Leïla Bengrine Lefevre

Unraveling the Tumor Microenvironment and PD-L1 Expression across Tissue Types in High-Grade Serous Ovarian Cancer in the NeoPembrOV/GINECO Phase II Randomized Trial

Abstract Purpose: To describe PD-L1 expression across tissue types and its associated tumor microenvironment and to investigate how it affects its predictive value for response to pembrolizumab in treatment-naïve patients with ovarian cancer included in the NeoPembrOV phase II trial (NCT03275506). Experimental Design: PD-L1 expression was assessed for 85 patients (56 on metastasis and 29 on tubo-ovary) using tumor proportion score (TPS) and immune cell (IC) score, considering positivity if ≥1% and high expression if ≥5%. RNA sequencing and multiplex immunofluorescence were conducted. The Australian Ovarian Cancer Study was used as an external validation cohort. Results: PD-L1 was primarily expressed by tumor cells in tubo-ovaries and by ICs in metastases. The IC score assessed on the metastases was associated with a longer progression-free survival in the pembrolizumab arm compared with the control arm. Compared with tubo-ovaries, metastases were enriched in T and B cells as well as in granzyme B (GZMB) CD8 cytotoxic T-cell signatures. In metastases, the IC score was associated with immune infiltration and overexpression of additional immune checkpoints, such as IDO1, LAG3, and ICOS, whereas TPS was associated with cell proliferation, immune infiltration, and IFN-γ pathways. In tubo-ovaries, TPS was associated with pathways linked to cell proliferation and antigen presentation but was depleted in activated immune pathways, and CD274 expression was correlated with hypoxia and PI3K/Akt/mTOR signaling. Conclusions: Distinct PD-L1 expression patterns across tissue types are associated with different biological pathways and tumor microenvironments in ovarian cancer, affecting PD-L1 predictive value. Our results provide novel insights into high-grade serous ovarian cancer biology for tailoring immunotherapy in patients with ovarian cancer.

ESR1 Mutation in Endocrine Treatment-Naïve Endometrial Cancer: Prevalence, Characteristics, and Prognostic Implications, Results from the UTOLA Phase II GINECO Trial

Abstract Purpose: Aromatase inhibitors (AI) are used to treat estrogen receptor (ER)–positive low-grade endometrioid endometrial cancer. In breast cancer, ESR1 mutations are rare at diagnosis (<5%) but are frequently acquired in AI-resistant cases and are considered one of the major resistance mechanisms to endocrine therapy. This study aimed to assess the prevalence of ESR1 mutations in hormonotherapy-naïve endometrial cancer samples and correlate them with molecular profiles, ER expression, and clinical outcomes. Experimental Design: A total of 147 patients with advanced endometrial cancer who had responded to first-line chemotherapy were recruited into the UTOLA trial. Archival endometrial cancer tumor tissues underwent sequencing of 127 genes, including ESR1. Only hotspot mutations in the ligand-binding domain were evaluated. ESR1 mutation prevalence was validated in the Genomics England dataset. In UTOLA, tumors were classified as POLE, MMR deficient, TP53abn, or no specific molecular profiles (NSMP) based on the Proactive Molecular Risk Classifier for Endometrial Cancer (PROMISE) classification. Results: Of 147 patients, 137 had sufficient tumor material for sequencing. ESR1 mutations were identified in eight tumors (6%), including Y537S/C/N (n = 4), L536H/P (n = 2), and E380Q (n = 2). A similar prevalence (3.5%) was found among 1,311 tumors in the Genomics England dataset. All ESR1 mutation cases were low-grade endometrioid endometrial cancer, ER-positive, and PR-positive, and classified as NSMP. Among patients with metastatic NSMP low-grade endometrioid endometrial cancer, 22% (8/37) harbored ESR1 mutations. Survival outcomes after platinum chemotherapy were similar between patients with ESR1 mutation endometrial cancer and ESR1 wild type (median, not reached vs. 25.3 months; P = 0.114). Conclusions: ESR1 mutations, while rare overall in treatment-naïve endometrial cancer, are more prevalent in patients with NSMP low-grade endometrioid endometrial cancer, potentially affecting AI efficacy. ESR1 status should be considered in selecting hormonotherapy and as a stratification factor in AI trials.

ctDNA for Prognostication and Monitoring in Patients with Metastatic Endometrial Carcinoma Treated with Olaparib: Validation in the GINECO-UTOLA Trial

Abstract Purpose: ctDNA may offer a noninvasive means to evaluate tumor response and anticipate disease dynamics before radiologic changes in advanced endometrial carcinoma. Experimental Design: This ancillary analysis included patients from the multicenter, randomized, phase II GINECO-UTerin OLAparib (UTOLA) trial (NCT03745950) evaluating olaparib/placebo as maintenance after first-line platinum-based chemotherapy. Plasma samples were collected at screening after chemotherapy (baseline), 3 months (M3), and progression. ctDNA detection was assessed by a validated methylation-based Droplet Digital PCR (MethddPCR) assay targeting DNA positions universally methylated in endometrial carcinoma. Results: Among 130 evaluable patients, ctDNA was detected in 25 of 129 (19%, 1 technical fail) at baseline, 15 of 80 (19%) at M3, and 33 of 52 (63%) at progression. Baseline ctDNA positivity was independently associated with poorer progression-free survival (PFS) [median 1.81 vs. 7.39 months; adjusted HR = 5.33 (3.17–8.97)] and overall survival (OS) [10.3 vs. 24.7 months; adjusted HR = 3.98 (2.28–6.91); adjusted for age, stage IV at diagnosis, p53abn subgroup, and residual measurable lesions after chemotherapy]. Patients with baseline ctDNA had median OS of 9.36 months under olaparib versus 19.6 months under placebo (log-rank P = 0.05). Patients with increasing ctDNA at M3 had median PFS of 1.67 months, versus 9.64 months without, and median OS of 18.8 versus 25.8 months. ctDNA rising was predictive of poor postprogression OS under olaparib but not under placebo (interaction test, P < 0.001). Conclusions: MethddPCR-ctDNA is an independent prognostic biomarker for OS in advanced/metastatic endometrial carcinoma. MethddPCR-ctDNA may identify patients unlikely to benefit from PARP inhibition, guide therapeutic decisions, and should be further evaluated as a new stratification parameter in future endometrial carcinoma trials.