Epithelial ovarian cancer (EOC) has a poor prognosis, with a 5 year survival rate of ~40%. A major therapeutic barrier is disease chemoresistant to the standard therapeutic regimen of taxol and platinum. Efficacious therapies in chemoresistant disease, a fuller understanding of the molecular mechanisms driving tumour evolution, and drugs to prevent emergence of chemoresistance are required. Bevacizumab, a VEGF inhibitor approved for second line treatment of recurrent, chemoresistant EOC, requires stratification biomarkers to identify patients who will benefit. We investigated VEGF gene promoter methylation in EOC from ICON7 clinical trial patients. Patients with low VEGF-B methylation showed better overall survival (OS) when treated with standard chemotherapy (HR=2.07(1.23, 3.49), P=0.006) but not bevacizumab (HR=1.77(0.99, 3.15), P=0.05), potentially identifying a patient sub-group which should not be treated with bevacizumab. An association between VEGF-C methylation and progression-free survival (PFS) in ICON7 patients (HR=0.66(0.52, 0.84), P=0.0006) was confounded by stage. ICGC patient samples collected at relapse showed elevated VEGF-C compared to primary presentation. In a CRISPR-generated VEGF-C null EOC model we demonstrated an autocrine mechanism by which VEGF-C deregulation and over-expression may drive EOC stage and metastasis by suppressing anoikis-induced apoptosis and promoting growth in non-adherent conditions. Further, we investigated the role of epigenetics and heterogeneity in acquisition of resistance to platinum-based therapies. Epigenetically defined sub-populations which show a reversibly resistant "drug tolerant" phenotype have been demonstrated to exist in chemosensitive tumour populations of many cancers. We isolated reversibly cisplatin tolerant populations (CTPs) from a cisplatin sensitive EOC cell line A2780, which were re-sensitised by EZH2 and HDAC inhibitors. Epigenomic profiling of these populations revealed heavily H3K27me3-modified chromatin in CTPs, preferentially in non-coding regions of the genome, which we hypothesise as a novel mechanism of tolerance for cisplatin as an initial stage for acquisition of stable resistance.