The Lancet October 26, 2026

First-line pembrolizumab versus carboplatin plus paclitaxel in participants with mismatch repair-deficient advanced or recurrent endometrial cancer (KEYNOTE-C93/GOG-3064/ENGOT-en15): interim analysis of a multicentre, open-label, randomised, active-comparator controlled, phase 3 trial

Brian M Slomovitz, Jianqing Zhu, Rutie Yin, Floor Backes, Mariusz Bidziński, Lucy Gilbert, et al.

Bottom Line

In patients with mismatch repair-deficient advanced or recurrent endometrial cancer, first-line pembrolizumab monotherapy significantly improved progression-free survival and reduced severe toxicity compared with standard carboplatin plus paclitaxel, offering a highly active, chemotherapy-free option.

Key Findings

1. At a median follow-up of 36.7 months, median progression-free survival (PFS) was significantly prolonged with pembrolizumab monotherapy compared to chemotherapy (38.2 months vs 8.4 months; HR 0.57, 95% CI 0.42-0.79, p=0.0002).

2. Overall survival (OS) numerically favored pembrolizumab but did not reach statistical significance at the interim analysis (HR 0.77, 95% CI 0.52-1.15, p=0.10).

3. Grade 3 or higher treatment-related adverse events occurred in only 18.1% of patients receiving pembrolizumab, compared to 50.7% of those receiving chemotherapy.

4. Treatment-related adverse events leading to treatment discontinuation were less frequent with pembrolizumab (9.4%) than with chemotherapy (14.6%).

5. Patient-reported outcomes favored pembrolizumab, demonstrating clinically meaningful improvements in global health status/quality of life at week 12, whereas patients receiving chemotherapy experienced significant declines, particularly regarding fatigue, neuropathy, and body image (alopecia).

Study Design

Design

RCT

Open-Label

Sample

299

Patients

Duration

36.7 mo

Median

Setting

24 countries

Population Adult participants with centrally confirmed mismatch repair-deficient (dMMR) inoperable stage III/IV, persistent, or recurrent endometrial cancer who had not received prior systemic chemotherapy for advanced disease.

Intervention Intravenous pembrolizumab 400 mg every 6 weeks for up to 18 cycles.

Comparator Intravenous carboplatin (AUC 5 or 6) plus paclitaxel 175 mg/m² every 3 weeks for six cycles.

Outcome Progression-free survival (PFS) assessed by blinded independent central review (BICR) and overall survival (OS).

Study Limitations

• The open-label study design may introduce bias into the patient-reported quality of life metrics and investigator-assessed outcomes, though the primary PFS endpoint was assessed by blinded independent central review.

• Overall survival data remain immature and did not meet statistical significance at this interim analysis; interpretation is heavily confounded by crossover, as 85.2% of progressing patients in the chemotherapy arm subsequently received immunotherapy off-study.

• The trial compares pembrolizumab monotherapy to chemotherapy alone, rather than the contemporary standard of care (chemotherapy combined with concurrent and maintenance immunotherapy), leaving the relative efficacy of completely omitting chemotherapy versus giving it upfront uncertain.

Clinical Significance

KEYNOTE-C93 demonstrates that first-line pembrolizumab monotherapy is a highly active, chemotherapy-free alternative for mismatch repair-deficient (dMMR) advanced endometrial cancer. By more than quadrupling the median progression-free survival and substantially sparing patients from the toxicities of platinum-based chemotherapy—such as severe alopecia and neuropathy—it offers an excellent regimen. However, clinicians must carefully select candidates for this monotherapy by weighing its superb tolerability against the rapid early disease control that concurrent chemoimmunotherapy (the current standard) might offer for highly symptomatic patients.

Historical Context

For decades, the standard of care for advanced or recurrent endometrial cancer was cytotoxic chemotherapy, typically carboplatin plus paclitaxel. The treatment paradigm recently shifted with landmark trials like NRG-GY018 (pembrolizumab plus chemotherapy) and RUBY (dostarlimab plus chemotherapy), which demonstrated that combining immune checkpoint inhibitors with upfront chemotherapy dramatically improves outcomes, particularly in the mismatch repair-deficient (dMMR) population. Because dMMR tumors are highly immunogenic and uniquely sensitive to PD-1 blockade, KEYNOTE-C93 was initiated to determine whether these patients could safely skip the toxic chemotherapy backbone altogether and rely solely on first-line single-agent immunotherapy.

Guided Discussion

High-yield insights from every perspective

Med Student

Medical Student

How does mismatch repair deficiency (dMMR) in endometrial cancer lead to increased susceptibility to immune checkpoint inhibitors like pembrolizumab?

Key Response

dMMR results in the inability to repair DNA replication errors, leading to a high mutational burden and microsatellite instability (MSI-H). This generates numerous neoantigens that are recognized by the immune system. Pembrolizumab, a PD-1 inhibitor, blocks the inhibitory signal that tumors use to evade T-cells, thereby unleashing a robust, pre-existing immune response against these highly immunogenic tumors.

Resident

A patient presents with newly diagnosed advanced endometrial cancer. Based on the KEYNOTE-C93 trial, what specific molecular testing must be ordered on the biopsy, and how will it dictate your first-line systemic therapy discussion?

Key Response

MMR testing (via immunohistochemistry for MLH1, PMS2, MSH2, MSH6) is mandatory. If the tumor is dMMR, the patient can be offered first-line pembrolizumab monotherapy. This allows for a conversation about avoiding the classic toxicities of carboplatin/paclitaxel (alopecia, neuropathy, myelosuppression) while achieving superior progression-free survival, effectively offering a highly active, chemotherapy-free option.

Fellow

Recent trials like NRG-GY018 and RUBY demonstrated the benefit of adding checkpoint inhibitors to carboplatin/paclitaxel in advanced endometrial cancer. Given KEYNOTE-C93 shows pembrolizumab monotherapy is superior to chemotherapy alone, how should we conceptualize the choice between IO monotherapy versus IO plus chemotherapy for dMMR patients?

Key Response

This highlights a major current clinical debate. KEYNOTE-C93 suggests chemotherapy might be safely omitted in dMMR patients, avoiding significant toxicity. However, without head-to-head trials of IO alone versus IO+chemo, choosing between regimens requires weighing tumor bulk, symptom burden (where chemo might add an early debulking benefit before IO takes full effect), and patient comorbidities/frailty.

Attending

The transition from cytotoxic chemotherapy to first-line pembrolizumab monotherapy in dMMR endometrial cancer represents a major paradigm shift. How does this alter the toxicity profile we must anticipate and manage in our aging or frail gynecologic oncology patients?

Key Response

Attendings must pivot their clinics from managing acute chemotherapy toxicities to vigilant monitoring for immune-related adverse events (irAEs) like pneumonitis, colitis, and endocrinopathies. While overall severe toxicity rates are lower with IO monotherapy, making it highly advantageous for frail patients, irAEs can be unpredictable, permanent (e.g., hypothyroidism), or fatal if unrecognized, requiring systematic re-education of both staff and patients.

Scholarly Review

Critical appraisal through the lens of expert reviewers and guideline development

PhD

KEYNOTE-C93 is an open-label trial with an interim analysis reporting significant PFS improvements. In trials evaluating progression via RECIST in an open-label setting, what specific methodological safeguards must be in place to minimize investigator bias, and how might the interim analysis timing inflate the observed effect size?

Key Response

In open-label trials, knowledge of treatment assignment can bias subjective tumor measurements; therefore, a Blinded Independent Central Review (BICR) is essential for PFS endpoints. Furthermore, early stopping for efficacy at interim analyses can sometimes overestimate the true treatment effect (the 'winner's curse'), requiring rigorous alpha-spending functions (e.g., O'Brien-Fleming) and sufficient follow-up to ensure durability.

Journal Editor

Considering the rapid approval of IO+chemo regimens (based on GY018/RUBY) during the enrollment period of KEYNOTE-C93, how does the continued use of a chemotherapy-alone control arm impact the trial's ethical standing and the ultimate clinical applicability of its overall survival (OS) data?

Key Response

A critical reviewer would flag that the standard of care shifted to IO+chemo during the trial. If crossover to IO was not permitted or if the control arm is no longer standard practice, it complicates the interpretation of the trial's outcomes. While the trial proves IO is better than chemo, the field has already moved past chemo alone, challenging the direct real-world applicability of the comparator arm in modern settings.

Guideline Committee

Current NCCN guidelines strongly recommend adding pembrolizumab or dostarlimab to carboplatin/paclitaxel for advanced dMMR endometrial cancer. Based on the KEYNOTE-C93 results, should pembrolizumab monotherapy receive a Category 1 recommendation, and should it replace IO+chemo as the preferred regimen?

Key Response

The committee must evaluate whether to give pembrolizumab monotherapy a Category 1 recommendation as an alternative to IO+chemo. Given the lack of direct comparison between IO alone and IO+chemo, guidelines will likely list IO monotherapy as a preferred, less toxic alternative (especially for patients who are poor candidates for chemotherapy), while IO+chemo may remain an option for those with bulky, highly symptomatic disease needing rapid cytoreduction.

Clinical Landscape

Noteworthy Related Trials

2022

KEYNOTE-775 Trial

n = 827 · NEJM

Tested

Lenvatinib plus Pembrolizumab

Population

Patients with advanced endometrial cancer previously treated with systemic therapy

Comparator

Chemotherapy (doxorubicin or paclitaxel)

Endpoint

Progression-free survival and Overall survival

Key result: Lenvatinib plus pembrolizumab significantly improved progression-free survival and overall survival compared to chemotherapy.

2023

NRG-GY018 Trial

n = 810 · NEJM

Tested

Pembrolizumab plus Carboplatin and Paclitaxel

Population

Patients with advanced or recurrent endometrial cancer

Comparator

Placebo plus Carboplatin and Paclitaxel

Endpoint

Progression-free survival

Key result: The addition of pembrolizumab to standard chemotherapy significantly prolonged progression-free survival in both dMMR and pMMR cohorts.

2023

RUBY Trial

n = 494 · NEJM

Tested

Dostarlimab plus Carboplatin and Paclitaxel

Population

Patients with primary advanced or recurrent endometrial cancer

Comparator

Placebo plus Carboplatin and Paclitaxel

Endpoint

Progression-free survival

Key result: Dostarlimab plus chemotherapy yielded unprecedented progression-free survival benefits in the dMMR population.

Tailored to your role

Want this tailored to you?

Add your specialty or training stage to get role-specific takeaways and more questions.

Personalize this analysis