Genomic profiling and identification of high‐risk tumours in uveal melanoma by array‐CGH analysis of primary tumours and liver metastases

Abstract

Abstract Purpose Current therapeutic efforts in Uveal Melanoma are directed toward detection of liver metastases at an earlier stage and adjuvant systemic therapy in high risk patients. Beside clinical and histological features, specific cytogenetic alterations, particularly monosomy 3 and gain of 8q, are strongly associated with metastasis. Methods A series of of 78 ocular tumours (OT) (median follow‐up 54 mo.)and of 66 liver metastases (LM), was analysed by CGH on a genome‐wide BAC/DNA microarray (CIT, INSERM U830), with a 1 Mbase average resolution. Correlations were looked for between genomic profiles of OT and the metastatic status of patients, and a prognostic classifier was built, in order to identify tumour profiles of high risk patients. Results Hierarchical clustering shows that status of chromosome 3 defines two groups of genomic profiles in OT and LM: group 1, with disomy 3, and group 2, with monosomy 3. Seven OT show a partial loss of chr 3, with a minimal deletion of 8.9 Mbase, distal to 3p25.3. Groups 1 and 2 can be subdivided in subgroups according to the presence of additional imbalances. Same genomic groups are found in OT and LM, but with different frequencies. The disomy 3 group represents 20% of metastases. A prognostic classifier including the status of chromosomes 3, 6p, 8p, 8q, 16q, and the position of breakpoint in 8q gains, leads to the best prediction performance in this sample (82%). Conclusion Genomic profiling by array‐CGH, combined with the allelic status of chromosome 3, should be a robust and reliable approach for identifying high‐risk patients eligible for LM screening and adjuvant systemic therapy.