Recent advances in endometrial cancer: a review of key clinical trials from 2015 to 2019

Introduction

In 2019, there will be an estimated 61,880 new cases of and 12,160 deaths from endometrial cancer¹. It is the most common gynecologic malignancy in the US and the only gynecologic cancer with increasing incidence and mortality². Tran and Gehrig published a comprehensive F1000 Faculty Review in January 2017³. They outlined the genetic bases of endometrial cancer, novel surgical treatment, molecular targeted therapeutics, and current clinical trials. We invite you to reference their review for the groundwork of this article. Since their publication, we have seen several important advances in understanding and therapy, including surgical staging and the utility of sentinel lymph node (SLN) mapping, adjuvant treatment for high-risk disease and HER2/neu-positive serous tumors, combination therapy for recurrent disease, molecular biology, and immunotherapy. We focus on these recent advances.

Surgical staging: sentinel lymph node evaluation

Lymph node involvement is a critical factor in determining prognosis and adjuvant therapy in endometrial cancer. As discussed in greater detail below, evidence supports a survival advantage for adjuvant chemotherapy in patients with stage III endometrial cancer; however, a similar benefit may not accrue to patients with locally advanced disease^4–6. Complete staging lymphadenectomies have been associated with morbidity that impacts quality of life, including lymphedema, lymphocele formation, and neuralgia^5,7,8. Several criteria have been proposed to define patients in whom the risk is low enough to safely omit a staging lymphadenectomy. Most recently, the Korean Gynecologic Oncology Group showed that their criteria—endometrioid histology, no evidence of deep invasion or enlarged lymph nodes on magnetic resonance imaging (MRI), and pre-operative CA-125 of less than 35 units/mL—resulted in a negative predictive value of 97.1%⁹; however, some criticize the cost and burden associated with obtaining pre-operative MRI and CA-125¹⁰. SLN biopsies offer a compromise and have been shown to reduce risks of lymphedema and lymphocele¹¹.

In 2017, the results of the Fluorescence Imaging for Robotic Endometrial Sentinel lymph node biopsy (FIRES) trial were published⁵. This large, multicenter, prospective cohort study enrolled patients with clinical stage I endometrial cancer and determined the sensitivity and negative predictive value of SLN mapping in detecting metastatic disease compared with the gold standard complete lymphadenectomy. FIRES investigators injected indocyanine green into the cervix and performed SLN mapping followed by pelvic lymphadenectomy with or without para-aortic lymphadenectomy. Pathologists assessed SLNs for metastases by hematoxylin and eosin (H&E) staining. If the results were negative by H&E, they performed cytokeratin immunohistochemistry (IHC) ultra-staging. They found an impressive 97% sensitivity and 99.6% negative predictive value⁵.

In the FIRES trial, approximately 14% of patients did not map and approximately 34% of patients mapped to only one side, which would necessitate approximately 48% of patients to have a lymphadenectomy due to mapping failure. Conversely, over half of patients with endometrial cancer can be spared a complete staging lymphadenectomy with SLN mapping. Some have raised concerns regarding missing isolated para-aortic metastases in SLN mapping¹², especially in those with the highest prevalence of para-aortic metastases (high-grade and deeply invasive tumors)¹³. One study showed that SLN biopsies detected more metastatic nodal disease compared with selective pelvic and para-aortic lymphadenectomies¹⁴. This was recently confirmed in a meta-analysis of 3,536 patients in six studies, showing higher positive pelvic nodal detection rates, no difference in para-aortic nodal detection rates, and no difference in overall recurrence or nodal recurrence rates in SLN compared with complete lymphadenectomy¹⁵. Additionally, the risk of para-aortic recurrence in patients with positive pelvic nodes and unassessed para-aortic nodes is only 4%¹⁶. In regard to high-risk histology, a prospective study examined women with grade 3 (endometrioid, serous, clear cell, and carcinosarcoma) endometrial cancer by SLN biopsy followed by full pelvic and para-aortic lymphadenectomy; they found 95% sensitivity and 98% negative predictive value and bilateral mapping rates of 58% and unilateral mapping rates of 40%, supporting the use of SLN in high-risk patients¹⁷. Additionally, 27% of patients in the FIRES trial and 27 to 100% of patients in the six studies included in the meta-analysis had high-grade histology^5,15.

In order to adequately evaluate nodal status, patients who do not map to one or both sides should have unilateral or bilateral complete lymphadenectomies, respectively. Additionally, all enlarged or suspicious lymph nodes should be removed, regardless of SLN mapping¹⁸. According to the Society of Gynecologic Oncology (SGO) Clinical Practice Statement, individual surgeons should consider performing full lymphadenectomy while their personal experience is accrued (20 to 30 cases) to ensure that they, and their pathologists, demonstrate acceptable sensitivity and negative predictive value. The decreased morbidity, high sensitivity, and high negative predictive value make SLN evaluation an appealing and standard option, which is supported by both the National Comprehensive Cancer Network (NCCN) and the SGO; however, some surgeons may choose to limit its use to intermediate-risk populations until further data accrue. Additionally, data to guide adjuvant treatment (that is, adding extended-field radiation) in patients with positive pelvic lymph nodes with unknown para-aortic nodal status are needed.

Adjuvant treatment for high-risk endometrial cancers

Most endometrial cancers will be cured with surgery alone. However, many advanced-stage and some early-stage cancers will recur. Multiple studies have characterized the risk of post-surgical recurrence and tried to identify adjunctive therapies to reduce it. High-risk factors include tumor grade, tumor size, cell type, depth of myometrial invasion, presence of lymphovascular space invasion (LVSI), and stage^13,19–21. Because of randomized controlled trials and retrospective series comparing adjuvant chemotherapy with radiation therapy, there has been optimism that chemotherapy may reduce the risk of distant recurrence and improve survival for patients with high-risk disease. However, as outlined below, chemotherapy has shown improvement in overall survival (OS) in advanced-stage disease but not in high-risk, early-stage patients. Radiation therapy is known to be quite effective in achieving local control, but improvement in OS has not been realized^6,19,22,23. Recently, the results of three important trials that evaluated combined radiation and chemotherapy aimed at reducing rates of both local and distant recurrence were presented: the results of GOG 249 were published in the Journal of Clinical Oncology in April 2019²⁴, Gynecologic Oncology Group (GOG) protocol 258 was presented in abstract form²⁵, and the results of Post-Operative Radiation Therapy in Endometrial Carcinoma 3 (PORTEC 3) were published in Lancet Oncology in March 2018²⁶. GOG 249 investigated the role of adjuvant chemotherapy and vaginal brachytherapy in early-stage, high-risk patients²⁴, whereas GOG 258^24–26 and PORTEC 3^24–26 examined the role of combined chemotherapy and radiation in higher-risk cohorts. Table 1 outlines the details of these and other key trials that guide adjuvant therapy in high-intermediate and high-risk endometrial cancer.

Treatment of recurrent endometrial cancers

Systemic recurrence of endometrial cancer is generally considered incurable and is associated with a poor prognosis. A variety of GOG studies have looked at single agents (etoposide, paclitaxel, dactinomycin, liposomal doxorubicin, pyrazoloacridine, topotecan, oxaliplatin, irofulven, flavopiridol, and bevacizumab) and alternating courses of megestrol acetate and tamoxifen in recurrent endometrial cancer, and response rates range from 0 to 31% and 6-month PFS rates from 0 to 43%^38–48. Hormonal therapy is an appealing option in recurrent endometrial cancer given that many endometrial cancers are hormonally driven and the relative lack of toxicity with hormonal therapy. However, response rates and PFS have been disappointing. Given cross-regulation between the estrogen receptor (ER) and PI3K/AKT/mTOR pathway and recent success of combination everolimus and an aromatase inhibitor in aromatase-refractory breast cancer, Slomovitz et al. hypothesized that combining everolimus and letrozole would result in improved response rates and PFS^49,50. In their phase II study, they found a clinical benefit rate of 40%, and 23% of patients obtained a complete response⁴⁹. The authors found that their oral regimen of everolimus 10 mg daily with letrozole 2.5 mg daily was well tolerated, and no patients discontinued the study because of toxicity. Metformin use and CTNNB1 mutations were associated with increased response rates, PFS, and OS, although these findings were not statistically significant. Of note, none of the patients (n = 7) with serous histology had a response. Patients with ER-positive and progesterone receptor (PR)-positive tumors had higher response rates than receptor-negative tumors, although 33% of ER-negative patients and 25% of PR-negative patients had responses. Given these promising results, the investigators evaluated this combined regimen compared with alternating megestrol acetate and tamoxifen in GOG 3007, which was recently published as an abstract⁵¹. They reported a 24% response rate in the everolimus/letrozole arm (PFS 6.4 months and OS 20.0 months) and a 22% response rate in the progestin/tamoxifen arm (PFS 3.8 months and OS 16.6 months). Hormonal combined therapies remain appealing options for recurrent endometrial cancer, and biomarkers are needed to better predict which patients will benefit.

Understanding the molecular biology of endometrial cancer

Advances in our classification of endometrial cancer with molecular profiling may enable us to better stratify risk and recommended therapy. MSI and polymerase-ε (POLE) lead to neoantigens, fragments of proteins expressed by tumor cells that may sensitize the immune system⁵². The molecular profile of a tumor may better define its prognosis and response to therapy than histology and stage alone. Talhouk et al.^53–55, The Cancer Genome Atlas Collaborative (TCGA)^53–55, and the GOG 210 investigators^53–55 have developed different molecular classification schema to better stratify patient risk.

The TCGA project determined four molecularly defined subgroups of endometrial cancer, which yielded excellent prognostic results⁵³. However, the methods required for classification are currently quite expensive and require special handling of the tissue, limiting applicability. Talhouk et al. developed a more pragmatic method, the Proactive Molecular Risk Classifier for Endometrial Cancer (ProMisE) classification, to identify molecularly distinct subgroups with a prognostic signature similar to that of the TCGA classification scheme^54,56. The four groups were MMR-deficient (MMR-D), POLE exonuclease domain mutations (POLE EDMs), p53 abnormal, and p53 wild-type. They assessed tumors in a step-wise fashion. The first assessment was based on MMR status, and MMR-D tumors were classified as MMR-D and MMR intact tumors were tested for POLE mutations. POLE mutants were classified as POLE EDM, and POLE wild-type tumors were assessed for p53 status. MMR intact, POLE wild-type tumors were classified as p53 wild-type or p53 abnormal (null/missense mutations). They validated the ProMisE classifier in a large confirmation cohort series and found that POLE EDMs had the best prognosis and p53 mutants the worst. The ProMisE classification system improved outcome prediction compared with European Society for Medical Oncology risk stratification, detecting those at risk for Lynch syndrome and potentially guiding clinical management⁵⁶.

GOG 210, an NRG Oncology/GOG study of the molecular classification for risk prediction in endometrioid endometrial cancer, was published in Gynecologic Oncology in August 2017⁵⁵. The investigators sought to develop a classification system for endometrioid endometrial cancers, the most common and potentially challenging endometrial cancer histology to classify. They assessed tumors for mismatch repair (MMR) defects (MSI, MMR IHC, and MLH1 methylation), POLE mutations, and loss of heterozygosity. Using four classifications (MMR deficient, copy number altered (CNA), copy number stable, and POLE mutant), they were able to stratify patients by PFS and OS. Their classification system remained statistically significantly related to clinical outcomes in multivariable analyses. CNA had the worst PFS and cancer-specific survival, whereas the POLE group had the best outcomes. The authors advocate prospective validation of this system and recommend that clinicians consider using it to universally screen for Lynch syndrome and identify additional prognostic information to guide treatment decisions.

Biomarkers and molecular alterations are increasingly being used to guide systemic therapy in the recurrent setting. Table 2 outlines common biomarkers in endometrial cancer, noting differences between type I (grade 1 or 2 endometrioid) and type II (grade 3 endometrioid, serous, clear cell, and carcinosarcoma) endometrial cancers as well as potential targeted therapies (Table 2).

In our practice, we test all primary endometrial cancers with IHC for MMR proteins and, if findings are consistent with Lynch syndrome (absence of staining and no evidence of promoter methylation), proceed to confirmatory testing. In addition, we perform MSI testing if there is a high clinical suspicion for Lynch syndrome even with intact proteins. We perform IHC for HER2/neu status in all serous cancers and confirm overexpression with fluorescence in situ hybridization. We have not yet adapted routine molecular classification; however, we optimistically await further results to guide systemic therapy by molecular class. In the recurrent setting, we recommend molecular profiling, testing for immunotherapy biomarkers (programmed death ligand-1 (PD-L1), MSI, and tumor molecular burden), HER2/neu status in serous cancer, and determining ER and PR status.

Immunotherapy in endometrial cancer

Programmed death 1 (PD-1) is an immune checkpoint receptor expressed on tumor-infiltrating T cells that, when activated by PD-L1, blocks T-cell activation and enables immune evasion^84,88–90. About one half of endometrial tumors evaluated in KEYNOTE-028, a phase 1b study, were PD-L1–positive⁸⁴. In this study, investigators evaluated the safety and efficacy of pembrolizumab, an anti–PD-1 monoclonal antibody, in patients with PD-L1–positive tumors. In this heavily pretreated cohort of women, 24 women were enrolled; three patients (13.0%) achieved a partial response and three patients achieved stable disease (median duration of 24.6 weeks). Over half of the patients developed a treatment-related toxicity, of which fatigue, pruritus, pyrexia, and anorexia were the most common. Pembrolizumab was well tolerated overall and resulted in durable response in a subset of patients; however, there remains a need to better identify biomarkers to predict durable response. In 2017, the US Food and Drug Administration (FDA) approved pembrolizumab for all advanced solid tumors that are MSI-high or MMR-D.

Makker et al. recently published an interim analysis of their phase 2 study of pembrolizumab plus lenvatinib in biomarker-unselected advanced endometrial cancer⁹¹. In the unselected cohort (85% microsatellite stable and 25% PD-L1–positive), 39.6% of patients responded with durable responses (65% had responses greater than 6 months, and median duration of response was not yet achieved)⁹¹. This trial led to FDA breakthrough therapy designation for pembrolizumab with lenvatinib in advanced endometrial cancer. The group has started enrollment for a phase 3 randomized clinical trial of pembrolizumab and lenvatinib versus doxorubicin or paclitaxel. Further research is ongoing to identify a combinatorial approach with other immunotherapies, radiation, or systemic treatment in order to augment the effects of immunotherapy in endometrial cancer.

Future directions and current clinical trials

Although there have been great advances in endometrial cancer in the past few years, there remains much left to do as we continue to refine our understanding and treatment of endometrial cancer. Focus will be on better classifying tumors in order to more appropriately enroll patients in clinical trials specific to that profile. We need to identify better biomarkers to guide personalized therapy as well as to explore combinatorial regimens for targeted therapeutics, hormonal therapy, and immunotherapy. Featured actively recruiting trials include a randomized phase III trial in women with endometrial cancer with high-intermediate risk factors to investigate the role of an integrated clinicopathological and molecular risk profile to guide adjuvant therapy (PORTEC 4-a, ClinicalTrials.gov Identifier: NCT03469674); a phase II trial of paclitaxel, carboplatin, and pembrolizumab in measurable advanced or recurrent endometrial cancer (ClinicalTrials.gov Identifier: NCT02549209); a phase II trial of vaginal cuff brachytherapy followed by adjuvant chemotherapy with carboplatin and dose-dense paclitaxel in patients with high-risk endometrial cancer (ClinicalTrials.gov Identifier: NCT03189446); a phase I study of the Wee I kinase inhibitor AZD1775 in combination with radiotherapy and cisplatin in cervical, upper vaginal, and uterine cancers (ClinicalTrials.gov Identifier: NCT03345784); a randomized phase II study comparing single-agent olaparib, single-agent cediranib, or combination cediranib/olaparib in women with recurrent, persistent, or metastatic endometrial cancer (GY012, ClinicalTrials.gov Identifier: NCT03660826); and multiple trials combining checkpoint inhibitors, immunotherapy, tyrosine kinase inhibitors, poly ADP ribose polymerase (PARP) inhibitors, and/or anti-angiogenic inhibitors (TSR042 Garnet study, UC1805, NRG-GY018, AMANDA study, KEYNOTE-077/ECHO-202, KEYNOTE-775, and ROSCAN).