Key Takeaways

• Combination immunotherapy with platinum-based doublet chemotherapy is the new standard of care in previously untreated advanced endometrial cancer with mismatch repair deficient/microsatellite instability‒high status.
• Combining a PARP inhibitor and durvalumab as maintenance therapy in advanced endometrial cancer is feasible and yields encouraging outcomes in patients with mismatch repair proficient status.
• In KEYNOTE-A18, addition of pembrolizumab to concurrent chemoradiation vs concurrent chemoradiation alone for patients with locally advanced cervical cancer significantly improved PFS and is likely to become the new standard of care in this setting.

In this commentary, adapted from a discussion between Brian Slomovitz, MD, MS, FACOG, and Keiichi Fujiwara, MD, PhD, the experts share their insights on seminal data highlighted at the 2023 International Gynecologic Cancer Society (IGCS) annual meeting. Drs Slomovitz and Fujiwara also share how they plan to integrate these results into their clinical practice.

Brian Slomovitz, MD, MS, FACOG:
In 2023, exciting key data for clinical trials were presented at international conferences around the world, and some of these also were highlighted during IGCS 2023. At IGCS 2023, we saw seminal data highlighted in endometrial cancer for NRG GY018, RUBY, AtTEnd, and DUO-E. Three of these studies evaluated the addition of an immune checkpoint inhibitor (ICI) to chemotherapy, and one looked at ICI therapy plus PARP inhibition.

Dr. Ramez Eskander and colleagues highlighted data previously presented at the 2023 Society of Gynecologic Oncology (SGO) annual meeting for the prospective, randomized phase III NRG GY018 study of carboplatin and paclitaxel with or without pembrolizumab followed by pembrolizumab or placebo maintenance for 2 years in patients with measurable stage III/IVA, stage IVB, or recurrent endometrial cancer. The study was designed for 2 prespecified efficacy endpoints (ie, in mismatch repair deficient [dMMR] population and mismatch repair proficient [pMMR] population). The 2 prespecified endpoints are a key difference in this study compared with other studies, where pMMR status was not a prespecified or powered endpoint.

In the dMMR population, no doubt, we can see that adding immunotherapy to chemotherapy in first-line management is practice changing, with unprecedented hazard ratios (HR) for median progression-free survival (PFS) (HR: 0.30). Moreover, in the pMMR population, the HR for median PFS was 0.54. Unfortunately, once PFS was reported, the trial ended, and overall survival (OS) data will not be part of the final analysis. Investigators also looked at whether methylation status had an effect on response to ICI—and the simple answer is no. It did not matter whether disease classified as dMMR was methylated or nonmethylated, and methylation status did not affect prognosis.

Also highlighted were data for the randomized phase III ENGOT-EN6/GOG-3031/RUBY trial evaluating carboplatin and paclitaxel with or without dostarlimab followed by dostarlimab or placebo maintenance for 3 years in patients with primary advanced or recurrent endometrial cancer. Again, this was another study that was presented at SGO 2023.

RUBY has a different statistical design than that of NRG GY018. As you may recall, the primary endpoint for RUBY was PFS by the investigator in the dMMR and overall populations, and OS analysis in the dMMR population was reported. Similar to NRG GY018, the HR for median PFS was 0.28 in the dMMR population and 0.64 in the overall population when adding dostarlimab. This data are unprecedented and are comparable to what we have come to expect with PARP inhibition in BRCA-mutant ovarian cancer. Although still preliminary, there was a clinically meaningful difference in OS in the dMMR population, with an HR of 0.30 in the dMMR population and 0.64 in the overall population.

What I found most interesting about the RUBY trial were the subgroup analyses, which investigated whether we could predict responses based on 4 molecular subtypes (eg, POLE, dMMR, P53, NSMP). POLE mutations are the least common subclassification seen with endometrial cancer and portend the best prognosis, and in this population it appeared that adding dostarlimab had no effect; all those patients did well. In addition, as discussed earlier, we saw unprecedented benefit for adding dostarlimab in the dMMR subgroup (HR: 0.31). In the P53-mutated population subgroup, the HR was 0.55. Finally, in the NSMP subgroup, the HR was 0.77, but the confidence interval did in fact cross 1. It is important to highlight that this study included carcinosarcomas, and we know based on prior work by Dr Mattew Powell that carcinosarcomas should be included in the subgroup of endometrial adenocarcinomas. These were results from small, not prespecified, subgroup analyses but were important nonetheless in helping inform which patients would do best with ICI therapy.

Together, I think these results for NRG GY018 and RUBY clearly show that patients with dMMR tumors need immunotherapy to attain the best overall outcome and have now changed the standard of care. We look forward to 2024 for additional data from the RUBY trial.

Another study we saw highlighted at IGCS was the phase III AtTEnd trial evaluating atezolizumab, a PD‑L1 inhibitor, together with the standard of care of carboplatin with paclitaxel followed by atezolizumab maintenance until disease progression vs carboplatin with paclitaxel followed by placebo maintenance in patients with newly diagnosed or recurrent endometrial cancer (N = 671). The coprimary endpoints were improving PFS per Response Evaluation Criteria in Solid Tumors (RECIST) v1.1 in the dMMR/microsatellite instability‒high and intention-to-treat (ITT) population and OS.

A key distinction of AtTEnd vs the previous 2 studies was the use of the anti‒PD-L1 atezolizumab. With a median follow-up of 26.2 months, the HR for median PFS was 0.36, and the 12-month and 24-month PFS rates were approximately 63% and 50% in patients receiving atezolizumab vs 23% and 16% in patients receiving placebo.

Although we believe that the AtTEnd OS data for all-comers in the dMMR population are clinically meaningful, we still need to see more mature data (43% maturity to date) to assess how the benefit carries out.

Similar to RUBY, the AtTEnd trial also included carcinosarcomas, but there were more nonendometrioid histologies in RUBY and AtTEnd than NRG GY018. Moreover, the platinum-free interval was ≥6 months in RUBY and AtTEnd and ≥12 months in NRG GY018.

In all, the HRs for PFS were similar among the 3 studies for the dMMR population (0.28, 0.30, and 0.36), with a signal in RUBY subgroups that patients with NSMP/P53wt disease may experience less benefit and opportunity for improvement in pMMR disease, representing an unmet need as we move forward. But how are we doing this? The next study highlighted at IGCS 2023 was DUO-E, which had shown provocative results in the pMMR population.

The phase III DUO‑E trial is exploring the addition of durvalumab, also an anti‒PD-L1, to carboplatin with paclitaxel followed by durvalumab with or without olaparib as maintenance in patients with newly diagnosed, previously untreated stage III/IV or recurrent endometrial cancer (N = 718). The primary endpoint was PFS by RECIST v1.1 with durvalumab with or without olaparib vs the control arm.

Unlike the 3 other studies, DUO-E added PARP inhibitor maintenance to ICI therapy, looking at whether adding olaparib could improve patient outcomes—and the study showed just that. In the ITT population, when comparing the median PFS in the control arm vs durvalumab arm and durvalumab plus olaparib arm, we saw a HR of 0.71 (P = .003) and 0.55 (P <.0001).

After a median follow-up of 18.6 months, OS data remain immature, with only 27% of OS events reported to date, but there was a trend favoring durvalumab plus olaparib maintenance compared with durvalumab maintenance or the control arm (HR: 0.59 vs 0.77).

Of importance, in the pMMR population (80%) there appears to be increased benefit for those receiving dual maintenance with durvalumab plus olaparib vs durvalumab alone, also compared with the control arm (HR: 0.57 vs 0.77). The data suggest that this is a group of patients who may in fact benefit from the addition of PARP inhibition.

Other future trials in advanced endometrial cancer that I am looking forward to being presented in the coming year explore the elimination of chemotherapy in the dMMR population with the phase III DOMENICA trial of first-line carboplatin and paclitaxel vs single-agent dostarlimab in dMMR advanced/metastatic endometrial cancer (NCT05201547), the phase III KEYNOTE-C93 trial of pembrolizumab vs chemotherapy in dMMR advanced/recurrent endometrial cancer (NCT05173987), and the phase III LEAP-001 trial of lenvatinib plus pembrolizumab vs carboplatin plus paclitaxel in patients with pMMR advanced/recurrent endometrial cancer (NCT03884101).

At IGCS 2023, we also discussed long-term data from the SIENDO trial. As you may recall, SIENDO is a phase III study evaluating the XPO1 inhibitor selinexor vs placebo maintenance in patients with metastatic or first-relapse advanced endometrial cancer, including endometrioid, serous, undifferentiated, or carcinosarcoma, who are currently in partial response/complete response after 12 weeks or longer of first-line carboplatin/taxane (N = 263). In the long-term follow-up from SIENDO, in median PFS results for the TP53wt population we saw unprecedented improvement with selinexor vs placebo maintenance of 27.4 months vs 5.2 months (HR: 0.42; 1-sided P = .0003). By contrast, in the TP53mut/abn population, there was no benefit with selinexor but potentially a detrimental effect, with median PFS of 4.2 months vs 5.4 months with placebo (HR: 1.34; 1-sided P = .9202). Looking deeper into the data, in the subgroup of patients with TP53wt and microsatellite stable/pMMR disease, the median PFS after 27.2 months of follow-up was not reached with selinexor and 4.9 months with placebo (HR: 0.32; 1-sided P = .0006).

Keiichi Fujiwara, MD, PhD:
Thank you, Dr Slomovitz. How will you apply these results for endometrial cancer in your practice? How will you select between immunotherapy agents (eg, pembrolizumab, dostarlimab, atezolizumab, and durvalumab with or without olaparib)?

Brian Slomovitz, MD, MS, FACOG:
Data for durvalumab with or without the PARP inhibitor olaparib (DUO-E) and the atezolizumab data (AtTEnd) were reported recently. Thus, I have not had a chance to incorporate this into my practice. As for the addition of dostarlimab or pembrolizumab to the standard of carboplatin and paclitaxel, I already am using both approaches. I believe the data to be consistent for both agents. With pembrolizumab, efficacy in dMMR and pMMR populations was a prespecified endpoint of the trial, so that may help determine which one to go with. But overall, I think both dostarlimab and pembrolizumab are active agents, and I use them interchangeably in my practice.

Keiichi Fujiwara, MD, PhD:
I think I can agree with that. I would like to highlight data presented at IGCS 2023 for localized cervical cancer. As you may recall, the current standard of care for locally advanced cervical cancer (LACC) is chemoradiation using weekly cisplatin.

At IGCS 2023, we saw data from the phase III INTERLACE trial evaluating induction chemotherapy with weekly paclitaxel 80 mg/m² and carboplatin AUC 2 for 6 weeks given on Days 1, 8, 15, 22, 29, and 36, followed by chemoradiotherapy (CRT) vs CRT using external-beam radiation in patients with newly diagnosed FIGO 2008 stage IB1N+, IB2, II, IIIB, and IVA squamous, adeno, and adenosquamous cervical cancer (N = 500). The serologic types included squamous, adeno, and adenosquamous cervical cancer; nodes above the aortic bifurcation on the imaging were not allowed; and there was no previous pelvic radiation therapy. The coprimary endpoints were PFS and OS.

Adherence to therapy was good in both groups, and patients tolerated treatments very well, although incidence of hematologic toxicities was more frequent in the induction chemotherapy group. As reported previously for the efficacy data, coprimary endpoints of OS and PFS were met in the induction chemotherapy group.

The HR for PFS was 0.65 (P = .013), and the 3-year and 5-year PFS rates were 75% and 73% with induction chemotherapy plus CRT vs 72% and 64% with CRT alone. For OS, we saw similar benefit with an HR of 0.61 (P  = .04), and the 3-year and 5-year OS rates were 88% and 80% with induction chemotherapy plus CRT vs 80% and 72% with CRT alone. Looking at the pattern of relapse, we see that the local/pelvic relapse rate was approximately 8% and 10% for induction chemotherapy plus CRT and CRT alone, but the distant relapse rate was almost double in the CRT-alone group at 20% vs 12% with induction chemotherapy plus CRT.

My perspective is that this type of induction chemotherapy can be easily applied anywhere in the world, so it will likely become part of the standard of care.

Another seminal study at IGCS 2023 for which I would like to share my takeaways is the phase III KEYNOTE-A18 trial. As you may recall, KEYNOTE‑A18 evaluated the efficacy and safety for the addition of pembrolizumab to concurrent CRT (cCRT) vs cCRT alone in patients with newly diagnosed high-risk, previously untreated LACC (N = 1060). The cCRT regimen consisted of 5 cycles (with an optional sixth dose) of weekly cisplatin 40 mg/m² with external-beam radiotherapy, followed by brachytherapy. High-risk disease was defined as having positive lymph node involvement, either pelvic or para‑aortic positive lymph nodes, or stage III and IVA disease. The experimental arm was adding pembrolizumab 200 mg every 3 weeks concurrently with chemoradiation, followed by pembrolizumab 400 mg every 6 weeks as maintenance for 15 cycles. The study coprimary endpoints were PFS per RECIST v1.1 (by investigator/histopathologic confirmation) and OS.

In KEYNOTE-A18, approximately 50% of patients were White, 30% were of Asian descent, and 20% comprised other races. More than 55% were stage III and IVA, which is a key difference from the INTERLACE trial, and 85% were node positive; 89% of the patients received high-tech radiation therapy such as intensity-modulated radiation therapy or volumetric-modulated arc therapy.

This was the first interim analysis to evaluate the treatment effects on PFS and OS. The 24-month PFS rate was 67.8% in the arm containing pembrolizumab and 57.3% in the arm with placebo. HR was 0.70 (P = .0020). In subgroup analyses, we saw benefit of adding pembrolizumab across all protocol-specified subgroups.

After median follow-up of approximately 18 months, and although data are not yet mature, the 24-month OS rate was 87.2% with pembrolizumab vs 80.8% with placebo (HR: 0.73). As for the safety analyses, addition of pembrolizumab to cCRT did not result in new safety signals.

Based on this data for KEYNOTE-A18, I believe we have another tool in our armamentarium using pembrolizumab to improve survival outcomes for LACC.

Brian Slomovitz, MD, MS, FACOG:
Thank you very much for those summaries, Dr Fujiwara. My question for you is what do you plan to do in your practice based on the positive results of INTERLACE and KEYNOTE-A18? Would you use INTERLACE or the KEYNOTE-A18 regimen?

Keiichi Fujiwara, MD, PhD:
For high-risk LACC, I will definitely use the induction chemotherapy based on the INTERLACE regimen, probably in combination with pembrolizumab once this treatment is approved.

Last, but not least, I would like to move on to highlight data presented for ovarian cancer during IGCS 2023. Professor Jonathan Ledermann presented data from the phase III ICON8B study comparing carboplatin, paclitaxel, and bevacizumab every 3 weeks (arm B1) vs dose-dense weekly paclitaxel plus bevacizumab every 3 weeks (arm B3) in patients with newly diagnosed high-risk epithelial ovarian cancer, either stage III (with residual disease or requiring new adjuvant chemotherapy) or stage IV (N = 590)—a very high‒risk population. Arm B2, with dose-dense weekly paclitaxel without bevacizumab, was discontinued because of futility. The primary endpoint of the study was PFS per study redesign, defined as a target HR of 0.75 for PFS at 87% power.

Dose-dense chemotherapy was shown to yield superior PFS and OS in Japanese patients compared with chemotherapy every 3 weeks. However, follow-up studies were not able to show a similar result in White patients. For example, in the GOG-0262 study, using dose-dense chemotherapy with bevacizumab failed to show a survival benefit, and the ICON8 study, using dose-dense chemotherapy without bevacizumab also failed in high-risk ovarian cancer. Nevertheless, primary debulking surgery was performed in 85% of the patient population. When we look at the result from ICON8B, both PFS and OS were better in the dose-dense chemotherapy group. The median PFS was 5.5 months longer (HR: 0.75; P = .002) in the dose-dense chemotherapy vs every-3-weeks group. Median OS was also 10.2 months longer (HR: 0.77; P = .02) in the dose-dense chemotherapy group vs every-3-weeks group. There also were no new safety signals.

My takeaway from ICON8B is that dose-dense chemotherapy remains the standard of care worldwide, not just for Japanese patients, although the Ovarian Cancer Consensus Conference of the Gynecologic Cancer InterGroup had concluded that it was only for Japanese patients. However, whether we should use bevacizumab remains controversial. My personal opinion is that we could go either way at this moment.

Your Thoughts?
What are your thoughts on the seminal trial results presented at recent gynecologic cancer meetings for endometrial, cervical, and ovarian cancers? Please answer the polling question and join the conversation by posting a comment in the discussion section.