Abstract
Background
Pembrolizumab–lenvatinib (P/L) is a standard option for patients with advanced or metastatic endometrial cancer (a/m EC) progressing after platinum-based chemotherapy, based on the KEYNOTE-775 trial. However, concerns remain regarding its real-world applicability, particularly in older, frailer patients and in aggressive histological subtypes underrepresented or excluded from pivotal trials. This study aimed to evaluate the effectiveness and tolerability of P/L in routine practice through the French early access program (EAP).
Patients and methods
We conducted a multicenter retrospective analysis of patients treated within the national EAP between March 2022 and December 2023. Eligible patients had histologically confirmed a/m EC progressing after platinum. Pembrolizumab was administered at 200 mg every 3 weeks, and lenvatinib was initiated at up to 20 mg daily. Progression-free survival (PFS) and overall survival (OS) were estimated using Kaplan–Meier methods, and prognostic factors explored with Cox regression.
Results
A total of 351 patients from 25 centers were included [median age 73 years; 21.5% Eastern Cooperative Oncology Group (ECOG) ≥2]. Carcinosarcomas represented 8.9% of cases. Median PFS was 6.0 months [95% confidence interval (CI) 5.7-8.3], and median OS was 21 months (95% CI 18.6-22.9). Mismatch repair-deficient status was strongly associated with prolonged PFS (16 versus 6 months, P < 0.001). Objective response rate at 6 months was 41%, with a clinical benefit rate of 68%. Independent adverse prognostic factors for PFS included ECOG >1, multiple metastatic sites, microsatellite stable status, and carcinosarcoma subtype. Reduced starting lenvatinib dose (37.2% of patients) did not adversely impact PFS. Permanent discontinuation of lenvatinib because of toxicity was observed in 96 patients (27.6%).
Conclusions
P/L demonstrates meaningful activity in a broad real-world population, including older patients with more aggressive disease features and those starting at reduced lenvatinib doses. These findings support the sustained benefit of this combination beyond clinical trial settings, provided that lenvatinib is carefully monitored and doses are personalized.
Key words: elderly population, endometrial cancer, pembrolizumab–lenvatinib, real-world efficacy data, reduced initiated dose impact, safety data
Highlights
-
•
P/L is standard of care after platinum failure in immunotherapy-naïve advanced EC (KEYNOTE-775).
-
•
KEYNOTE-775 raises real-world concerns due to toxicity and a highly selected population.
-
•
LARENA real-world data show clinical benefits in older patients.
-
•
Reduced starting doses of lenvatinib (37.2%) did not impair PFS.
-
•
Lower starting doses led to fewer dose reductions for AEs versus KEYNOTE-775 (59.7% versus 68.7%).
Introduction
Endometrial cancer (EC) is the fourth most common cancer in women worldwide, with ∼420 000 new cases and 97 000 deaths reported annually.1 Its incidence is steadily increasing, largely driven by aging populations and rising obesity rates.2 However, this trend is not homogeneous across Europe: although incidence continues to rise in several Southern and Eastern European countries (e.g. Slovakia and Slovenia) and in postmenopausal women in many Northern and Western European countries, a stabilization or decline has been observed in others, including Denmark, France, and Switzerland, particularly among premenopausal and perimenopausal women, possibly reflecting the wider use of combined oral contraceptives.3,4 In The Netherlands, incidence of the endometrioid subtype peaked in 2004 before declining, whereas nonendometrioid subtypes have continued to increase.5 For patients with advanced or metastatic endometrial cancer (a/m EC), first-line treatment currently consists of carboplatin–paclitaxel chemotherapy, with or without the addition of immunotherapy depending on molecular characteristics.6,7
Following progression after platinum-based chemotherapy, the combination of pembrolizumab, a programmed cell death protein 1 inhibitor, and lenvatinib, a multitargeted tyrosine kinase inhibitor, has become a standard therapeutic option for immunotherapy-naïve patients. This recommendation is based on the results of the phase III KEYNOTE-775 trial, which demonstrated a significant improvement in both median progression-free survival (mPFS) [7.2 versus 3.8 months, hazard ratio (HR) 0.56, 95% confidence interval (CI) 0.47-0.66, P < 0.001] and median overall survival (mOS) (18.3 versus 11.4 months, HR 0.62, 95% CI 0.51-0.75, P < 0.001) compared with standard chemotherapy, regardless of mismatch repair (MMR) status.8
However, the KEYNOTE-775 results have raised concerns regarding the real-world applicability of this combination. Firstly, the toxicity profile was substantial, with grade ≥3 adverse events (AEs) reported in 88.9% of patients, lenvatinib dose reductions required in 66.5%, and treatment discontinuation occurring in 33% of cases. Secondly, the trial population was highly selected, with a relatively young median age (64 years) and the exclusion of aggressive histological subtypes such as carcinosarcomas, potentially limiting the generalizability of the findings.
In France, a temporary early access program (EAP) was implemented between March 2022 and December 2023 to provide pembrolizumab–lenvatinib (P/L) to patients with advanced or metastatic EC. The present study aims to evaluate real-world management practices, as well as the efficacy and tolerability of P/L, using data from this French EAP in a heterogeneous population representative of routine clinical and clinicopathological profiles.
Patients and methods
Patients and treatment
The French EAP was accessible to all patients and centers in France. To be eligible, the following criteria were mandatory: women aged ≥18 years; treated for a histologically proven advanced or recurrent endometrial cancer whose disease was progressing during or following prior platinum-based chemotherapy at any stage and who were not eligible for curative surgery or radiotherapy. All patients treated as part of the EAP were prospectively registered at their site in a national registry. Lenvatinib was delivered to patients by their hospital pharmacy every month, at a maximum starting dose of 20 mg daily. Pembrolizumab was used at standard doses (200 mg i.v. every 3 weeks). Treatment and dose reductions were managed by oncologists, following drug labels.
Data collection
This study was approved by the Institut Curie Institutional Review Board, which granted a waiver of informed consent owing to the retrospective nature of the study. A synopsis of the study was distributed to French cancer centers within the ARCAGY-GINECO academic cooperative group. Twenty-five centers—including university hospitals, comprehensive cancer centers, and private hospitals within the network—agreed to participate.
Medical oncologists at participating sites retrospectively collected data from electronic medical records for all patients enrolled in the EAP at their institutions. AEs were graded according to the Common Terminology Criteria for Adverse Events (U.S Department of Health and Human Services, National Institutes of Health, National Cancer INstitute, Bethesda. MD, version 5.0), and tumor responses were assessed by imaging using the RECIST v1.1 (Oxford. United Kingdom).
The data cut-off date was 31 December 2023. Pseudonymized individual patient data were subsequently manually reviewed at Institut Curie for quality control and internal consistency, and data queries were issued to participating centers when necessary.
Statistics
Baseline patient characteristics (demographics and medical history) were described. Qualitative variables were summarized using counts, percentages, and the number of missing values. Quantitative variables were described using the number of observations, mean, standard deviation, median, interquartile range (25th-75th percentiles), minimum, maximum, and the number of missing values.
Simple comparisons of categorical variables were carried out using the chi-square test or Fisher’s exact test (two-tailed), as appropriate.
Progression-free survival (PFS) was defined as the time from the first administration of lenvatinib plus pembrolizumab to the date of the first documented disease progression according to RECIST 1.1 or death from any cause. Patients without an event were censored at the date of last follow-up. Overall survival (OS) was defined as the time from the first administration of lenvatinib plus pembrolizumab to the date of death from any cause or the date of last follow-up.
PFS and OS were estimated using the Kaplan–Meier method. Estimated survival rates at times of interest are presented with their 95% CIs. Comparisons of PFS or OS curves according to patient status were carried out using the log-rank test. Median follow-up was estimated using the reverse Kaplan–Meier method.
Univariable and multivariable analyses were conducted using Cox proportional hazards regression models. The following potential prognostic factors for PFS were explored: age at initiation of P/L, Eastern Cooperative Oncology Group (ECOG) performance status, histological subtype, MMR status determined by immunohistochemistry (IHC) or molecular testing, number of metastatic sites, initial lenvatinib dose, and prior (neo)adjuvant platinum exposure.
All explanatory variables associated with the outcome at a conservative P value threshold of 0.20 (P < 0.20) were included in the multivariable model, except for age and prior (neo)adjuvant platinum exposure, which were included regardless of significance given its established importance in the literature. A backward selection procedure was then applied to remove nonsignificant variables (P ≥ 0.05).
The proportional hazards assumption was assessed using Schoenfeld residuals. To evaluate the robustness of the analyses, a sensitivity analysis was carried out using multiple imputation with the R package Amelia (Harvard University, Cambridge, Massachusetts). Ten imputed datasets (M = 10) were generated and the same Cox regression models were applied to each dataset. Estimates were combined using Rubin’s rules to obtain pooled results. Analyses will be carried out using R software (Vienna. Austria, version 4.4.1).
Results
Patients’ and tumor characteristics
A total of 351 patients from 25 participating centers met the study eligibility criteria and were included in the analysis. Baseline demographic and clinicopathological characteristics are summarized in Table 1.
Table 1.
Patients’ characteristics in LARENA and KEYNOTE 775 trial
| Variable | LARENA (N = 351) | KEYNOTE-775—lenvatinib + pembrolizumab arm (n = 411) |
|---|---|---|
| Age, median (min, max) | 73 (27-90) | 63 (22-93) |
| ECOG | ||
| 0/1 | 244 (78.5%) | 411 (100%)a |
| ≥2 | 67 (21.5%) | 0 (excluded by protocol) |
| NA | 20 | 0 |
| Number of metastatic sites | ||
| 0/1 | 79 (24.1%) | NR |
| ≥2 | 249 (75.9%) | NR |
| Prior treatments for advanced cancer | ||
| 0 | 47 (14.3%) | N/A (≥1 prior platinum line required) |
| 1 | 151 (45.7%) | NR |
| ≥2 | 132 (40%) | NR |
| NA | 1 | NR |
| Type of metastatic sites | ||
| Lymph nodes | 208 (60.5%) | NR |
| Peritoneum | 180 (52.3%) | NR |
| Lung or pleura | 168 (49.3%) | NR |
| Liver | 93 (27.3%) | NR |
| Bone | 61 (17.9%) | NR |
| Prior immunotherapy exposure | 9 (2.6%) | 0 (excluded by protocol) |
| Histological subtypes | ||
| Low-grade endometrioid carcinoma | 79 (22.6%) | 138 (33.6%)b |
| High-grade endometrioid carcinoma | 66 (18.9%) | 136 (33.1%) |
| Endometrioid carcinoma of unspecified grade | 24 (6.9%) | – |
| Serous carcinoma | 79 (22.6%) | 75 (18.3%) |
| Clear cell carcinoma | 15 (4.3%) | 12 (2.9%) |
| Carcinosarcoma | 31 (8.9%) | 0 (excluded by protocol) |
| Mixed features | 26 (7.4%) | 28 (6.9%) |
| Other | 30 (8.6%) | 21 (5.2%) |
| NA | 1 | NR |
| MMR status (by IHC) | ||
| dMMR | 46 (15.6%) | 65 (15.8%) |
| pMMR | 248 (84.4%) | 346 (84.2%) |
| NA | 57 | 0 |
| p53-mutated profile (by IHC) | ||
| No | 134 (43.8%) | NR |
| Yes | 172 (56.2%) | NR |
| NA | 45 | NR |
| Time from the last administration of platinum-based chemotherapy (median, months, Q1, Q3) | 6 (3,12) | NR |
| Platinum-based neoadjuvant therapy | ||
| No | 319 (90.9%) | NR |
| Yes | 32 (9.1%) | NR |
| Platinum-based adjuvant therapy | ||
| No | 104 (29.6%) | NR |
| Yes | 247 (70.4%) | NR |
| Prior pelvic irradiation | ||
| No | 171 (48.7%) | 235 (57.2%) |
| Yes | 180 (51.3%) | 176 (42.8%) |
| Lenvatinib dose initiation | ||
| Standard dose initiation (20 mg) | 204 (62.8%) | 411 (100%)c |
| Reduced dose initiation (<20 mg) | 121 (37.2%) | 0 |
Source: Makker et al.8
ECOG, Eastern Cooperative Oncology Group; dMMR, mismatch repair-deficient; IHC, immunohistochemistry; MMR, mismatch repair; NR, not reported in the primary KEYNOTE-775 publication; N/A, not applicable (protocol eligibility criterion); pMMR, proficient mismatch repair.
KEYNOTE-775 enrolled only ECOG 0 of 1 patients (ECOG 0: 59.9%; ECOG 1: 40.1%).
Combines low-grade and unspecified-grade endometrioid carcinoma (‘nonhigh-grade’ category).
Fixed lenvatinib starting dose of 20 mg per protocol (early dose reduction at initiation was not permitted).
Proficient mismatch repair (pMMR) was identified by IHC in 248 patients (84.4%), whereas 46 patients (15.6%) had mismatch repair-deficient (dMMR) disease. The most frequent histological subtypes were low-grade endometrioid carcinoma (22.6%), high-grade endometrioid carcinoma (18.9%), serous carcinoma (22.6%), carcinosarcoma (8.9%), and clear cell carcinoma (4.3%). A p53-abnormal profile, as assessed by IHC, was observed in 172 patients (56.2% of those with known p53 IHC status). Among the 265 assessable patients, only 11 (4.1%) exhibited a ‘double-classifier’ profile combining a p53-mutated profile and dMMR status, further supporting the enrichment of this cohort in high-risk histological phenotypes. Indeed, p53 abnormality was significantly associated with histological subtype (P < 0.001) and was numerically more frequent in carcinosarcomas and serous carcinomas. The median age at treatment initiation was 73 years (range 27-90). ECOG performance status was 0-1 in 244 patients (78.5%) and ≥2 in 67 patients (21.5%).
Most patients presented with more than one metastatic site (75.9%), the most common locations being lymph nodes (60.5%), peritoneum (52.3%), and lung or pleura (49.3%). One hundred fifty-one patients (45.7%) had received one prior line of systemic therapy and the median time since the last platinum-based chemotherapy regimen was 6 months (Q1 = 3, Q3 = 12).
Lenvatinib was initiated at a daily dose of 20 mg in 215 patients (62.3%), 14 mg in 77 patients (22.3%), and 10 mg in 52 patients (15.1%) (Table 2).
Table 2.
Treatment dose management, overall, and according to age/ECOG performance status subgroups
| Variable | Overall (N = 351) | <70 years and ECOG 0/1 (n = 105) | <70 years and ECOG ≥2 (n = 19) | ≥70 years and ECOG 0/1 (n = 151) | ≥70 years and ECOG ≥2 (n = 56) | P valuea |
|---|---|---|---|---|---|---|
| Pembrolizumab | ||||||
| Dose initiation | 200 mg (all patients) | – | – | – | – | – |
| Treatment interruption, n (%)b | 58 (16.7) incl. 48 (82.8) for toxicity | – | – | – | – | – |
| Permanent discontinuation, n (%)b | 235 (67.1) Incl. 30 (12.8) for toxicity | – | – | – | – | – |
| Lenvatinib | ||||||
| Dose initiation, n (%) | ||||||
| 20 mg | 215 (62.3) | 76 (74.5) | 12 (63.2) | 87 (58.4) | 24 (43.6) | 0.001 |
| 14 mg | 77 (22.3) | 11 (10.8) | 3 (15.8) | 40 (26.8) | 19 (34.5) | |
| 8/10 mgc | 53 (15.4) | 15 (14.7) | 4 (21.1) | 22 (14.8) | 12 (21.8) | |
| NA | 6 | 3 | 0 | 2 | 1 | |
| Dose reduction (during treatment), n (%) | ||||||
| No | 140 (40.3) | 42 (41.2) | 10 (52.6) | 53 (35.3) | 28 (50.0) | 0.17 |
| Yes | 207 (59.7) | 60 (58.8) | 9 (47.4) | 97 (64.7) | 28 (50.0) | |
| NA | 4 | 3 | 0 | 1 | 0 | |
| Treatment interruption, n (%)b | 198 (57.1) Incl. 165 (83.8) for toxicity | – | – | – | – | – |
| Permanent discontinuation, n (%)b | 265 (76.6) Incl. 96 (36.2) for toxicity | – | – | – | – | – |
Age/ECOG subgroup analysis restricted to the 331 patients with known age and ECOG performance status at treatment initiation (20 patients with missing ECOG status excluded).
ECOG, Eastern Cooperative Oncology Group.
P value for comparison across the four age/ECOG subgroups (Fisher’s exact test with simulated P value, 2000 replicates); not computed for variables without subgroup breakdown.
Not assessed by age/ECOG subgroup.
Combines 8 mg (n = 1, 0.3%) and 10 mg (n = 52, 15.1%) initiation doses in the overall cohort, to match the granularity of the age/ECOG subgroup analysis; 16 mg was not used as an initiation dose (n = 0).
Treatment efficacy
At the data cut-off date, the median follow-up was 14 months (range 0.2-19.5). In the overall population, mPFS was 6.0 months (95% CI 5.7-8.27) (Figure 1A). PFS was significantly longer in patients with dMMR compared with those with pMMR disease (16 versus 6 months, P < 0.001) (Supplementary Figure S1A, available at https://doi.org/10.1016/j.esmoop.2026.108585). At database lock, 137 deaths had occurred, yielding a mOS of 21 months (95% CI 18.6-22.9) (Figure 1B). dMMR status did not significantly impact OS (Supplementary Figure S1B, available at https://doi.org/10.1016/j.esmoop.2026.108585). Among 295 patients with measurable disease, best overall responses included complete response in 26 patients (8.8%), partial response (PR) in 95 patients (32.2%), and stable disease (SD) in 81 patients (27.5%) (Supplementary Figure S2, available at https://doi.org/10.1016/j.esmoop.2026.108585). At 6 months, the objective response rate was 41% (95% CI 35% to 47%) and the clinical benefit rate was 68% (95% CI 63% to 74%).
Figure 1.
(A) Progression-free survival in the overall population. (B) Overall survival in the overall population.
Prognostic factors for PFS, objective response rate, and clinical benefit rate
Univariate and multivariate analyses for PFS are presented in Figure 2A and B. In multivariate analysis, independent adverse prognostic factors for PFS were ECOG performance status >1 (HR 1.72, 95% CI 1.19-2.50), more than one metastatic site (HR 1.67, 95% CI 1.21-2.3), and carcinosarcoma subtype (HR 1.86, 95% CI 1.13-3.07), whereas dMMR status was a favorable prognostic factor (HR 0.51, 95% CI 0.34-0.77). In a four-category composite analysis combining age (<70 versus ≥70 years) and ECOG performance status (0-1 versus ≥2), Kaplan–Meier analysis of PFS revealed significant differences across the four subgroups (global log-rank P = 0.004; Supplementary Figure S3, available at https://doi.org/10.1016/j.esmoop.2026.108585), confirming that impaired performance status (ECOG ≥2), rather than chronological age, is associated with poorer PFS. The administration of P/L as first-line treatment in the metastatic setting following prior exposure to platinum-based chemotherapy in the neoadjuvant and/or adjuvant setting did not significantly impact PFS (Figure 2B). Regarding ORR, the number of metastatic sites was the only factor significantly associated with tumor response (P = 0.03), whereas age and ECOG status had no significant impact. Regarding clinical benefit rate, both ECOG performance status (P = 0.03) and the number of metastatic sites (P < 0.01) were significantly associated with achieving a durable clinical benefit (Supplementary Figure S4A and B, available at https://doi.org/10.1016/j.esmoop.2026.108585).
Figure 2.
(A) Univariate analysis for PFS. (B) Multivariate analysis for PFS.
CI, confidence interval; dMMR, mismatch repair-deficient; ECOG PS, Eastern Cooperative Oncology Group performance status; HR, hazard ratio; PFS, progression-free survival; pMMR, proficient mismatch repair.
Notably, reduced lenvatinib starting dose did not significantly impact PFS. Indeed, the mPFS was 6.1 months (95% CI 5.36-7.81) without dose reduction (20 mg) versus 7.71 months (95% CI 5.74-10.77) with dose reduction (8/10/14 mg) (P = 0.3) (Supplementary Figure S5, available at https://doi.org/10.1016/j.esmoop.2026.108585).
Among the nine patients previously treated with immunotherapy, whose prior exposure was described as part of an exploratory descriptive analysis, three achieved a PR, three had SD, and three were not assessable.
Safety
Temporary discontinuation of lenvatinib occurred in 198 patients (57.1%), of whom 165 (83.8%) discontinued due to treatment-related toxicity. Permanent discontinuation of lenvatinib because of toxicity was observed in 96 patients (27.6%). Overall, 207 patients (59.7%) required at least one lenvatinib dose reduction.
Temporary discontinuation of pembrolizumab occurred in 58 patients (16.7%), including 48 cases (82.8%) related to toxicity. Permanent discontinuation of pembrolizumab was reported in 235 patients (67.1%), of whom 30 (12.8%) discontinued due to toxicity (Table 2).
The most frequent grade 3-4 AEs were hypertension (10.9%), fatigue (5%), diarrhea (4.7%), anorexia (4.3%), arthralgia (2.6%), and weight loss (2.2%). No treatment-related deaths were reported (Table 3).
Table 3.
CTCAE 5.0 AEs
| Adverse events | Patients without eventsa | Grades |
|||
|---|---|---|---|---|---|
| 1a | 2a | 3a | 4a | ||
| Anemia | 324 (96) | 2 (0.6) | 7 (2.1) | 5 (1.5) | 0 (0) |
| Appetite | 244 (74) | 31 (9.5) | 39 (12) | 14 (4.3) | 0 (0) |
| Arthralgia | 284 (83) | 22 (6.5) | 26 (7.6) | 9 (2.6) | 0 (0) |
| Constipation | 313 (92) | 24 (7.0) | 5 (1.5) | 0 (0) | 0 (0) |
| Diarrhea | 220 (65) | 67 (20) | 37 (11) | 16 (4.7) | 1 (0.3) |
| Fatigue | 146 (45) | 73 (23) | 88 (27) | 16 (5.0) | 0 (0) |
| Hypertension | 227 (69) | 19 (5.7) | 49 (15) | 33 (10.0) | 3 (0.9) |
| Hyperthyroidism | 324 (94) | 13 (3.8) | 8 (2.3) | 0 (0) | 0 (0) |
| Hypothyroidism | 197 (63) | 46 (15) | 71 (23) | 1 (0.3) | 0 (0) |
| Nausea | 282 (83) | 37 (11) | 17 (5.0) | 4 (1.2) | 0 (0) |
| Neutropenia | 339 (97) | 3 (0.9) | 3 (0.9) | 2 (0.6) | 1 (0.3) |
| Proteinuria | 306 (93) | 11 (3.3) | 9 (2.7) | 4 (1.2) | 0 (0) |
| Vomiting | 318 (94) | 13 (3.8) | 8 (2.4) | 1 (0.3) | 0 (0) |
| Weight decrease | 214 (66) | 57 (18) | 46 (14) | 7 (2.2) | 0 (0) |
AEs, adverse events; CTCAE, Common Terminology Criteria for Adverse Events.
n (%).
Discussion
In this large real-world cohort derived from the French EAP, P/L demonstrated clinically meaningful activity in patients with a/m EC, confirming the efficacy observed in the pivotal KEYNOTE-775 trial8 while providing important insights into its use in a less selected population.
In KEYNOTE-775, P/L significantly improved PFS and OS compared with physician’s choice chemotherapy, with a mPFS of 7.2 months and a median OS of 18.3 months in the overall population. The 5-year follow-up confirmed a significant advantage of P/L over chemotherapy, with improved PFS and OS rates of 9.8% versus 3.3% (HR 0.56, 95% CI 0.48-0.66) and 19.9% versus 7.7% (HR 0.66, 95% CI 0.57-0.77), respectively.9 In our real-world cohort, mPFS was 6.0 months and median OS reached 21 months, suggesting comparable outcomes, as supported by the overlapping 95% CIs.8 OS may also have been influenced by postprogression treatments. The slightly shorter PFS observed in our study may reflect the inclusion of patients with poorer prognostic features, including older age (median age 73 versus 64 years), worse performance status (over 20% of patients presenting with ECOG performance status ≥2), a high metastatic burden, and aggressive histological subtypes that were excluded or under-represented in KEYNOTE-775, with the example of carcinosarcomas that represented nearly 9% of cases, whereas these tumors were excluded from KEYNOTE-775. These differences highlight the broader clinical spectrum captured in routine practice and reinforce the external validity of our findings.
Importantly, the magnitude of benefit according to dMMR status was consistent with trial findings. In our cohort, dMMR status was strongly associated with prolonged PFS, reaching a median of 16 months, and remained an independent prognostic factor in multivariate analysis. Interestingly, dMMR status did not translate into a substantial OS advantage, likely reflecting the impact of subsequent treatment lines including immunotherapy in this subgroup. Notably, a reduced starting dose of lenvatinib did not adversely impact PFS, suggesting flexibility in treatment initiation without compromising efficacy. Of note, the carcinosarcoma subtype was identified as an independent negative prognostic factor.
In routine care, toxicity remains a major challenge with the P/L combination. In KEYNOTE-775, the incidence of grade ≥3 AEs was 88.9%, with high rates of dose reductions and treatment discontinuations. In our real-world cohort, dose reductions due to AEs occurred less frequently (59.7%) than in the KEYNOTE-775 (68.7%),9 reflecting both the impact of proactive toxicity management strategies and the lower dose initiation of lenvatinib. Moreover, permanent discontinuation due to toxicity occurred in a relatively limited proportion of patients, and no treatment-related deaths were reported.
These findings suggest that careful dose adaptation and close monitoring allow safe administration of this regimen in routine practice, including in older patients. The absence of a negative impact of lower starting doses on PFS, used in 37.2% of patients, may be explained by fewer subsequent lenvatinib dose reductions, resulting in an overall exposure similar to the 13.4 mg/day reported in KEYNOTE-775.9 These findings, consistent with other real-world data,10, 11, 12, 13, 14 further support individualized dosing strategies guided by pharmacological monitoring, allowing a better balance between efficacy and safety. When such monitoring is not feasible, an initial dose of 14 mg/day may be considered, as it most closely approximates the median dose exposure reported in KEYNOTE-775.
In the era of increasing use of immunotherapy in combination with carboplatin–paclitaxel in the first-line metastatic setting,6,7 together with the rapidly expanding role of antibody–drug conjugates, P/L remains a relevant treatment option as initial therapy for advanced/recurrent disease in patients experiencing early relapse after prior platinum-based (neo)adjuvant chemotherapy. This approach is supported by the PFS and OS benefits reported in the corresponding subgroup of the LEAP-001 trial (PFS HR 0.52-0.60, OS HR 0.64-0.67)15 and is consistent with our real-world findings, in which the use of P/L following prior (neo)adjuvant platinum exposure was not associated with poorer PFS.
Although KEYNOTE-775 included only immunotherapy-naïve patients, our cohort included a small subset of patients previously exposed to immune checkpoint inhibitors. Given the limited number of patients in this subgroup, no firm conclusions can be drawn regarding the efficacy of P/L after prior immunotherapy exposure. Although responses were observed in some patients, these findings should be interpreted with caution. Further studies in larger cohorts are needed to determine whether P/L retains meaningful clinical activity in this setting.16,17
The strengths of this study include its large sample size, nationwide multicenter design, and inclusion of a heterogeneous population representative of real-world clinical practice. Limitations include its retrospective nature, potential missing data, and lack of centralized radiological review. Nevertheless, sensitivity analyses carried out after multiple imputation of missing data on prognostic factors yielded consistent results, confirming the robustness of the main model estimates. Additionally, OS analyses may be confounded by subsequent therapies.
Overall, our findings confirm the effectiveness of P/L in routine clinical practice and extend the applicability of KEYNOTE-775 results to a broader population, including elderly patients, those with poor performance status, and aggressive histological subtypes. These real-world data support P/L as a robust therapeutic option after platinum failure and underscore the importance of personalized toxicity management to optimize outcomes.
Acknowledgements
We would like to thank Institut des Cancers des Femmes by Institut Curie and ARCAGY-GINECO for sponsoring this study. We declare that Eisai provided financial support for the statistical budget.
Funding
This work was supported by Eisai, which provided partial financial support to Institut Curie for the statistical analyses of the LARENA study.
Disclosure
The authors have declared no conflicts of interest.
Supplementary data
References
- 1.Global Cancer Observatory. https://gco.iarc.fr/en Available at.
- 2.Rahib L., Wehner M.R., Matrisian L.M., Nead K.T. Estimated projection of US cancer incidence and death to 2040. JAMA Netw Open. 2021;4(4) doi: 10.1001/jamanetworkopen.2021.4708. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Bray F., Dos Santos Silva I., Moller H., Weiderpass E. Endometrial cancer incidence trends in Europe: underlying determinants and prospects for prevention. Cancer Epidemiol Biomarkers Prev. 2005;14(5):1132–1142. doi: 10.1158/1055-9965.EPI-04-0871. [DOI] [PubMed] [Google Scholar]
- 4.Rodríguez-Palacios D.Á., Colorado-Yohar S.M., Velten M., Vaamonde-Martín R.J., Ballesta M., Chirlaque M.-D. Incidence and trend of type I and II endometrial cancer in women from two population-based European cancer registries (1998–2012) Int J Environ Res Public Health. 2022;19(7):3789. doi: 10.3390/ijerph19073789. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Van Weelden W.J., Bretveld R., Romano A., et al. Trends over time in the incidence and use of hormonal therapy in endometrial cancer: a population-based study in the Netherlands. Int J Gynecol Cancer. 2021;31(7):1014–1020. doi: 10.1136/ijgc-2021-002444. [DOI] [PubMed] [Google Scholar]
- 6.Mirza M.R., Chase D.M., Slomovitz B.M., et al. Dostarlimab for primary advanced or recurrent endometrial cancer. N Engl J Med. 2023;388(23):2145–2158. doi: 10.1056/NEJMoa2216334. [DOI] [PubMed] [Google Scholar]
- 7.Eskander R.N., Sill M.W., Beffa L., et al. Pembrolizumab plus chemotherapy in advanced endometrial cancer. N Engl J Med. 2023;388(23):2159–2170. doi: 10.1056/NEJMoa2302312. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Makker V., Colombo N., Casado Herráez A., et al. Lenvatinib plus pembrolizumab for advanced endometrial cancer. N Engl J Med. 2022;386(5):437–448. doi: 10.1056/NEJMoa2108330. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Makker V., Colombo N., Casado A., et al. Lenvatinib plus pembrolizumab in previously treated advanced endometrial cancer: 5-year outcomes from the randomized, phase 3 study 309/KEYNOTE-775. J Immunother Cancer. 2026;14(2) doi: 10.1136/jitc-2025-013713. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Yan G., Du Y., Zhang H., et al. Efficacy and safety of lenvatinib plus pembrolizumab in patients with advanced and recurrent endometrial cancer: a systematic review and meta-analysis. Front Immunol. 2024;15 doi: 10.3389/fimmu.2024.1404669. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Corr B.R., Thomas S.M., Haight P.J., et al. Evaluation of the combination lenvatinib and pembrolizumab in endometrial cancer; a real world multi-institutional review of practice patterns, efficacy and tolerability. Gynecol Oncol. 2025;201:69–75. doi: 10.1016/j.ygyno.2025.08.008. [DOI] [PubMed] [Google Scholar]
- 12.How J.A., Patel S., Fellman B., et al. Toxicity and efficacy of the combination of pembrolizumab with recommended or reduced starting doses of lenvatinib for treatment of recurrent endometrial cancer. Gynecol Oncol. 2021;162(1):24–31. doi: 10.1016/j.ygyno.2021.04.034. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Pawsey A., Mahalingam P., Senthivel N., et al. Real world outcomes in patients with recurrent, advanced, or metastatic endometrial cancer treated with lenvatinib plus pembrolizumab. Clin Oncol. 2025;37 doi: 10.1016/j.clon.2024.10.008. [DOI] [PubMed] [Google Scholar]
- 14.Ta Q.D., Do C., Sharma K., et al. Reduced dosing of lenvatinib with pembrolizumab for advanced endometrial cancer. Gynecol Oncol. 2025;196:36–41. doi: 10.1016/j.ygyno.2025.03.034. [DOI] [PubMed] [Google Scholar]
- 15.Marth C., Moore R.G., Bidziński M., et al. First-line lenvatinib plus pembrolizumab versus chemotherapy for advanced endometrial cancer: a randomized, open-label, phase III trial. J Clin Oncol. 2025;43(9):1083–1100. doi: 10.1200/JCO-24-01326. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Podder V., Coleman R.L., Eskander R.N., Flint M., Monk B.J., Slomovitz B.M. Checkpoint inhibitor rechallenge in advanced endometrial cancer: revisiting the immune landscape beyond first-line therapy. Int J Gynecol Cancer. 2025;35(11) doi: 10.1016/j.ijgc.2025.102124. [DOI] [PubMed] [Google Scholar]
- 17.Lheureux S., Matei D.E., Konstantinopoulos P.A., et al. Translational randomized phase II trial of cabozantinib in combination with nivolumab in advanced, recurrent, or metastatic endometrial cancer. J Immunother Cancer. 2022;10(3) doi: 10.1136/jitc-2021-004233. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.