The tumor immune microenvironment correlates with neoadjuvant chemotherapy response in high-grade serous ovarian cancer.

Abstract

e17605 Background: Platinum-based neoadjuvant chemotherapy (NACT) plus interval debulking surgery (IDS) and adjuvant chemotherapy has been an established alternative for advanced epithelial ovarian cancer (EOC). The chemotherapy response score (CRS) has been applied to assess the efficacy of NACT in high-grade serous ovarian cancer (HGSOC). The purpose of this study was to assess the effect of NACT on tumor immune microenvironment (TIME) of HGSOC, and the correlation between the TIME and CRS. Methods: All patients underwent NACT with IDS. Formalin-fixed paraffin-embedded tissue samples were collected from pretreatment biopsy specimens and the corresponding post-NACT surgical specimens. The densities and percentage of various cell populations within the TIME were analyzed by multiplex immunohistochemistry (3DMedcines Inc.), including specific tumor infiltrating lymphocytes (TILs) (CD8+T cell, FOXP3+CD4+Treg cell and CD20+B cell), tumor associated macrophages (TAMs), NK cells (CD56 dim and CD56 bright). Patients were divided into three groups according to CRS based on pathological reaction of surgical specimens: CRS1 shows no or minimal response, CRS2 shows moderate response and CRS3 usually shows complete or near-complete response. Statistical analyses were performed using GraphPad Prism 7.0. P < 0.05 was considered statistically significant. Results: A total of 25 HGSOC patients were enrolled in this study, 32% (8/25) of patients with CRS 1, 64% (16/25) with CRS 2, and 4% (1/25) with CRS 3. After NACT, patients with CRS 2 or 3 had a significant increase of the densities of CD20+B cells in stroma (pre-NACT 15.71/mm3 vs. post-NACT 87.82/mm3, p = 0.048). We also observed borderline significant increases of percentage of CD20+B cells, the densities and percentage of CD8+T cells in stroma after NACT (pre-NACT vs. post-NACT, 0.32% vs.1.54%, 56.35/mm3 vs. 311.35/mm3, 0.95% vs. 5.48%, respectively). Meanwhile, no significant change in CD4+T, TAMs and NK cells. In patients with CRS 1, we observed a significant increase of the densities and percentage of CD4+FOXP3+Treg in stroma (pre-NACT vs. post-NACT, 1.38/mm3 vs. 19.75/mm3, p = 0.01; 0.04% vs. 0.41%, p = 0.017, respectively), while CD8+T, CD20+B, TAMs and NK cells did not change significantly. Additionally, we found a significant difference in CD56 bright NK infiltration when comparing CRS 2 or 3 group with CRS1 group in post-NACT (0.5/mm3 vs.12/mm3, p = 0.049; 0.005% vs. 0.23%, p = 0.048). Conclusions: These data suggest immune infiltrate in TIME may serve as a biomarker to predict response to NACT in HGSOC. The infiltration of CD8+ T and CD20+ B cell was more pronounced in CRS2 or 3, while the enhanced infiltration of FOXP3+CD4+Treg cells is associated with immunosuppressive TIME and poor response (CRS1). The mechanism of lymphocyte regulation is not completely understood, thereby deserving further investigation.