Dr Eskander on Interpreting Biomarker Analyses From the DUO-E Trial in pMMR Endometrial Cancer

“The take-home message is that these data are principally important because they’re going to help us inform future drug development and potentially the design of future trials. [However,] we can’t take these data and use them to justify who we are or are not going to [administer] this treatment [to]. That’s not the intent of this analysis.”

Ramez N. Eskander, MD, an assistant professor in the Departments of Obstetrics, Gynecology, and Reproductive Sciences at the University of California (UC) San Diego School of Medicine; as well as the director for oncology of the clinical trials office at the UC San Diego Moores Cancer Center, discussed the need for careful interpretation of findings from a post hoc exploratory analysis of the phase 3 DUO-E trial (NCT04269200) evaluating biomarker and histological heterogeneity, the presence of baseline circulating tumor DNA, and efficacy in patients with mismatch repair–proficient (pMMR) recurrent endometrial cancer receiving olaparib (Lynparza) and durvalumab (Imfinzi) plus chemotherapy.

At the 2025 SGO Annual Meeting on Women’s Cancer, Kathleen N. Moore, MD, MS, of the Stephenson Cancer Center, presented data from this post hoc exploratory analysis. Results showed that in the pMMR subpopulation (n = 575), 84% of patients were positive for at least 1 biomarker. In biomarker-positive subgroups, the addition of durvalumab to chemotherapy demonstrated a consistent progression-free survival (PFS) benefit compared with chemotherapy alone. Furthermore, the addition of olaparib to durvalumab and chemotherapy showed a potential additive PFS benefit, with these effects observed across molecular subtypes.

FDA approvals exist for chemotherapy plus immunotherapy for the management of advanced-stage or recurrent pMMR endometrial cancer, but the DUO-E study comprised a heterogeneous patient population, Eskander began. Previously published efficacy data showed that the trial met its PFS end point, Eskander reported. However, several presentations since then have explored subgroups and integrated molecular data, which have introduced nuance and discordance, Eskander stated.

For example, DUO-E suggested PD-L1 expression may enhance response, but this was not observed in the phase 3 NRG-GY018 (NCT03914612) and RUBY (NCT03981796) trials, Eskander detailed. Similarly, conflicting signals regarding the role of TP53 mutations as predictive biomarkers were seen both across different DUO-E treatment arms and in DUO-E compared with RUBY, he added.

The challenge lies in interpreting these biomarkers when many patients fall into multiple subgroups, especially in post hoc analyses where confounding variables cannot be controlled, Eskander continued. As such, the relevance of individual biomarkers remains uncertain, he stated. These data are valuable for guiding future drug development and trial design but are not sufficient to inform clinical decision-making on patient selection for current treatment regimens, he emphasized, adding that doing so could lead to false discovery and inappropriate clinical conclusions.