Predictors of treatment benefit and prognosis in cutaneous malignant melanoma

This thesis describes work designed to identify genomic prognostic and predictive markers for melanoma using nucleic acids extracted from formalin-fixed (FFPE) tumour samples and emerging technologies for gene expression profiling. I report use of the cDNA-mediated annealing, selection, extension and ligation (DASL) assay using a 502-gene Human Cancer panel to identify prognostic markers. The study identified over-expression of SPP1 and DNA repair genes in primary tumours as associated with reduced relapse-free survival time. These genes remained associated with survival in analyses adjusted for current prognostic markers. I also report use of the DASL assay to study gene expression of tiny metastatic specimens, allowing assessment of gene expression in matched primary and metastatic tumours. I report studies of driver mutations in BRAF and NRAS as prognostic markers in which V600K BRAF mutations, were independently associated with survival in multivariate analysis. Gene expression profiling using the DASL assay identified a number of genes differentially expressed in BRAF and NRAS mutated tumours, for example ETV1 and TYRO3, providing biological insight into these tumours. Ulceration of a primary tumour is a poor prognostic sign, but paradoxically also predicts response to interferon- adjuvant therapy. I report greater numbers of macrophages in ulcerated tumours and both over-expression of interleukin-6 and deranged expression of genes in the interferon Jak-STAT signalling pathway. These factors may contribute to mechanisms of responsiveness to interferon- in ulcerated tumours. I report an investigation of predictive markers for dacarbazine therapy. The Fluidigm quantitative Real-time PCR system was used to identify over-expression of MGMT as independently associated with chemotherapy resistance and shorter survival after starting chemotherapy. Following validation of these findings, prognostic and predictive markers were identified which could potentially be used clinically to provide additional information to patients with melanoma, allowing a more personalised approach to melanoma treatment.