NRG-GY018
Pembrolizumab versus placebo added to paclitaxel and carboplatin in stage III-IV or recurrent endometrial cancer (NRG-GY018, KEYNOTE-868)
Phase 3N = 816NCT03914612
- Population
- Newly diagnosed stage III or IVA endometrial cancer with measurable disease, or stage IVB or recurrent disease, of any histology except carcinosarcoma; previous adjuvant chemotherapy allowed if the chemotherapy-free interval was at least 12 months
- Design
- Double-blind, placebo-controlled phase 3 trial, randomised 1:1 within two independent cohorts defined by central immunohistochemistry (dMMR and pMMR). Paclitaxel 175 mg/m2 and carboplatin (AUC 5) with pembrolizumab 200 mg or placebo every 3 weeks for 6 cycles, then pembrolizumab 400 mg or placebo every 6 weeks for up to 14 cycles. Stratified by MMR status, ECOG performance status and previous adjuvant chemotherapy. ClinicalTrials.gov lists 813 participants; the NEJM paper reports 816 enrolled (225 in the dMMR cohort, 591 in the pMMR cohort) and the Nature Medicine paper 810.
- Primary endpoint
- PFS
- Arms
- Pembrolizumab plus paclitaxel-carboplatin (experimental): Pembrolizumab 200 mg plus paclitaxel 175 mg/m2 and carboplatin (AUC 5) every 3 weeks for 6 cycles, then pembrolizumab 400 mg every 6 weeks for up to 14 cycles
- Placebo plus paclitaxel-carboplatin (control): Placebo plus paclitaxel 175 mg/m2 and carboplatin (AUC 5) every 3 weeks for 6 cycles, then placebo every 6 weeks for up to 14 cycles
- People
- Eskander RN, NRG Oncology
PFSHR 0.30 (95% CI 0.19–0.48)Δ 12-month PFS+36 pointsPFSHR 0.54 (95% CI 0.41–0.71)Δ median PFS+4.4 monthsPFSHR 0.34 (95% CI 0.22–0.53)PFSHR 0.57 (95% CI 0.44–0.74)
Show the numbers
| Endpoint | Analysis | Pembrolizumab plus paclitaxel-carboplatin | Placebo plus paclitaxel-carboplatin | Effect |
|---|---|---|---|---|
| PFSdMMR cohort (centrally determined; analysis cohort of the NEJM report)Investigator-assessed progression-free survival according to RECIST 1.1 | primarymedian follow-up 12 monthsEskander 2023 | 74% at 12 mo | 38% at 12 mo | HR 0.30 (95% CI 0.19–0.48)P<0.001Kaplan-Meier estimates of freedom from progression or death at 12 months. |
| PFSpMMR cohort (centrally determined; analysis cohort of the NEJM report)Investigator-assessed progression-free survival according to RECIST 1.1 | primarymedian follow-up 7.9 monthsEskander 2023 | median 13.1 months | median 8.7 months | HR 0.54 (95% CI 0.41–0.71)P<0.001 |
| PFSdMMR population, intention-to-treat as stated in the Nature Medicine report (n = 222: 110 pembrolizumab, 112 placebo)Investigator-assessed progression-free survival according to RECIST 1.1 | interimmedian follow-up 14.4 monthsEskander 2025 | — | median 8.3 months (6.5–12.3) | HR 0.34 (95% CI 0.22–0.53)P<0.0001Median PFS was not reached in the pembrolizumab arm. The Nature Medicine report gives different numbers from the NEJM report (HR 0.30) for the same interim data cut-off; the paper states an intention-to-treat analysis with 29 events (26.4%) in the pembrolizumab group and 60 (53.6%) in the placebo group. |
| PFSpMMR population, intention-to-treat as stated in the Nature Medicine report (n = 588: 294 pembrolizumab, 294 placebo)Investigator-assessed progression-free survival according to RECIST 1.1 | interimmedian follow-up 10 monthsEskander 2025 | median 13.1 months (10.6–19.5) | median 8.7 months (8.4–11.0) | HR 0.57 (95% CI 0.44–0.74)P<0.0001The Nature Medicine report gives a different hazard ratio from the NEJM report (0.54) for the same interim data cut-off; events were 95 (32.3%) with pembrolizumab and 138 (46.9%) with placebo. |
| OSdMMR population (n = 222) | interimEskander 2025 | — | — | HR 0.55 (95% CI 0.25–1.19)P=0.0617 (1-sided, nominal)Immature: information fraction 18.0% (10 deaths with pembrolizumab, 17 with placebo). |
| OSpMMR population (n = 588) | interimEskander 2025 | — | — | HR 0.79 (95% CI 0.53–1.17)P=0.1157 (1-sided, nominal)Immature: information fraction 27.2% (45 deaths with pembrolizumab, 54 with placebo). |
| OSdMMR population (ad hoc analysis) | exploratorymedian follow-up 22.5 monthsEskander 2025 | — | — | HR 0.57 (95% CI 0.31–1.04)P=0.0323 (1-sided, nominal)Ad hoc analysis, information fraction 29.3% (17 deaths with pembrolizumab, 27 with placebo). |
| OSpMMR population (ad hoc analysis) | exploratorymedian follow-up 20.8 monthsEskander 2025 | — | — | HR 0.80 (95% CI 0.59–1.08)P=0.0683 (1-sided, nominal)Ad hoc analysis, information fraction 46.4% (77 deaths with pembrolizumab, 92 with placebo). |
| PFSdMMR population (n = 222)Progression-free survival per RECIST 1.1 by blinded independent central review | exploratoryEskander 2025 | — | median 14.1 months | HR 0.45 (95% CI 0.27–0.73)P=0.0005 (1-sided, nominal)Exploratory endpoint. Median PFS was not reached with pembrolizumab; the placebo-arm median has a 95% CI of 8.5 months to not reached, which is not encoded here. |
| PFSpMMR population (n = 588)Progression-free survival per RECIST 1.1 by blinded independent central review | exploratoryEskander 2025 | median 19.5 months (13.1–28.0) | median 11 months (9.0–11.5) | HR 0.64 (95% CI 0.49–0.85)P=0.0008 (1-sided, nominal)Exploratory endpoint. |
| Endpoint | Analysis | Pembrolizumab plus paclitaxel-carboplatin | Placebo plus paclitaxel-carboplatin | Effect |
|---|---|---|---|---|
| PFSdMMR cohort (centrally determined; analysis cohort of the NEJM report)Investigator-assessed progression-free survival according to RECIST 1.1 | primarymedian follow-up 12 monthsEskander 2023 | 74% at 12 mo | 38% at 12 mo | HR 0.30 (95% CI 0.19–0.48)P<0.001Kaplan-Meier estimates of freedom from progression or death at 12 months. |
| PFSpMMR cohort (centrally determined; analysis cohort of the NEJM report)Investigator-assessed progression-free survival according to RECIST 1.1 | primarymedian follow-up 7.9 monthsEskander 2023 | median 13.1 months | median 8.7 months | HR 0.54 (95% CI 0.41–0.71)P<0.001 |
| PFSdMMR population, intention-to-treat as stated in the Nature Medicine report (n = 222: 110 pembrolizumab, 112 placebo)Investigator-assessed progression-free survival according to RECIST 1.1 | interimmedian follow-up 14.4 monthsEskander 2025 | — | median 8.3 months (6.5–12.3) | HR 0.34 (95% CI 0.22–0.53)P<0.0001Median PFS was not reached in the pembrolizumab arm. The Nature Medicine report gives different numbers from the NEJM report (HR 0.30) for the same interim data cut-off; the paper states an intention-to-treat analysis with 29 events (26.4%) in the pembrolizumab group and 60 (53.6%) in the placebo group. |
| PFSpMMR population, intention-to-treat as stated in the Nature Medicine report (n = 588: 294 pembrolizumab, 294 placebo)Investigator-assessed progression-free survival according to RECIST 1.1 | interimmedian follow-up 10 monthsEskander 2025 | median 13.1 months (10.6–19.5) | median 8.7 months (8.4–11.0) | HR 0.57 (95% CI 0.44–0.74)P<0.0001The Nature Medicine report gives a different hazard ratio from the NEJM report (0.54) for the same interim data cut-off; events were 95 (32.3%) with pembrolizumab and 138 (46.9%) with placebo. |
| OSdMMR population (n = 222) | interimEskander 2025 | — | — | HR 0.55 (95% CI 0.25–1.19)P=0.0617 (1-sided, nominal)Immature: information fraction 18.0% (10 deaths with pembrolizumab, 17 with placebo). |
| OSpMMR population (n = 588) | interimEskander 2025 | — | — | HR 0.79 (95% CI 0.53–1.17)P=0.1157 (1-sided, nominal)Immature: information fraction 27.2% (45 deaths with pembrolizumab, 54 with placebo). |
| OSdMMR population (ad hoc analysis) | exploratorymedian follow-up 22.5 monthsEskander 2025 | — | — | HR 0.57 (95% CI 0.31–1.04)P=0.0323 (1-sided, nominal)Ad hoc analysis, information fraction 29.3% (17 deaths with pembrolizumab, 27 with placebo). |
| OSpMMR population (ad hoc analysis) | exploratorymedian follow-up 20.8 monthsEskander 2025 | — | — | HR 0.80 (95% CI 0.59–1.08)P=0.0683 (1-sided, nominal)Ad hoc analysis, information fraction 46.4% (77 deaths with pembrolizumab, 92 with placebo). |
| PFSdMMR population (n = 222)Progression-free survival per RECIST 1.1 by blinded independent central review | exploratoryEskander 2025 | — | median 14.1 months | HR 0.45 (95% CI 0.27–0.73)P=0.0005 (1-sided, nominal)Exploratory endpoint. Median PFS was not reached with pembrolizumab; the placebo-arm median has a 95% CI of 8.5 months to not reached, which is not encoded here. |
| PFSpMMR population (n = 588)Progression-free survival per RECIST 1.1 by blinded independent central review | exploratoryEskander 2025 | median 19.5 months (13.1–28.0) | median 11 months (9.0–11.5) | HR 0.64 (95% CI 0.49–0.85)P=0.0008 (1-sided, nominal)Exploratory endpoint. |
Data table
| Endpoint | HR | 95% CI |
|---|---|---|
| PFS | 0.3 | 0.19–0.48 |
| PFS | 0.54 | 0.41–0.71 |
| PFS | 0.34 | 0.22–0.53 |
| PFS | 0.57 | 0.44–0.74 |
| OS | 0.55 | 0.25–1.19 |
| OS | 0.79 | 0.53–1.17 |
| OS | 0.57 | 0.31–1.04 |
| OS | 0.8 | 0.59–1.08 |
| PFS | 0.45 | 0.27–0.73 |
| PFS | 0.64 | 0.49–0.85 |
- Pembrolizumab plus paclitaxel-carboplatin (experimental)
- Placebo plus paclitaxel-carboplatin (control)
Median-only curves assume a constant hazard (exponential) and fit plateaus, as seen with some immunotherapies, poorly.
Data table (modelled values)
| Months | Pembrolizumab plus paclitaxel-carboplatin | Placebo plus paclitaxel-carboplatin |
|---|---|---|
| 0 | 100.0% | 100.0% |
| 8 | 65.5% | 52.9% |
| 16 | 42.9% | 27.9% |
| 25 | 26.6% | 13.6% |
| 33 | 17.4% | 7.2% |
- Placebo plus paclitaxel-carboplatin (control)
Median-only curves assume a constant hazard (exponential) and fit plateaus, as seen with some immunotherapies, poorly.
Data table (modelled values)
| Months | Placebo plus paclitaxel-carboplatin |
|---|---|
| 0 | 100.0% |
| 5 | 65.9% |
| 10 | 43.4% |
| 16 | 26.3% |
| 21 | 17.3% |
Appears in
Sources
Eskander RN, Sill MW, Beffa L, Moore RG, Hope JM, Musa FB, et al. Pembrolizumab plus Chemotherapy in Advanced Endometrial Cancer. N Engl J Med. 2023;388(23):2159-2170.
Eskander RN, Sill MW, Beffa L, Moore RG, Hope JM, Musa FB, et al. Pembrolizumab plus chemotherapy in advanced or recurrent endometrial cancer: overall survival and exploratory analyses of the NRG GY018 phase 3 randomized trial. Nat Med. 2025;31(5):1539-1546.