Uveal melanoma: non-invasive predictive testing.

Uveal melanoma (UM) is the most common intraocular tumor in the western population. Overall survival has not improved during the past decades, despite a shift towards more conservative eye treatments.1 Currently, the choice of treatment for UM depends on size and location of the tumor, secondary effects of the tumor on the eye (for instance inflammation or neovascular glaucoma), the status of the fellow eye, and patients’ choice. Irrespective of primary treatment of UM, close to half of the patients will develop metastases.2 UM spreads via haematogenous dissemination with a high tendency to metastasize to the liver in 90-95% of patients. More accurate prognosticators are under development; these include chromosomal profile, expression profile and nowadays BAP1 mutation status. All of these modalities require tumor tissue to be of any value.

Tumor tissue can be obtained by enucleation or FNAB (fine needle aspiration biopsy) during plaque placement, before or after proton beam or stereotactic radiotherapy. Obtaining tumor tissue by FNAB is invasive and poses operative risks; many patients might therefore refuse collection of tumor material, preventing the opportunity to gather information on tumor profile. In this way patients at risk of developing metastatic disease cannot be identified optimally. Development of new distant metastasis, even years after control of the primary tumor, is due to the presence of circulating tumor cells at the time of initial diagnosis.3 Biomarkers for early detection could provide a solution for this condition. Unfortunately, significant biomarkers for early detection are not yet available. Liver function tests (gamma-glutamyl transpeptidase (γGT) and lactate dehydrogenase (LDH) serum levels obtained annually or semi-annually for liver metastasis are being widely used, but do not improve survival.4,5 Other options that have been explored as serum biomarkers include S-100b (neural crest marker), melanoma inhibitory activity (MIA), tissue polypeptide specific antigen (TPS) and osteopontin (OPN).6 With newer and more sensitive techniques, circulating tumor cells can be detected and may also be potentially useful for diagnosis, identification of high risk patients and monitoring disease progression.7,8

In this issue of Journal of Ophthalmic Vision Research, Demirci and co-workers show a different approach for detection of metastatic risk.9 They evaluate the microarray expression of genes predictive for UM metastasis and genes that are differentially expressed in UM metastases, in normal whole human blood, and in human tissues that are prone to metastatic involvement by UM. The expression of a large panel of genes is analyzed in healthy normal whole blood, liver, lung and skin, by screening different gene expression databases. The evaluated genes in this study which were taken from the gene expression profiling signature described by Harbour and coworkers were based on a maximal difference in expression of the microenvironment of low and high risk tumors.10 The other panel of genes that has been analyzed in this study is the differentially expressed panel previously described by Meir et al., who already stated that the gene expression pattern of UM liver metastases demonstrated a resemblance to normal liver tissue.11 Although there was no overlap in genes between both panels, the hypothesis that homing of tumor cells requires a specific microenvironment is interesting. The panel of genes used in this study is not useful in detecting similarities in different microenvironments, however searching for circulating tumor cells and the concept of tropism remain subjects for further research.