Endometrial cancer (EC) is the most common gynecological malignancy in developed countries, with a majority of cases occurring postmenopause (1). Over the past decade, the prevalence of EC and disease-associated mortality rates has been continuously rising worldwide (1, 2). The prognosis of patients with primary advanced EC is poor, and available therapeutic options are limited. At the moment, no standard therapy has been globally accepted as the standard of care for advanced and recurrent EC after the failure of platinum-based chemotherapy.

Since the proposal of molecular subgroups based on the analysis of the Cancer Genome Atlas (TCGA) in 2013, it has played a crucial role in formulating the most appropriate therapy and identifying prognostic factors (3). More recently, monotherapy with an immune checkpoint inhibitor (ICIs), including antiprogrammed death 1 (PD-1) and antiprogrammed death ligand-1 (PD-L1), has demonstrated a substantial improvement in oncologic outcomes in patients with deficient mismatch repair (MMRd) EC (4, 5). Although several trials highlighted the limited efficacy of anti-PD-1/PD-L1 agents alone, combined therapy of PD-1/PD-L1 inhibitors and antiangiogenic agents is an attractive approach in proficient mismatch repair (pMMR) EC (6, 7).

In 2009, a phase I study presented the first clinical evidence of olaparib, a PARP inhibitor, having antitumor activity in patients associated with breast cancer (BRCA)1/2 mutations (8). In patients without BRCA mutations, the therapeutic value of PARP inhibitors is associated with homologous recombination deficiency (HRD) (9). Defects in homologous recombination repair (HRR) genes, including BRCA1/2, BRIP1, and RAD51C, have the potential to produce HRD (10, 11). A systematic review reported that BRCA1/2 pathogenic variants were identified in 4.3% of EC patients (10). In a large-scale study of molecular profiles of 1,475 EC cases, pathogenic variants in HRR genes were observed in 34.4% of cancers (12). Finally, alterations in the phosphatase and tensin homolog (PTEN) gene occur in up to 78% of EC (13). In preclinical studies, loss of PTEN is associated with HRD, sensitizing tumors to PARP inhibitors (14).

Although several clinical trials have either been completed or are ongoing, evidence supporting the PARP inhibitors, alone or in combination with other drugs, in advanced EC still remains limited. Currently, the majority of relevant studies are phase II clinical trials with small sample sizes. As such, this meta-analysis aimed to synthesize the available data to investigate the effectiveness and safety of PARP inhibitors for the treatment of advanced or recurrent EC.

This is an up-to-date meta-analysis on the efficacy and safety of PARP inhibitors for patients with advanced or recurrent EC. We included 12 studies with a total of 1,594 patients, and the most reliable comparative information comes from the HRs of randomized trials. The main findings from the pooled analyses were as follows: (1) the results showed that the combination of PARP and PD-L1/PD-1 inhibitors could significantly prolong PFS versus placebo; (2) in the intention-to-treat population, there was no PFS/OS difference between PARP inhibitor monotherapy and placebo; (3) the ORR was notably elevated in the combined therapy group compared with PARP inhibitor monotherapy; and (4) the most common AEs included fatigue (226, 54.5%), nausea (219, 52.8%), anemia (219, 52.8%), diarrhea (127, 30.6%), and alopecia (121, 29.2%).

Recently, novel immunotherapeutic strategies have been proposed to enhance clinical efficiency. PARP inhibitors are among the most promising agents and have already demonstrated compelling results in ovarian cancer (28). Furthermore, as it has been reported that the accumulation of DNA damage caused by PARP inhibitors may reverse the immune escape by upregulating PD-L1, combining PD-1/PD-L1 inhibitors with PARP inhibitors is an attractive approach for EC patients (28). Our study suggests that the combination of PARP with PD-L1/PD-1 inhibitors could significantly prolong PFS versus placebo for advanced or recurrent EC patients, and ORR was notably elevated in the combined therapy group compared with PARP inhibitor monotherapy. Studies investigating PD-1/PD-L1 inhibitor monotherapy in MMRd EC revealed a median PFS ranging from 4.4 to 25.7 months, along with an ORR of 26.7% to 57.1% (5, 29, 30,), which were better than those reported in this meta-analysis. The outcomes of our study appear better than those of studies involving PD-1/PD-L1 inhibitors in pMMR EC, which indicated a median PFS of 1.8 to 1.9 months and an ORR of 3.0% to 13.4% (4, 29, 30). Several previous studies have found that the combination of PARP inhibitors, such as olaparib and niraparib, with PD-1/PD-L1 inhibitors might have modest clinical effects (17–19). The addition of the anti-PD-L1 antibody durvalumab with olaparib to platinum-based chemotherapy significantly improved PFS for patients with advanced or recurrent EC compared to durvalumab or chemotherapy alone, demonstrating a potential role for PARP inhibitors (18).

The efficacy of PARP inhibitors was also tested in association with other drugs, such as antiangiogenesis and AKT inhibitors (16, 21, 22). Some studies demonstrate that chronic hypoxia induced by antiangiogenic agents, such as bevacizumab, can lead to HRD within tumors, potentially enhancing sensitivity to PARP inhibitors (16, 31, 32). The combination of antiangiogenic agents with PARP inhibitors has also been assessed in a phase II trial (NCT03660826) (20). In this three-arm randomized trial, PFS was 3.8 months for cediranib alone, 2 months for olaparib, and 5.5 months for olaparib and cediranib. However, the between-arm differences were not statistically significant (20). In this meta-analysis, we included two trials that investigated the combination of antiangiogenic agents with PARP inhibitors. The pooled ORR was found to be 0.138 (95% CI = 0.072–0.203) for advanced or recurrent EC, which is similar to that observed with the combined therapy including a PD-1/PD-L1 inhibitor.

While generally manageable with successful management strategies, AEs occurred more frequently and remain a significant concern with combination immunotherapies. The results of this study indicated that the most commonly reported AEs of any grade included fatigue (226, 54.5%), nausea (219, 52.8%), anemia (219, 52.8%), diarrhea (127, 30.6%), and alopecia (121, 29.2%). Although the frequency of adverse events is typically higher in combination chemotherapies, grade ≥ 3 AEs were reported with low incidence, with the most frequent being anemia (88, 21.2%), decreased neutrophil count (76, 18.3%), fatigue (31, 7.5%), decreased platelet count (28, 6.7%), and decreased white blood cell count (17, 4.1%).

In the phase III DUO-E trial, the overall incidence of treatment-emergent AEs of grade ≥ 3 in the control, durvalumab, and durvalumab + olaparib arms was 56.4%, 54.9%, and 67.2%, respectively (18). Although there was a higher rate of AEs of grade ≥ 3 in the durvalumab + olaparib arm, the safety profiles were generally consistent with the known profiles of the individual components of the regimen (18). However, in the phase II DOMEC trial (19), the combination of durvalumab and olaparib was well tolerated, with no new safety profiles observed, consistent with those in previous studies (19, 29, 30).

Targeted therapies, such as olaparib, which inhibit PARP, induce synthetic lethality in tumor cells harboring HRD. The percentage of BRCA-mutated endometrial cancers is very low; therefore, HRD is an emerging concept in EC and appears to play a role in its pathogenesis. HRD is largely restricted to TP53-mutated ECs, and approximately 20%–40% of p53-abnormal tumors harbor HRD. The UTOLA trial suggests that olaparib may provide the greatest PFS benefit in the population with TP53 mutations (25).

For patients with pMMR disease, the magnitude of PFS benefit is lower, and the efficacy of PD-1 and PD-L1 inhibitors is variable. In the recurrent setting, the combination of pembrolizumab and lenvatinib is a promising therapeutic option for the pMMR/microsatellite-stable (MSS) subset, despite the significant morbidity associated with this pharmacological regimen. Moreover, exploratory subgroup analyses of the DUO-E trial suggest that the addition of maintenance olaparib to the combination of durvalumab plus chemotherapy may improve outcomes in the pMMR and PD-L1–positive patient populations (25). Consequently, assessing PD-L1 expression in recurrent pMMR/MSS cases could help identify patients who might benefit from the combination of PARP with ICIs (33).

There has been strong interest in utilizing a PARP inhibitor alone or in combination with other agents for managing EC. Several ongoing trials are testing the safety and effectiveness of PARP inhibitors in combination with chemotherapy (e.g., NCT04159155, NCT03603184), ICI (e.g., NCT04269200, NCT03016338, NCT03981796), and antiangiogenics (e.g., NCT03660826). Moreover, further biomarker-driven trials aimed at guiding the therapeutic selection and testing targeted agents are warranted.

This research has several limitations. First, owing to restricted data availability, subgroup analyses based on factors such as MMR status, PD-L1 expression, BRCA mutation, and HRD status were not feasible. Second, several clinical trials have not yet reached their final OS endpoints; thus, our OS analyses are based on a limited number of immature datasets and should be interpreted with caution. Updated analyses will be needed once these trials mature. Third, the majority of included trials were nonrandomized, single-arm, and unblinded, and therefore carried high risks of selection bias and performance bias. Further large-scale RCTs are required to confirm these results.

In conclusion, our study provided preliminary evidence that the combination of PARP inhibitors with PD-L1/PD-1 inhibitors exhibits modest activity accompanied by substantial toxicity, particularly in heavily pretreated and predominantly pMMR advanced or recurrent EC. Definitive conclusions regarding survival benefits, however, require further phase III data and biomarker-driven analyses.