contrary to nevi, the majority of melanomas harbor chromosomal imbalances. Comparative genomic hybridization-based and fluorescence in situ hybridization (FISH) tests can help differentiating malignant from benign tumors. In the present study, eight-bacterial artificial chromosome (BAC) probes targeting chromosomes 6, 8, 9 and 11 were tested by FISH, and compared with a commercial four-color FISH probe set targeting chromosomes 6 and 11 in a first set of 62 tissue microarray-included melanocytic tumors (47 melanomas and 15 nevi). A second set of 108 tumors (70 melanomas and 38 nevi) was analyzed with the eight-probes kit, and manual counting was compared with the newly developed automated FISH signals counting and with semi-quantitative visual detection of chromosomal imbalances. Intra-tumor heterogeneity was also evaluated in 12 melanomas and 10 patients with paired melanoma samples. Testing the tumors from the first set with the commercial kit and the eight-probes test permitted to correctly identify 45/47 and 47/47 melanomas, respectively. In the second tumor set, 65/70 malignant tumors presented at least one chromosomal imbalance, whereas none was detected in the nevi. The agreement between manual and automated signals counting was better in good-quality FISH slides compared with poor-quality slides. Semi-quantitative visual appreciation of chromosomal imbalances also reached strong agreement with exact manual counting. In addition, a frequent cytogenetic heterogeneity within melanomas and between paired tumors was noticed in patients with metastatic melanomas. To conclude, FISH testing targeting chromosomes 6, 8, 9 and 11 enabled to differentiate the majority of melanomas from nevi but was difficult to automate. Tumor cytogenetic heterogeneity was frequent and could impair FISH testing.
Keywords: chromosome; fluorescence in situ hybridization; melanoma; nevus; tumor heterogeneity.