Cancer Medicine (May 2021)

Tumor genomic, transcriptomic, and immune profiling characterizes differential response to first‐line platinum chemotherapy in high grade serous ovarian cancer

  • Johanne I. Weberpals,
  • Trevor J. Pugh,
  • Paola Marco‐Casanova,
  • Glenwood D. Goss,
  • Natalie Andrews Wright,
  • Prisni Rath,
  • Jonathon Torchia,
  • Alexander Fortuna,
  • Gemma N. Jones,
  • Martine P. Roudier,
  • Laurence Bernard,
  • Bryan Lo,
  • Dax Torti,
  • Alberto Leon,
  • Kayla Marsh,
  • Darren Hodgson,
  • Marc Duciaume,
  • William J. Howat,
  • Natalia Lukashchuk,
  • Stanley E. Lazic,
  • Doreen Whelan,
  • Harmanjatinder S. Sekhon

DOI
https://doi.org/10.1002/cam4.3831
Journal volume & issue
Vol. 10, no. 9
pp. 3045 – 3058

Abstract

Read online

Abstract Background In high grade serous ovarian cancer (HGSOC), there is a spectrum of sensitivity to first line platinum‐based chemotherapy. This study molecularly characterizes HGSOC patients from two distinct groups of chemotherapy responders (good vs. poor). Methods Following primary debulking surgery and intravenous carboplatin/paclitaxel, women with stage III–IV HGSOC were grouped by response. Patients in the good response (GR) and poor response (PR) groups respectively had a progression‐free intervals (PFI) of ≥12 and ≤6 months. Analysis of surgical specimens interrogated genomic and immunologic features using whole exome sequencing. RNA‐sequencing detected gene expression outliers and inference of immune infiltrate, with validation by targeted NanoString arrays. PD‐L1 expression was scored by immunohistochemistry (IHC). Results A total of 39 patient samples were analyzed (GR = 20; PR = 19). Median PFI for GR and PR patient cohorts was 32 and 3 months, respectively. GR tumors were enriched for loss‐of‐function BRCA2 mutations and had a significantly higher nonsynonymous mutation rate compared to PR tumors (p = 0.001). Samples from the PR cohort were characterized by mutations in MGA and RAD51B and trended towards a greater rate of amplification of PIK3CA, MECOM, and ATR in comparison to GR tumors. Gene expression analysis by NanoString correlated increased PARP4 with PR and increased PD‐L1 and EMSY with GR. There was greater tumor immune cell infiltration and higher immune cell PD‐L1 protein expression in the GR group. Conclusions Our research demonstrates that tumors from HGSOC patients responding poorly to first line chemotherapy have a distinct molecular profile characterized by actionable drug targets including PARP4.

Keywords