Historically, endometrial cancer has been classified into 2 Bokhman types based on histologic and molecular characteristics—type I (endometrioid) and type II (nonendometrioid).⁷ Type I tumors are most commonly endometrioid adenocarcinomas and represent up to 90% of all endometrial cancers.⁸ The most common molecular alterations in type I tumors affect the PI3K/PTEN/AKT pathway and lead to increased cell proliferation and survival. Type I tumors express the estrogen and progesterone receptors and are estrogen dependent. Typically, they are low grade and have a more favorable prognosis if diagnosed in the early stage.

The type II tumors are predominantly serous and clear-cell adenocarcinomas and most commonly exhibit alterations involving the tumor suppressors p53 and/or p16, leading to cell cycle disruption and genetic instability. Patients with type II tumors generally have high-grade disease that is not dependent on estrogen. Unfortunately, these patients have worse prognoses and a lower survival rate.

In 2013, The Cancer Genome Atlas proposed a new classification system for endometrial cancer based on molecular characteristics observed in 373 endometrial carcinomas.⁹ This system divided endometrial cancers into 4 categories: POLE ultramutated, MSI hypermutated, copy-number low (endometrioid), and copy-number high (serous like). I will discuss MSI in greater depth shortly.

These molecular classifications have important implications for prognosis and treatment. PFS was most prolonged in patients with POLE-ultramutated disease and shortest in those with copy-number high disease, whereas those with MSI hypermutated or copy-number low disease fell between these extremes (log-rank P = .02). Regarding treatment, the investigators noted that approximately one quarter of tumors classified by pathologists as high-grade endometrioid disease exhibited a molecular phenotype more characteristic of uterine serous carcinomas—suggesting that patients with this molecular phenotype might benefit more from treatments used for type II serous disease (eg, chemotherapy).

Malignancies with a dMMR system accumulate mutations in microsatellites—short, repetitive DNA sequences susceptible to mismatch errors.¹⁰ There are 4 genes encoding proteins critical to the MMR process: MLH1, MSH2, MSH6, and PMS2. Inactivation of at least 1 of these genes via germline or somatic mutations and/or epigenetic silencing leads to dMMR that, in turn, leads to MSI.

In an analysis of MSI status for 7919 tumor and matched normal pairs encompassing 23 different malignancies, endometrial cancers exhibited the greatest frequency of MSI-high status (28.3%).¹¹ This is important because MSI suggests susceptibility to immune checkpoint inhibition, which I will discuss shortly.¹⁰

It is important to assess MSI status in endometrial cancers. ESMO consensus guidelines recommend IHC testing using antibodies against the 4 proteins listed above. In the event of indeterminate IHC findings, molecular testing with MSI PCR is recommended using one of 2 panels assessing microsatellite markers, where MSI is defined as loss of stability in ≥ 2 of 5 markers. The poly-A panel is preferred due to its greater sensitivity and specificity; this panel looks at 5 poly-A mononucleotide repeats (BAT-25, BAT-26, NR-21, NR-24, NR-27). The alternative panel looks at 2 mononucleotide (BAT-25, BAT-26) and 3 dinucleotide (D5S346, D2S123, D17S250) repeats. Next-generation sequencing represents an alternative molecular testing option, with the advantage being that it can also assess tumor mutational burden—another potential biomarker for immunotherapy efficacy.

Endometrial cancers that are MSI‑high/dMMR exhibit a markedly higher mutational load, reflected in a 7-fold higher number of neoantigens compared with microsatellite-stable (MSS) tumors.¹² These tumor‑specific neoantigens are highly immunogenic, resulting in greater numbers of tumor‑infiltrating lymphocytes and, remarkably, a compensatory upregulation of immune checkpoints (ie, PD-1/PD-L1) by tumor cells.¹³ Thus, patients with endometrial cancers are potential candidates for treatment with immune checkpoint inhibitors.

Early evidence of clinical activity with immune checkpoint inhibitor–based therapy in advanced endometrial cancer came from KEYNOTE-016, a pivotal multicohort phase II trial. Le and colleagues¹⁴ evaluated the PD-1 inhibitor pembrolizumab in 41 patients with progressive metastatic carcinoma with or without dMMR. The 3 trial cohorts comprised patients with colorectal adenocarcinomas with dMMR; pMMR colorectal adenocarcinomas; and noncolorectal malignancies with dMMR, with this final cohort including 2 patients with endometrial cancer. All patients received pembrolizumab at 10 mg/kg every 14 days.

In the noncolorectal cohort with dMMR, after a median follow-up of 21 weeks, the immune-related ORR was 71%. The 20-week immune-related PFS rate was 67%.

Two years after publication of KEYNOTE-016, Ott and colleagues¹⁵ published an analysis from the multicohort phase Ib KEYNOTE-028 trial, which evaluated the safety and efficacy of pembrolizumab in patients with PD-L1–positive advanced solid tumors. In 23 evaluable patients with PD-L1–positive endometrial cancer, the ORR was 13% (all PRs).

In 2017, a pooled analysis of KEYNOTE-016 and 4 other multicohort, single-arm trials led to the FDA approval of pembrolizumab for treatment of patients with unresectable or metastatic, MSI-high/dMMR solid tumors that progressed following previous treatment and had no satisfactory alternative options.¹⁶ Of the 149 patients in this analysis, 14 had endometrial cancer and exhibited an ORR of 36%.¹⁷ This marked the first time the FDA issued a tissue/site-agnostic approval.¹⁶

The ongoing multicohort phase II KEYNOTE‑158 trial is evaluating pembrolizumab in patients with advanced solid tumors who had progressed on standard therapy.¹⁸ Patients received pembrolizumab 200 mg every 3 weeks for 2 years or until they experienced disease progression, unacceptable toxicity, or consent withdrawal. The trial included cohort D, which enrolled patients with advanced endometrial cancer regardless of biomarker status, and cohort K, which enrolled patients with advanced noncolorectal MSI-high/dMMR malignancies.^19,20

Cohort K included 49 patients with MSI-high/dMMR endometrial cancer.²⁰ Among these 49 patients, the ORR was 57.1%, which included 8 CRs. Of interest, the median duration of response was not yet reached. The median PFS was 25.7 months and the median OS was also not yet reached.

Among these 49 patients with MSI-high/dMMR endometrial cancer, any-grade treatment‑related AEs were observed in 79.6%.¹⁹ In addition, any-grade immune‑related AEs were observed in 32.7% of patients, of whom 8.2% experienced a grade ≥ 3 event. The most common grade ≥ 3 immune-related AEs were colitis (2.0%), type 1 diabetes mellitus (2.0%), and hepatitis (2.0%).

At ESMO 2020, my colleagues and I²¹ presented results from the phase I GARNET trial. This is a large trial of 290 patients with recurrent/advanced endometrial cancer with or without dMMR who were treated with the investigational PD-1 inhibitor dostarlimab (formerly TSR-042).²¹ The expansion cohorts were divided by MMR status, with 129 patients being classified as dMMR vs 161 as pMMR. Both expansion cohorts received dostarlimab at 500 mg every 3 weeks for 4 cycles followed by 100 mg every 6 weeks until disease progression. The primary endpoint was ORR.

The safety population in GARNET consisted of all patients enrolled and dosed (dMMR, n = 126; pMMR, n = 145).²¹ The efficacy population included only those with ≥ 6 months of follow-up and ≥ 1 measurable lesion at baseline (dMMR, n = 103; pMMR, n = 142).²¹

Looking at baseline characteristics, the most notable difference between the cohorts is histology. In the dMMR cohort, the most common histology was endometrioid carcinoma type I (68.0%), whereas the majority of patients in pMMR cohort had type II endometrial carcinoma (76.8%), with serous histology being the most common (58.0%). All patients had received at least 1 previous line of therapy.

With a median follow-up of 16.3 months in the dMMR cohort and 11.5 months in the pMMR cohort, the ORR was 44.7% and 13.4%, respectively.²¹ CRs were observed in 10.7% and 2.1% of each respective cohort. The disease control rate—defined as a CR, PR, or stable disease for ≥ 12 weeks—was quite impressive at 57.0% in the dMMR cohort and 35.2% in the pMMR cohort.

Among the patients with response, 89.1% of the dMMR cohort and 63.2% of the pMMR cohort remained in response at data cutoff.²¹

The safety profile observed with dostarlimab is consistent with that of other immune checkpoint inhibitors. Grade ≥ 3 treatment‑related AEs occurred in 13.5% of the dMMR cohort and 19.3% of the pMMR cohort.²¹ Regarding immune checkpoint inhibitor-related AEs, the most common any-grade events in the dMMR and pMMR cohorts were hypothyroidism (5.6% and 7.6%, respectively), diarrhea (4.8% and 3.4%), and increased aspartate aminotransferase (1.6% and 2.8%). Grade ≥ 3 immune-related treatment-related AEs occurred at low rates, with the most common events being increased aspartate aminotransferase, diarrhea, and increased amylase (all ≤ 1.6% in each cohort). Discontinuations due to treatment-related AEs occurred in 4.0% of the dMMR cohort and 6.9% of the pMMR cohort.

To summarize, the GARNET trial observed durable antitumor activity with dostarlimab in patients with dMMR or pMMR advanced/recurring endometrial cancer. Of importance, dMMR status identified those patients most likely to benefit from immune checkpoint inhibition.

We turn now to data on 2 other immune checkpoint inhibitors, avelumab and durvalumab, in endometrial cancer.

The PD-L1 inhibitor avelumab was evaluated in a multicohort phase II trial that enrolled patients with endometrial cancer with dMMR (n = 15) or with pMMR (n = 16).²² All patients had been treated with ≥ 1 previous chemotherapy. Although patients with POLE mutation were eligible to enroll in the dMMR cohort, there were no patients harboring this mutation; patients with a POLE mutation were ineligible to enroll in the pMMR cohort.

The coprimary endpoints were the proportion of patients achieving PFS of ≥ 6 months and ORR. Secondary endpoints included OS, PFS, and safety.

The ORR was 26.7% in the dMMR cohort but only 6.25% in the pMMR/non‑POLE cohort.²² PFS of ≥ 6 months was observed in 40.00% and 6.25%, respectively.

The single-arm phase II PHAEDRA trial is evaluating the activity of the PD-L1 inhibitor durvalumab in 2 cohorts of patients with advanced endometrial cancer that is either pMMR or dMMR and that progressed after 1-3 previous lines of chemotherapy; patients assigned to the dMMR cohort could be previously untreated.²³ The primary endpoint was ORR per the Response Criteria in Solid Tumors for Immunotherapy, and secondary endpoints included the disease control rate at 16 weeks, immune-related AEs, and biomarker assessments.

At ASCO 2019, Antill and colleagues²³ presented preliminary data from 71 patients. The ORR was 43% in the dMMR cohort and only 3% in the pMMR cohort. In the dMMR cohort, the ORR was 52% in those receiving durvalumab in the first-line setting, 31% in the second line, and 0% in the third line.

The disease control rate at 16 weeks was 60% in the dMMR cohort and 20% in the pMMR cohort.

Collectively, these data indicate that there is a remarkable difference in the activity of checkpoint inhibitors in the patients with endometrial cancer that is dMMR vs pMMR. There is clearly room to improve the activity of checkpoint inhibitors in pMMR disease. Next, I discuss the activity of several combination regimens in this population, notably immune checkpoint inhibitors combined with antiangiogenic agents or chemotherapy.

The combination of pembrolizumab plus the oral multikinase inhibitor lenvatinib was approved in 2019 in the United States, Australia, and Canada for patients with advanced endometrial carcinoma that is neither MSI-high nor dMMR, that progressed following previous systemic therapy, and who were not candidates for curative surgery or radiation.^17,24 This approval was based on the ongoing single-arm phase Ib/II KEYNOTE‑146 trial, which is evaluating this regimen in patients with various solid tumors, including a cohort with endometrial carcinoma.

This cohort enrolled patients with pathologically confirmed metastatic endometrial carcinoma that had been treated with ≤ 2 previous lines of systemic therapy (≥ 2 permitted at discretion of sponsor), excluding those previously treated with lenvatinib or PD-(L)1 inhibition.²⁵ All patients were centrally assessed for MSI/dMMR status. Patients received pembrolizumab 200 mg every 3 weeks plus lenvatinib 20 mg/day, to a maximum of 35 treatments with pembrolizumab. The primary endpoint was ORR at Week 24, with secondary endpoints including ORR, duration of response, OS, PFS, and disease control rate.

The primary efficacy analysis included 124 patients, of whom 108 had previously treated disease (dMMR, n = 11; pMMR, n = 94). In the previously treated population, the percentage of those with PD-L1–positive disease was 63.6% in the dMMR and 48.9% in the pMMR cohorts. A minority of patients had been treated with ≥ 3 previous lines of therapy for endometrial cancer (dMMR, 9.1%; pMMR, 10.6%).

In the previously treated population, the ORR at Week 24 was 38.0%.²⁵ When we look at the cohorts divided by MSI/dMMR status, we see a higher ORR at Week 24 in those with dMMR compared with those with pMMR (63.6% vs 36.2%, respectively). The median duration of response was 21.2 months in the dMMR cohort and was not yet reached in the pMMR cohort.

Although the ORR at Week 24 was higher in those with dMMR, responses were observed regardless of MSI/dMMR status or histology.²⁵

Although this regimen is quite active in both dMMR and pMMR patient cohorts, the investigators observed a number of important AEs. Approximately 3 of 5 patients (62.9%) experienced treatment-related AEs necessitating lenvatinib dose reductions, and 18.5% of patients had to discontinue study treatment due to treatment-related AEs.²⁵ More than one half (57.3%) experienced immune-related AEs, of which hypothyroidism was the most common (47.6%). In addition, there were 2 deaths deemed related to treatment (sepsis from Escherichia coli infection, n = 1; intracranial hemorrhage, n = 1).

As previously mentioned, the results from KEYNOTE-146 led to the approval of pembrolizumab plus lenvatinib in the United States, Australia, and Canada for patients with advanced endometrial carcinoma that is not MSI-high nor dMMR, that progressed following prior systemic therapy, and who were not candidates for curative surgery or radiation.²⁴

This combination is under investigation in specific settings of endometrial cancer. For example, the ongoing phase III KEYNOTE-775 trial is comparing second-line treatment with pembrolizumab plus lenvatinib vs physician’s choice of chemotherapy in 780 patients with advanced endometrial cancer.^26,27 Of those planned 780 patients, approximately 120 will have dMMR and 660 will have pMMR disease. The coprimary endpoints are PFS and OS. We are anticipating that the primary competition date will be in February 2022.

The ongoing phase III ENGOT/LEAP-001 trial is another important study comparing first-line treatment with pembrolizumab plus lenvatinib vs chemotherapy in 720 patients with newly diagnosed stage III/IV or recurrent endometrial cancer that can be dMMR or pMMR.^28,29 The coprimary endpoints are PFS and OS, and the estimated primary completion date is April 2023.

We have additional data on another combination of anti–PD-1 therapy plus antiangiogenic agents. At ASCO 2020, Lheureux and colleagues³⁰ presented results from a randomized phase II trial that compared nivolumab plus cabozantinib vs nivolumab alone in patients with recurrent epithelial endometrial carcinoma that had been treated with ≥ 1 previous platinum-based chemotherapy.

The primary endpoint of PFS was significantly improved with nivolumab plus cabozantinib compared with nivolumab alone (5.3 vs 1.9 months, respectively; HR: 0.59; 95% CI: 0.35-0.98; P = .09, meeting predefined PFS significance level of α = 0.10).³⁰ The ORR was 25.0% with nivolumab plus cabozantinib vs 11.1% with nivolumab alone.

Regarding safety, common AEs in the combination arm included diarrhea, increased aspartate aminotransferase, increased alanine aminotransferase, fatigue, and hypertension.

The other combination strategy under investigation is immune checkpoint inhibition plus chemotherapy. We do not yet have data on this approach, but I will quickly review some important ongoing trials. The basic rationale for combining these approaches is that chemotherapy can augment immune stimulation and promote a favorable tumor microenvironment for enhancing the action of PD-(L)1 inhibitors.¹³

The phase III NRG-GY018 trial is comparing pembrolizumab plus paclitaxel/carboplatin vs placebo plus paclitaxel/carboplatin in 810 patients with stage III/IV or recurrent endometrial cancer.³¹ The primary endpoint is PFS with secondary endpoints including ORR, OS, quality of life, and safety.

As of November 2020, this trial is suspended due to COVID-19.

Another important trial is the phase III AtTEnd/ENGOT-en7 trial comparing atezolizumab plus paclitaxel/carboplatin vs placebo plus paclitaxel/carboplatin in 550 patients with advanced endometrial cancer that could either be newly diagnosed or recurring.³² The coprimary endpoints are OS and PFS, and the estimated primary completion date is July 2021.

Finally, the phase III RUBY trial is comparing dostarlimab plus chemotherapy vs placebo plus chemotherapy in 470 patients with recurrent or advanced endometrial cancer.³³ The primary endpoint is PFS and the estimated primary completion date is October 2021.

As discussed, immune checkpoint inhibitor–based therapy is quite active in endometrial cancer, but it also comes with the risk of immune-related AEs. There are 3 key points to emphasize about managing immune-related AEs. First, it is imperative to identify these AEs as early as possible—the sooner these are recognized and treated, the better the outcome will be. Second, use a multidisciplinary approach to management: When immune‑related AEs such as hypothyroidism and hyperthyroidism are identified, consult with endocrinology. Likewise, when severe diarrhea and/or colitis is identified, consult with gastroenterology. Third, use steroids wisely to manage these immune-related AEs—know when to initiate, to augment, to switch, and to taper.

Detailed recommendations regarding the management of a variety of organ-based immune-related AEs associated with immune checkpoint inhibitors have recently been updated by the National Comprehensive Cancer Network^® (NCCN^®). To receive case-specific guidance from the NCCN Guidelines^® on irAE management for unique patient scenarios, please visit the interactive algorithm tool developed by Clinical Care Options in partnership with the NCCN^®.