Molecular subgroup of high-grade serous ovarian cancer (HGSOC) as a predictor of outcome following bevacizumab.

Abstract

5502 Background: HGSOC is a histopathological diagnosis and may represent multiple diseases at a molecular level. We investigated whether distinct molecular subgroups may influence treatment choice. Methods: mRNA was extracted from 265 macrodissected formalin fixed paraffin embedded HGSOCs from Scottish patients (pts) treated with primary debulking then platinum based chemotherapy. Transcriptional analysis was performed using the Ovarian DSA microarray. This was repeated using 283 UK samples from the ICON7 study [first line paclitaxel/carboplatin +/- concomitant and maintenance bevacizumab (bev) for 12 months]. Results: Unsupervised hierarchical clustering (Scottish tumours) identified three major subgroups, two with angiogenic gene upregulation and one with angiogenic gene repression and immune gene upregulation. This latter ‘immune’ subgroup had a superior overall survival (OS) compared to the other two subgroups combined [HR = 0.66 (0.46-0.94)]. A 63-gene expression signature to prospectively identify this subgroup was generated and validated as prognostic for OS in an independent dataset [HR = 0.32 (0.19-0.54)]. As the immune subgroup had repressed angiogenesis-related gene expression we hypothesised that these pts would benefit less from bev [power to detect interaction > 2 in predicted direction for progression free survival (PFS) in ICON7 was 88% (α=0.1, one-tail)]. In ICON7 the gene signature showed a difference in impact of bev on PFS between the immune and proangiogenic subgroups (1-sided test for interaction, p=0.016). For the immune group (41% of cases), the addition of bev conferred a worse PFS [HR = 1.73 (1.12-2.68)] and OS [HR = 2.00 (1.11-3.61)] compared to chemotherapy alone. In the proangiogenic group there was a non-significant trend to improved PFS for the addition of bev (median 17.4 vs 12.3 months in controls). Conclusions: An immune molecular subgroup of HGSOC has superior survival to other HGSOC. The addition of bev appears to significantly reduce PFS and OS in this subgroup. Patients in the proangiogenic subgroups have a trend towards a PFS benefit from bev. These data suggest a mechanism for stratification of bev therapy and should be validated in additional datasets.