Tumor board presentation of a woman with metastatic, hormone receptor-positive, mismatch repair-deficient endometrial cancer

A 65-year-old woman presented to an outside hospital with vaginal bleeding in 2009. Dilatation and curettage showed endometrial endometrioid carcinoma, International Federation of Gynecology and Obstetrics (FIGO) grade 2. She underwent exploratory laparotomy with pelvic washings, total abdominal hysterectomy with bilateral salpingo-oophorectomy, bilateral pelvic and para-aortic lymphadenectomy, and an omental biopsy. Pathology was read as uterine carcinosarcoma, homologous type, invading 3.5 cm into a 4-cm-thick myometrium, with involvement of the endocervical epithelium (no stromal involvement) and lymphovascular invasion. All 14 of the sampled lymph nodes (6 left pelvic, 1 left common iliac, 1 left aortic, 3 right pelvic, 2 right common iliac, and 1 right aortic) were negative for carcinoma. The tumor-node-metastasis (TNM) pathologic stage using current staging (2018 FIGO Cancer Report¹) was pT1b N0 American Joint Committee on Cancer [AJCC] Cancer Staging Manual, eighth edition), FIGO stage IB. Postoperative positron emission tomography/computed tomography (CT) imaging studies demonstrated no evidence of disease.

The patient was otherwise healthy. Her father had lung cancer, but there was no other history of cancer in the family. She received pelvic radiotherapy and 4 cycles of paclitaxel 135 mg/m² plus ifosfamide 1600 mg/m², with mesna and granulocyte-colony–stimulating factor support.

Five years later (at age 71 years), she developed abdominal pain and vomiting and was managed conservatively for a partial small bowel obstruction. CT scans of the abdomen and pelvis showed a large, heterogeneously enhancing mass in the left suprarenal space, measuring 8.9 × 8.6 cm, likely originating from the left adrenal gland. A lobular mass in the posterior left lung base measuring 3.2 × 1.6 cm was also noted. A CT scan of the chest revealed multiple, bilateral pulmonary nodules, the largest measuring 2.4 × 2.2 cm, some with central cavitation, most consistent with metastatic disease. A fine-needle aspirate of the suprarenal mass showed poorly differentiated carcinoma with extensive necrosis.

At that time, the patient presented to a university hospital. Review of the original pathology reclassified the tumor as endometrial endometrioid carcinoma, FIGO grade 3 (Fig. 1A), with the suprarenal mass morphologically consistent with the patient's primary. Estrogen receptor (ER) and progesterone receptor (PR) were strongly and diffusely positive (Fig. 1B,C).

Subtyping of high-grade endometrial carcinomas is diagnostically challenging for pathologists, with even expert gynecologic pathologists reaching consensus in only 63% to 72% of tumors.^{2, 3} The most common disputes include endometrioid versus serous carcinoma, serous versus clear cell carcinoma, and endometrioid/clear cell/undifferentiated carcinoma versus carcinosarcoma.² Morphologically, high-grade endometrial carcinoma can show extensive solid growth, whereas low-grade endometrioid carcinoma may have a spindled morphology,⁴ both of which may mimic carcinosarcoma with a homologous sarcomatous component. Although immunohistochemistry can be useful in subtyping endometrioid, serous, and clear cell carcinomas, it is generally unhelpful for distinguishing carcinosarcomas from pure endometrial carcinomas. Because carcinosarcomas are considered a subtype of endometrial carcinoma⁵ rather than a mixed epithelial and mesenchymal tumor, as previously hypothesized, it is not surprising that the sarcoma (as well as the carcinoma) portions of the tumor may show cytokeratin and PAX8 expression, eliminating the utility of these stains in differentiating between carcinosarcoma and pure endometrial carcinoma.^{6, 7}

The College of American Pathologists recommends performing ER immunohistochemistry on stage III, IV, and recurrent endometrial carcinomas to predict response to endocrine therapy.⁸ Unlike in breast carcinoma, there is not a uniform reporting system for hormone receptors in endometrial carcinoma; however, it is recommended that the number of positive cells and staining intensity be documented.⁹

The patient was presented with the option of chemotherapy with carboplatin plus paclitaxel. She declined chemotherapy. She was also presented with the option of participation in a clinical trial, Gynecologic Oncology Group (GOG) 3007 (a randomized phase II trial of everolimus and letrozole or hormonal therapy [medroxyprogesterone acetate/tamoxifen], in women with advanced, persistent, or recurrent endometrial cancer; Clinicaltrials.gov identifier NCT02228681). She elected to participate in GOG 3007 and was randomized to therapy with medroxyprogesterone acetate/tamoxifen. She was treated with oral tamoxifen 20 mg twice daily (continuously), with the addition of medroxyprogesterone acetate 200 mg daily on alternating weeks (days 8-14 and 22-28 of a 28-day cycle). She had a partial response to therapy (Fig. 2) and remained on study for 19 cycles (about 1.5 years), at which time she had mild disease progression in the lung. She was switched to single-agent letrozole without further response.

Of note, had her tumor not been reclassified as an endometrioid endometrial carcinoma, she would have been ineligible for GOG 3007 because that trial, as well as many current ongoing trials for women with endometrial cancer (including those testing immunotherapy), excludes carcinosarcomas.

Whereas, historically, progestins were front-line therapy for metastatic endometrial cancer, currently, they are most commonly used in the second or later lines, although it is not clear that their activity is fully maintained in this setting. Single-agent progestins (most frequently oral medroxyprogesterone acetate or megestrol acetate),^{10, 11} aromatase inhibitors (letrozole, anastrozole),^{12, 13} selective ER modulators (tamoxifen, arzoxifene),^{14, 15} fulvestrant,^{16, 17} or combination therapies, such as progestin/tamoxifen^{18, 19} or everolimus (an mTOR inhibitor)/letrozole,^{20, 21} have all been investigated and have shown some efficacy. A 2017 meta-analysis evaluated 16 studies that included second-line endocrine therapy and reported an objective response rate of 18.5%.²² When patients with stable disease were also included, the clinical benefit rate was 36%. By comparison, a randomized trial of tamoxifen versus anastrazole for first-line therapy of metastatic breast cancer (almost one-half of patients had hormone receptor-positive tumors and, in slightly greater than one-half, the tumor hormone receptor status was unknown) demonstrated an objective response rate of approximately 33% in both arms.²³ The current National Comprehensive Cancer Network guidelines include multiple endocrine therapy options for endometrial cancer, but they note that these therapies are typically reserved for patients who have low-grade endometrioid tumors with more indolent behavior.²⁴

More recently, the combination of everolimus plus letrozole and the combination of tamoxifen plus medroxyprogesterone acetate were evaluated in GOG 3007, the study in which our patient enrolled. Although this was a randomized, phase 2 study of women with advanced or recurrent endometrial cancer and included both first-line and second-line settings, it is important to note that the 2 regimens were not directly compared. Preliminary results were presented at the 2018 Society for Gynecologic Oncology Annual Meeting. The progression-free survival was 6.3 months for the everolimus/letrozole arm and 3.8 months for the medroxyprogesterone acetate/tamoxifen arm.²⁵ However, in patients who had received prior chemotherapy, progression-free survival was 3.3 months and 3.2 months, respectively. Overall survival for the medroxyprogesterone acetate/tamoxifen group was 16.6 months but was not yet calculable for the everolimus/letrozole group. In addition, a phase 2 trial of letrozole plus ribociclib (a cyclin-dependent kinase inhibitor) was recently published. Of the 20 patients with endometrial cancer enrolled on the trial, 55% were still on treatment at 12 weeks.²⁶ Similar to the studies discussed above, those patients who had low-grade endometrioid tumors seemed to derive the most benefit.

Discussion regarding the use of endocrine therapy before chemotherapy in the metastatic setting is ongoing. A 2017 meta-analysis evaluated results from 23 studies of hormonal therapy in the first-line setting and found that the overall response rate (ORR) was highest in patients who had hormone receptor-positive tumors, at 32.5%,²² and that low-grade tumors had higher response rates than high-grade tumors. The ORR for ER-negative tumors was only 9.2%. The authors also reported that single-agent aromatase inhibitors produced the lowest response rates of the endocrine therapy classes, with a mean ORR of only 8.6%. For patients treated on GOG 3007 in the first-line setting, the objective response rate was 53% for everolimus/letrozole compared with 43% for tamoxifen/medroxyprogesterone acetate, and progression-free survival was 21.6 months in the everolimus/letrozole arm but only 6.6 months in the medroxyprogesterone acetate/tamoxifen arm.²⁵ Given the limitations of cross-trial comparisons, it is difficult to determine whether an endocrine regimen may be similar to front-line chemotherapy in patients with endocrine-sensitive tumors. A head-to-head trial of front-line chemotherapy versus endocrine therapy seems unlikely at this point, but we believe that there is a subset of patients with hormone receptor-positive endometrioid tumors for whom first-line endocrine therapy may be appropriate and may be as effective as and better tolerated than chemotherapy.

Most ongoing trials are looking at novel combinations that include hormonal therapy rather than new single-agent hormonal therapies. Several studies are evaluating the use of cyclin-dependent kinase inhibitors (palbociclib, ribociclib) in combination with letrozole (Clinicaltrials.gov identifiers NCT03675893, NCT02730429, NCT03008408, and NCT04049227) or fulvestrant (Clinicaltrials.gov identifier NCT03643510). A recent window-of-opportunity trial evaluating the combination of medroxyprogesterone acetate with entinostat (a histone deacetylase inhibitor) showed no difference in PR expression from before to after treatment, although some decrease in Ki-67 expression was seen, suggesting that this combination could be of interest in the future.²⁷

As in breast and prostate cancer, hormones play an important role in endometrial cancer development and treatment. During the menstrual cycle, proliferative effects of estrogens on the endometrium are counteracted by progestins in a strictly regulated balance. In case of excessive estrogen secretion, proliferative effects may not be sufficiently controlled by progesterone, and endometrial cancer can develop; unopposed estrogen exposure in the postmenopausal setting also increases the risk for endometrial cancer.²⁸ In 1961, Kelley and Baker first described the use of synthetic progestins to treat 21 patients with recurrent endometrial cancer.²⁹ Later, other hormonal drugs, including tamoxifen and aromatase inhibitors, were shown to exhibit antitumor effects in endometrial cancer. The introduction of endocrine treatment also initiated efforts to identify predictive biomarkers for the selection of patients who would benefit most from such therapy, which is generally less toxic than chemotherapy.

At the time of progression on letrozole, the patient remained asymptomatic and still did not desire chemotherapy. Further testing on the tumor was performed at this time because mismatch repair (MMR) testing was not standard at the time of her original presentation. The endometrial carcinoma showed loss of expression for MLH1 (Fig. 1D) and PMS2, retained expression for MSH2 and MSH6, and exhibited MLH1 promoter hypermethylation. Next-generation sequencing (NGS) showed mutations in MAP2K1 (MEK1), PIK3CA, PTEN, and MED12. She was switched to single-agent pembrolizumab. She did not experience any toxicity and had further disease shrinkage (Fig. 3); after 2 years of therapy, there is no evidence of progression.

The National Comprehensive Cancer Network currently recommends universal MMR protein/microsatellite instability (MSI) testing for all endometrial carcinomas,⁴⁴ regardless of patient age, personal/family history, or tumor subtype. Approximately 30% of endometrial carcinomas are MMR-deficient, thus identifying this subset helps screen for Lynch syndrome (2%-6% of MMR-deficient tumors),^{45, 46} guide management based on molecular classification,⁴⁷ and recognize patients who are eligible for targeted therapy with immune checkpoint inhibitors.⁴⁸ The 2 techniques for detecting MMR deficiency include immunohistochemistry for MMR proteins (MLH1, PMS2, MSH2, and MSH6) and molecular testing for MSI. Because these 2 methods are highly concordant (94%; P < .001),⁴⁹ immunohistochemistry is generally the preferred screening test because of its cost effectiveness and accessibility.

MMR proteins repair DNA base-base mismatches and small insertion-deletions (including those that alter microsatellite length) to prevent propagation of these mutations. They form heterodimers, pairing MLH1 with PMS2 and pairing MSH2 with MSH6. PMS2 and MSH6 must bind to their respective partners; however, MLH1 and MSH2 can form heterodimers with other proteins.⁵⁰ In normal cells as well as MMR-proficient carcinomas, all 4 MMR proteins show diffuse (typically strong) nuclear expression, whereas defects in one or more proteins result in loss of nuclear staining for that protein(s). Four abnormal staining patterns may be observed: 1) loss of MLH1 and PMS2, 2) loss of PMS2, 3) loss of MSH2 and MSH6, and 4) loss of MSH6. All patterns may occur in either the germline or somatic setting, but combined MLH1 and PMS2 loss is generally a result of sporadic MLH1 promoter hypermethylation, which can be confirmed by methylation-specific polymerase chain reaction (PCR)/multiplex ligation-dependent probe amplification.⁵¹ If MLH1 promoter hypermethylation is absent or another abnormal staining pattern is present, genetics counseling with germline testing is recommended.

Alternatively, MSI testing directly detects unstable microsatellites (ie, microsatellites with variable length between tumor cells) and is either PCR-based or NGS-based.⁵² The PCR-based assay requires matched tumor and normal tissue, tests several (from 5 to 10) microsatellite loci, and classifies each locus as stable or unstable. The tumor is then assigned an overall MSI status (MSI-high, MSI-low, microsatellite-stable) based on the percentage of unstable loci. However, because this test was originally designed for colorectal carcinoma screening, it is not as sensitive for endometrial carcinoma,⁵³ Furthermore, PCR-based testing is less sensitive for detecting MSH6 deficiency, which is more common in endometrial carcinomas than in colorectal carcinomas.^{54, 55} In contrast, NGS-based tests can be performed on tumor without matched normal tissue, evaluate hundreds of microsatellite loci, and result in a quantitative score, which then can be grouped into 3 categories: MSI-H, MSI-indeterminate, and microsatellite-stable.⁵² Although NGS-based MSI testing circumvents many of the shortcomings of PCR-based assays, NGS is often costly and has a longer turnaround time than the other modalities.

The Cancer Genome Atlas categorized endometrial cancers into 4 distinct molecular subtypes: POLE/ultramutated, MSI-H, copy number low, and copy number high.⁵⁶ These subtypes have been correlated with the risk of recurrence. Up to 30% of primary and recurrent endometrial tumors are MSI-H (MMR-deficient).^{57, 58} Cancers in the MSI-H category, as discussed above, have a deficiency in their DNA MMR proteins, which leads to a high DNA mutational burden. It is believed that this increases the tumor immunogenicity by creating more neoantigens that can be identified and targeted by T cells and is responsible for the efficacy of immune checkpoint inhibitors for the treatment of these tumors.⁵⁹ Immune checkpoint inhibitors are also likely to be effective in the treatment of ultramutated POLE-mutant tumors, although these are less commonly found in the metastatic setting.⁶⁰

Pembrolizumab, the first site-agnostic therapy approved by the US Food and Drug Administration (FDA), was approved in 2017 for the treatment MSI-H/MMR-deficient solid tumors⁶¹ regardless of tumor origin. It targets the PD-1 (programmed death-1) receptor on the surface of T cells, thereby blocking the immune-suppressing ligands PD-L1 and PD-L2 (programmed death ligand-1 and programmed death ligand-2, respectively) on the cancer cells from interacting with PD-1 and inactivating the T cells. The KEYNOTE 158 trial (Clinicaltrials.gov identifier NCT02628067), which enrolled patients with pretreated MSI-H/MMR-deficient tumors of multiple histologies, included 49 patients with endometrial cancer. Of these, 8 patients had a complete response, and 20 had a partial response, yielding an overall response rate of 57.1%. Response duration at the time of reporting was from 2.9 to ≥27 months.⁶² A second, small, prospective trial of pembrolizumab only in patients with MSI-H/MMR-deficient, recurrent endometrial cancers (n = 25) who had received at least one prior line of chemotherapy found an overall response rate of 58%; interestingly, the response rate was 100% in the 6 patients with somatic loss of MMR proteins and only 44% in patients with MLH1 promotor methylation.⁶³ Subsequently, dostarlimab (an anti–PD-1 monoclonal antibody) has also been approved, both specifically for second-line treatment of MSI-H/MMR-deficient endometrial cancer and for MMR-deficient solid tumors of any histology that have progressed on or after prior therapy and for which there is no satisfactory alternative treatment option. Although immune checkpoint inhibitors can produce severe or fatal autoimmune toxicity, for many patients, such as ours, they are well tolerated.^{64, 65}

The current FDA approval for pembrolizumab in patients with MSI-H/MMR-deficient tumors, including endometrial cancer, is for those whose cancer has progressed after prior treatment and who have no satisfactory alternative treatment options. An obvious direction of research is the use of immune checkpoint inhibitors in patients with endometrial cancer who have not received prior cytotoxic therapy. On June 29, 2020, the FDA approved pembrolizumab for first-line use in the treatment of patients with MSI-H/MMR-deficient colorectal cancer based on results of the KEYNOTE-177 trial (Clinicaltrials.gov identifier NCT02563002), which randomized 307 patients with MSI-H/MMR-deficient colorectal cancer to receive either pembrolizumab 200 mg intravenously every 3 weeks or investigator's choice of 5-fluorouracil-based therapy with or without bevacizumab or cetuximab. Pembrolizumab was given to patients on the chemotherapy arm at the time of progression. The median progression-free survival was 16.5 months for those receiving pembrolizumab versus 8.2 months for those receiving chemotherapy (hazard ratio, 0.60; P = .0002). Among responders, 83% of patients who received pembrolizumab versus 35% of those who received chemotherapy had ongoing responses at 24 months.⁶⁶ Toxicity was lower in the pembrolizumab arm, with 56% versus 78% of patients having grade 3 adverse events, respectively. Fourteen percent of patients in the pembrolizumab group and 12% of those in the chemotherapy group discontinued therapy because of adverse events. Health-related quality of life showed a clinically meaningful improvement from baseline to week 18 in scores on the European Organization for Research and Treatment of Cancer Quality of Life Questionnaire Core 30 global health status/quality of life for pembrolizumab versus chemotherapy (P = .0002), and the median time to deterioration for global health status/quality of life was longer with pembrolizumab than with chemotherapy (hazard ratio, 0.61; P = .019).⁶⁷ This paradigm (chemotherapy vs single-agent immunotherapy up front in MSI-H/MMR-deficient tumors) is also being explored in patients with endometrial cancer. The GOG-3064/ENGOT-en15 trial should open shortly: this is a phase 3 randomized trial of pembrolizumab versus platinum doublet chemotherapy limited to participants with MMR-deficient advanced or recurrent endometrial carcinoma in the first-line setting.

Although single-agent immune checkpoint inhibitors produce a low response rate in patients with advanced endometrial cancer that is not MSI-H/MMR-deficient, a combination therapy, pembrolizumab plus lenvatinib (a tyrosine kinase inhibitor targeting vascular endothelial growth factor receptor as well as other receptor tyrosine kinases) has recently been FDA-approved in this setting. A response rate of 36.2% was reported with the combination for microsatellite-stable tumors.⁶⁸ The addition of pembrolizumab to standard front-line chemotherapy (paclitaxel and carboplatin) is also being investigated (Clinicaltrials.gov identifier NCT03914612) in a randomized trial as is the addition of dostarlimab (Clinicaltrials.gov identifier NCT03981796)^{69, 70} and durvalumab (Clinicaltrials.gov identifier NCT04269200); all of these trials are open to patients regardless of MSI tumor status. It is hoped that the results of these trials will help improve outcomes for women with advanced endometrial cancer.

Gini F. Fleming's institution receives funding for institutional trials from Tesaro/GSK, Compugen, Incyte, AbbVie, Eisai, Celldex, AstraZeneca, Corcept, Plexxicon, and Astellas for which she serves as principal investigator; and honoraria from UpToDate outside the submitted work. Katherine C. Kurnit participate on a Data Safety Monitoring Board or Advisory Board at LEAP Therapeutics through the Gynecologic Oncology Group Foundation outside the submitted work. Jennifer A. Bennett, Kathryn A. Mills, and Willem Jan vanWeelden made no disclosures.