Uveal melanoma (UM) is the most common intraocular malignancy in adults. Despite successful ocular treatment, about 50% of patients succumb to metastatic dissemination, which occurs haematogenously and mainly affects the liver. On the basis of clinical, histopathological and genetic features of the primary tumour it is possible to predict if the individual patient is at high risk (HR) or low risk (LR) of developing metastases. However, the mechanisms responsible for the development of metastatic disease in UM are still largely unknown; therefore no adjuvant treatment is currently offered to HR patients to prevent development of fatal disease. As the time to discovery of clinically detectable metastases can range from months to decades, a secreted biomarker(s) that could be routinely tested in blood is much needed. The scope of the work presented in this thesis was to use proteomics as a tool to identify potential novel, UM-specific biomarkers. Moreover, the proteomic data acquired would complement genomic and transcriptomic information already generated by the Liverpool Ocular Oncology Research Group, with the ultimate aim of increasing our understanding of UM development and dissemination. The aim of Chapter 2’s project was to compare the proteome of UM tissue samples at HR versus LR of developing metastatic disease using isobaric tags for relative and absolute quantitation (iTRAQ) labelling and mass spectrometry (MS). The quantification of proteins in our samples, proteomic analysis and further validation by immunohistochemistry has led to the identification of two novel prognostic and potentially therapeutic target, S100A6 and the tumour suppressor PDCD4. In Chapter 3 we focused on proteins released in the conditioned medium (secretome) of short-term cultures of HR and LR UM cells, as well as normal melanocytes. Using a label-free quantitative proteomic approach, almost 2000 proteins were identified and quantified, with more than 30% of these identified as secreted and/or previously described in exosomes. Using these data, an 18-protein signature able to discriminate between HR and LR UM was identified. Further validation will be necessary in secretome samples and in the peripheral blood of UM patients, but this has the potential of being translated into a clinically useful assay to detect early development of metastatic disease. As reported in Chapter 4, we also conducted a pilot clinical study on circulating tumour cells (CTC) in UM, using the CellSearch® platform with the novel melanoma kit to enumerate CTC in the peripheral blood of UM patients at LR, HR or with overt metastatic disease. CTC were detected in metastatic and HR tumours and were not present in LR UM, however, the number of CTC detected varied widely, calling into question the clinical value of using this platform in UM patients. The research detailed in Chapter 5 had a direct clinical value, as it addressed the procedures undertaken during the acquisition and processing of prognostic biopsies from UM tumours treated conservatively. The modifications introduced led to a significant improvement of the success rate of such prognostic biopsies for risk stratification, which is essential for clinical management, follow-up and research purposes. In conclusion, the work conducted throughout this PhD has provided further insight into the molecular characteristics that can differentiate between HR and LR UM, identifying novel potential biomarkers that will need validation in the clinical setting.