Introduction NSGO-OV-UMB1/ENGOT-OV-30 – cohort A is a single-arm, open-label phase II study of the combination of immune checkpoint inhibitors: durvalumab (anti-PD-L1) and oleclumab (anti-CD73), in relapsed high-grade serous ovarian cancer (HGSOC). The clinical efficacy data, presented at ESGO2021, showed the combination had effect, but the disease-control rate was not correlated with intratumoral CD8 and PD-L1 expression. Identification of responding patients by use of single cell-profiling through biomarker enrichment is needed to better select patients for immunotherapeutic strategies.

Methods Whole blood samples from the patients (n=25) were taken at regular intervals (pre-treatment, every 56 days, and at progression). Total leukocytes were isolated and fixed. Immunophenotyping with a 40 metal-tagged antibody panel with the ability to define multiple T-cell populations was done on cell suspensions on a CyTOF® XT mass cytometer. After data acquisition, the data was analyzed with a combination of R, Cytobank and MATLAB to identify predictive and response biomarkers.

Results Preliminary analysis identified 34 immune cell subsets present in all samples (n=37). Compared to the baseline samples, samples taken on day 56 of the treatment period contained higher proportions of classical monocytes (p=0.007), and lower proportions of central memory CD8⁺ T-cells (p=0.04) and effector memory CD4⁺ T-cells (p=0.0498). At baseline, the long-term survivors (≥16 weeks) demonstrated higher proportions of total T-cells (p=0.0302), total CD4⁺ T-cells (p=0.0221), and naïve CD4⁺ and CD8⁺ T-cells (p=0.0011 and p=0.0312, respectively).

Conclusions The analysis reveals immunological responses to durvalumab and oleclumab immunotherapy in patients with recurrent and metastatic HGSOC and suggests potential predictive biomarkers for categorizing patients into predefined response subgroups. Further investigation of both discoveries is underway.