With approximately 4 years of extended follow-up, pembrolizumab (Keytruda) plus carboplatin and paclitaxel continued to demonstrate an overall survival (OS) benefit over chemotherapy alone in both mismatch repair–deficient (dMMR) and mismatch repair–proficient (pMMR) advanced or recurrent endometrial cancer, according to updated results from the phase 3 NRG-GY018 trial (KEYNOTE-868; NCT03914612) presented at the 2026 ASCO Annual Meeting.1
In the dMMR cohort (n = 223), at a median follow-up of 49 months across both treatment arms, pembrolizumab plus chemotherapy reduced the risk of death by 44% compared with placebo plus chemotherapy (HR, 0.56; 95% CI, 0.34-0.92; log-rank P = .0124). In the pMMR cohort (n = 586), at a median follow-up of 44 months, OS was directionally favorable for pembrolizumab plus chemotherapy, with an HR of 0.86 (95% CI, 0.69-1.08; log-rank P = .1072). This corresponded to a median OS of 44.4 months (95% CI, 37.8-52.1) vs 35.1 months (95% CI, 30.2-43.8) with placebo.
Subgroup analyses of OS in both the dMMR and pMMR cohorts were generally consistent with the overall findings.
"Importantly, this OS benefit persisted despite substantial post-study immune checkpoint inhibitor [ICI] utilization in the placebo arms of both of these populations. This is the largest proportion of the placebo-treated patients who went on to receive a subsequent ICI strategy amongst all of the contemporary phase 3 trials in this clinical setting. Taken together, these data add great confidence to the established United States and international approvals of this regimen.” -Ramez Eskander, MD, presenter, gynecologic oncologist and professor of Obstetrics, Gynecology, and Reproductive Sciences at UC San Diego Health.
What prior data have been reported from NRG-GY018?
NRG-GY018 previously met its primary end point of progression-free survival (PFS), demonstrating statistically significant improvements with the addition of pembrolizumab to chemotherapy followed by pembrolizumab maintenance in patients with advanced or recurrent endometrial cancer, regardless of MMR status.1
In the dMMR cohort (n = 225), pembrolizumab plus chemotherapy reduced the risk of disease progression or death by 70% compared with placebo plus CP (HR, 0.30; 95% CI, 0.19-0.48; P < .001), and in the pMMR cohort (n = 591), it reduced the risk by 46% (HR, 0.54; 95% CI, 0.41-0.71; P < .001).2 Adverse effects were consistent with the known profiles of the individual agents. Of note, at the time of the primary analysis, OS data were immature but directionally favored pembrolizumab plus chemotherapy.1
Based on these results, on June 17, 2024, the FDA approved pembrolizumab with carboplatin and paclitaxel, followed by single-agent pembrolizumab, for adult patients with primary advanced or recurrent endometrial carcinoma.3 The regimen was simultaneously included in treatment guidelines for endometrial cancer in the United States and Europe.1
4-Year Results From NRG-GY018: OS Sustained Across MMR Status/Post-ICI
- At a median follow-up of 49 months in the dMMR cohort, pembrolizumab plus carboplatin and paclitaxel demonstrated a sustained OS benefit over carboplatin and paclitaxel alone (HR, 0.56; 95% CI, 0.34-0.92), with a 48-month OS rate of 78.6% (95% CI, 69.6%-85.2%) vs 60.4% (95% CI, 50.0%-69.2%).
- In the pMMR cohort, at a median follow-up of 44 months, the OS HR was directionally favorable at 0.86 (95% CI, 0.69-1.08), with a median OS of 44.4 months (95% CI, 37.8-52.1) vs 35.1 months (95% CI, 30.2-43.8) for placebo plus chemotherapy.
- The OS benefit in both cohorts persisted despite substantially greater post-study ICI use in the placebo plus chemotherapy arm compared with the pembrolizumab plus chemotherapy arm (93.2% vs 34.2% in the dMMR cohort; 81.1% vs 42.9% in the pMMR cohort).
How was NRG-GY018 designed?
NRG-GY018 was a multicenter, randomized, double-blind, placebo-controlled, phase 3 trial. Eligible patients had measurable stage III or IVA disease, measurable or nonmeasurable stage IVB disease, or recurrent EC, with a pathology report confirming institutional MMR immunohistochemistry results.1 Additional eligibility criteria included an ECOG performance-status score of 0 to 2 and no prior chemotherapy except prior adjuvant chemotherapy completed at least 12 months before study entry.
A total of 810 patients (pMMR, n = 588; dMMR, n = 222) were randomly assigned 1:1 to either:
- Arm 1: Intravenous (IV) placebo plus 175 mg/m² of IV paclitaxel and IV carboplatin at AUC 5 every 3 weeks for 6 cycles, followed by IV placebo every 6 weeks for up to 14 additional cycles
- Arm 2: IV pembrolizumab at 200 mg plus paclitaxel and carboplatin at the above dosages every 3 weeks for 6 cycles, followed by 400 mg of IV pembrolizumab every 6 weeks for up to 14 additional cycles
Patients were stratified by MMR status (dMMR vs pMMR), ECOG performance status (0 or 1 vs 2), and prior adjuvant chemotherapy (yes vs no). The study’s primary end point was PFS per RECIST 1.1 criteria by investigator assessment in both the pMMR and dMMR cohorts. The focus of this long-term follow-up analysis was updated OS and post-study ICI therapy in both cohorts.
Regarding the trial's statistical design, Eskander noted that OS was a predefined secondary end point without a hierarchical strategy or alpha allocation. "This trial was essentially 2 clinical trials in 1," he explained. "The dMMR and the pMMR populations had a separate statistical analysis plan, separate assumptions, and separate alpha allocations because we understood the biologic strength of checkpoint inhibition in a biomarker-directed dMMR population and wanted to have great confidence [in] the benefit if it existed in the pMMR cohort."
