Commentary
Article
Author(s):
Abdulrahman Sinno, MD, highlights key trials that have impacted the treatment paradigm in endometrial and ovarian cancer throughout 2023.
Positive data from several phase 3 trials readout in 2023 have elucidated the benefit of adding checkpoint inhibitors with or without PARP inhibitors to standard chemotherapy approaches for patients with advanced or recurrent endometrial cancer, including those with mismatch repair–deficient (dMMR) and mismatch repair–proficient (pMMR) disease. Phase 3 trials included the NRG-GY018/KEYNOTE-868 (NCT03914612), RUBY (NCT03981796), and DUO-E (NCT04269200) studies, and investigations such as the phase 2/3 ROCSAN trial (NCT03651206) aim to expand these approaches to patients with less common and more aggressive tumors, according to Abdulrahman Sinno, MD.
“2023 was a big year for immune checkpoint inhibitors, and one of the big things is including checkpoint inhibitors in the front line,” Sinno said in an interview with OncLive® regarding a recent OncLive State of the Science Summit™ on gynecologic oncology, which he chaired. “I do look forward to a time where we use these checkpoint inhibitors in both dMMR and pMMR tumors in the upfront setting.”
Sinno adds that these insights may pave the way for tailored treatment approaches and continued research into the molecular characteristics of endometrial cancer, aiming to optimize outcomes and refine therapeutic strategies. Sinno is an associate professor of Clinical Obstetrics Gynecology and Reproductive Sciences at the University of Miami Miller School of Medicine, as well as the director of Surgical Research and Education at Sylvester Comprehensive Cancer Center in Miami, Florida.
In the interview, Sinno discusses significant findings from recent trials in advanced endometrial and ovarian cancer, focusing on their implications for treatment approaches and molecularly stratified patient populations. He expands further on the shift towards precision medicine approaches in endometrial cancer management in a concurrent interview.
Sinno: NRG-GY018 is a randomized study of standard paclitaxel and carboplatin vs paclitaxel and carboplatin plus pembrolizumab [Keytruda] in patients with advanced stage endometrial cancer. These are patients with stage III/IV measurable disease; stage IIIa or IVa with measurable disease or IVb recurrent whether there’s measurable disease or not. Patients in that trial may have received prior radiation therapy or hormonal therapy, so that wasn’t an exclusionary criterion. Patients in whom both radiation and chemotherapy were planned must have received that radiation ahead of entering the study. The patients were randomly assigned paclitaxel and carboplatin plus placebo, or [the chemotherapy regimen] plus pembrolizumab followed by maintenance pembrolizumab for 12 months. Stratification was by pMMR vs dMMR status, [ECOG] performance status, and whether patients had measurable disease. The primary end point was progression-free survival [PFS]. Interestingly, in this study there were 591 pMMR patients and [225] dMMR patients.
As of the data cutoff, the median PFS in the [control] group of patients who were dMMR or microsatellite instability [MSI]–high wasn’t reached and the HR for disease progression or death was 0.30. That’s something we haven’t seen in a very long time and indicates that this is an extremely effective treatment in the dMMR group. Whenever you have survival curves and HRs that are that impressive, it begs the question of whether this is practice changing. Based upon these data, we had the [FDA approval] of [single-agent] pembrolizumab in dMMR disease.
Interestingly, we still see a benefit [with the pembrolizumab regimen] in the pMMR group. The HR is not as impressive in the [pMMR group] as it is in the dMMR group because we do expect the dMMR group to have a better response to checkpoint inhibitors. However, we do still see a stratified HR for disease progression or death of 0.54 and this confidence interval doesn’t cross 1.
We see similar survival curves with the RUBY trial. One of the biggest differences [between NRG-GY018 and RUBY] would be the inclusion of carcinosarcoma in the RUBY trial. The HRs were also [comparable]. In the overall population, that HR was 0.64 and in the dMMR group it was 0.30. Interestingly, we do see the separation of these curves late. Up until 12 months, most of these curves are close to each other in the overall population. Right after 12 months is when the separation really starts to occur.
These are studies that included long-term follow-up and maintenance, which did make the difference. If you looked at 12-month PFS in the overall population, you’re not going to see much of a difference because the curves separate later. This was great [with the] study design.
[Looking at] the molecular subclassification of these patients, only [approximately] 1.2% harbored POLE mutations, which is very interesting. We see repeatedly that the POLE percentage in the population of interest has not been very common. However, this site included a high percentage of [patients with] TP53 wild type [disease]. It’s good to have these data because it allows us to understand what population we’re dealing with. Looking at these data put together, [the majority of] patients had pMMR disease, so the fact that you still see this benefit [with dostarlimab-gxly (Jemperli) plus chemotherapy] is impressive.
When you look specifically at what subgroups benefited [in a post-hoc analysis], the no specific molecular profile [NSMP] group was the only group where that confidence interval crossed 1. [These data are] hypothesis-generating, but we should be cautious interpreting them. I’m not saying that the NSMP group doesn’t benefit from this regimen overall. Within other categories, [the HRs] were significant in the dMMR/MSI-H subgroup and p53-mutant subgroup. Future trials will be designed with this in mind and be powered appropriately.
What is it about an NSMP, specifically, [that may not lead to benefit with chemoimmunotherapy]? How does having TP53 wild-type disease potentially make you less [responsive] to a checkpoint inhibitor vs p53-mutant disease? These data are interesting because we do think of patients with p53 mutations as having [poor] prognosis in general. These data are enough to build upon, but not enough to [inform] clinical decision-making.
The ROCSAN trial is evaluating anti-PD1 therapy plus niraparib in recurrent ovarian or uterine carcinosarcoma [following] the first line [of chemotherapy]. Patients underwent a biopsy and circulating tumor DNA [(ctDNA) assessment] and were randomly assigned 2:2:1 to [either] niraparib monotherapy, niraparib in combination with dostarlimab, or investigator’s choice of chemotherapy [n = 59]. Then the [best] of those 2 [experimental] arms continues enrollment—the best arm was the niraparib [combination] arm, and 170 patients will be randomly assigned to that.
The first part was only done in France. Patients were stratified according to ECOG performance status, previous lines of therapy, initial FICO stage, and ovarian vs uterine carcinosarcoma. Recruitment is going to be 36 months in this trial. It’s interesting that we have ctDNA correlates included in this study, and most of these trials are being designed with ctDNA correlates in mind. I am excited to see the results of those moving forward.
DUO-E is investigating the combination of immune checkpoint inhibitors and PARP inhibitors in newly diagnosed stage III/IV recurrent endometrial cancer. Patients should not have had any prior lines of chemotherapy, [although] adjuvant chemotherapy was allowed if [patients were] more than 12 months from the last treatment to relapse. We needed to have known MMR status, prior radiation was allowed, and all histologies were included except sarcomas. Patients were stratified by MMR status, disease status, and geographic region.
These are studies that were designed earlier, so we weren’t stratifying by molecular classification. Patients were randomly assigned 1:1:1 and the whole study essentially used combinations of olaparib [Lynparza] and durvalumab together. The first arm [consisted of] platinum-based chemotherapy [followed by placebo maintenance], the second arm was durvalumab with [chemotherapy] and [placebo] maintenance, and the third arm was [durvalumab plus chemotherapy followed by olaparib maintenance therapy.] The primary end point was PFS between arms A and B. Secondary end points included PFS between arms A and C, overall survival, PFS2, safety, and patient-reported outcomes as well.
Interestingly, the initial press release in May said that durvalumab in combination with platinum-based chemotherapy followed by either durvalumab plus olaparib or durvalumab alone did show a PFS advantage. [However], the data published in October of 2023 showed a PFS benefit observed in the durvalumab arm [vs control arm]; the HR was 0.71 and this was statistically significant. We also saw a benefit with durvalumab plus olaparib vs durvalumab alone, and the HR was 0.55.
[However], we don’t have a specific head-to-head [comparison] between [arms] B and C. That wasn’t what the study design was for [but] the curves are interesting to look at. If you look at it by MMR status, there wasn’t any benefit in the dMMR group with durvalumab plus olaparib [vs] durvalumab alone—the curves are almost overlapping. [Conversely], in the pMMR group, which was approximately 80% of the population, there did appear to be some benefit [with the durvalumab and olaparib regimen]. The HR vs control was 0.77, and the HR vs durvalumab alone was 0.76, and the confidence interval didn’t officially cross 1.
The outcomes [from this trial] might not be as impressive overall, but when you subclassify by molecular status, there’s some interesting information there. When you have a patient who is dMMR, the magnitude of the effect that you’re going to have from adding the checkpoint inhibitor is going to be so overwhelming that adding the PARP inhibitor on top of that isn’t really going to help. In the pMMR group, we can see the sensitization and the synergy between the PARP inhibitor and the checkpoint inhibitor, because the checkpoint inhibitor isn’t as effective in that setting. I’m interested to see where this is going to go in the future.