Objective Women with Lynch syndrome have a risk up to 40–60% of developing endometrial cancer, which is higher than their risk of developing colorectal or ovarian cancer. To date, no data on the outcomes of patients with Lynch syndrome diagnosed with non-endometrioid endometrial cancer are available. The goal of this study was to evaluate the outcome of patients with Lynch syndrome diagnosed with non-endometrioid endometrial cancer.
Methods Data from consecutive patients diagnosed with Lynch syndrome and with a histological diagnosis of non-endometrioid endometrial cancer were retrospectively collected in two referral institutes in Italy. A case–control comparison (applying a propensity matching algorithm) was performed in order to compare patients with proven Lynch syndrome and controls. Inclusion criteria were: (a) histologically-proven endometrial cancer; (b) detection of a germline pathogenic variant in one of the MMR genes; (c) adequate follow-up. Only carriers of pathogenic or likely pathogenic variants (ie, class 5 and 4 according to the InSiGHT classification) were included in the study. Survival outcomes were assessed using KaplanMeier and Cox models.
Results Overall, 137 patients with Lynch syndrome were collected. Mean patient age was 49.2 (10.9) years. Genes involved in the Lynch syndrome included MLH1, MSH2, and MSH6 in 43%, 39%, and 18% of cases, respectively. The study population included 27 patients with non-endometrioid endometrial cancer, who were matched 1:2 with patients with sporadic cancers using a propensity matching algorithm. After a median follow-up of 134 months (range 1–295), 2 (7.4%) of the 27 patients developed recurrent disease (3 and 36 months) and subsequently died of disease (7 and 91 months). Patients diagnosed with Lynch syndrome experienced better disease-free survival (HR 7.86 (95% CI 1.79 to 34.5); p=0.006) and overall survival (HR 5.33 (95% CI 1.18 to 23.9); p=0.029) than controls.
Conclusions Non-endometrioid endometrial cancer occurring in patients with Lynch syndrome might be associated with improved oncologic outcomes compared with controls. Genetic/molecular profiling should be investigated in order to better understand the mechanism underlying the prognosis.
- endometrial neoplasms
- Lynch syndrome II
- uterine neoplasms
Statistics from Altmetric.com
Patients with non-endometrioid endometrial cancer have to be screened for Lynch syndrome
Patients with Lynch syndrome-related non-endometrioid endometrial cancer experienced a better oncologic outcome than sporadic cases
Further study evaluating the omission of adjuvant treatments in MSI/MMR-deficient tumors are warranted
Uterine cancer is one of the most common gynecological cancers in developed countries. It is estimated that more than 61 000 new cases will be diagnosed in the year 2019 in the USA.1 2 Estimated uterine cancer-related mortality increased dramatically in the last decade, from 7780 deaths in the year 2009 to 12 160 deaths in the year 2019.1 2 Endometrial cancer is characterized by an overall good prognosis, with a 5-year survival rate of 80% for patients following appropriate treatment.3 Classically, risk factors predicting poor survival outcomes included age, depth of myometrial invasion, nodal involvement, extra-uterine spread, and histology.4 This latter feature, per se, has important implications in the prognosis.4 It is estimated that, compared with endometrioid endometrial cancer, non-endometrioid endometrial cancer is characterized by aggressive biological features, metastatic stage at presentation, and intrinsic chemoresistance.5 6
Interestingly, endometrial cancer might be related to a genetic predisposition. Overall, hereditary causes contribute to 2–10% of endometrial cancer cases.7 Lynch syndrome is an autosomal dominant genetic disorder characterized by a high risk of developing cancer, in particular colorectal and endometrial cancer. Lynch syndrome is caused by the presence of a germline pathogenic variant in one of the four DNA mismatch-repair (MMR) genes (mutL homolog 1 (MLH1), mutS homolog 2 (MSH2), mutS homolog 6 (MSH6), or postmeiotoic segregation increased 2 (PMS2)).8 Women with Lynch syndrome have a risk of 40–60% of developing endometrial cancer, higher than their risk of developing colorectal or ovarian cancer.8 Generally, Lynch syndrome-related endometrial cancer is characterized by endometrioid histology and an overall good prognosis.8 However, although uncommon, Lynch syndrome-related non-endometrioid endometrial cancer might occur. To date, no data investigating the outcomes of patients with Lynch syndrome diagnosed with non-endometrioid endometrial cancer are available. We aim to review the outcomes of non-endometrioid endometrial cancer occurring in the context of Lynch syndrome. We sought to compare the outcomes of Lynch syndrome-related non-endometrioid endometrial cancer with a selected cohort of patients with sporadic non-endometrioid endometrial cancer. As a secondary endpoint we aimed to assess the factors predicting adverse outcomes among patients with non-endometrioid endometrial cancer.
In this retrospective study we collected data for consecutive patients diagnosed with endometrial cancer after undergoing surgery with proven Lynch syndrome at two Italian referral centers: the Fondazione IRCCS Istituto Nazionale dei Tumori in Milano and the Ospedale di Circolo Fondazione Macchi, Università degli Studi dell’Insubria in Varese. The study was approved by the Institutional Review Boards of both institutions.
All patients included in the study received treatment in the two institutions from January 1973 to December 2017. All patients who were alive signed written consent for research purposes. Patients who did not consent to use their clinical information for research purposes were excluded. Inclusion criteria were: (a) histologically-proven endometrial cancer; (b) detection of a germline pathogenic variant in one of the MMR genes; (c) appropriate follow-up information.
Patients were identified over the years among members of the families enrolled in the Registry of Hereditary Digestive Tract Tumor Unit of Fondazione IRCCS Istituto Nazionale Tumori, Milan, Italy and the Registry of Hereditary Cancers of University of Insubria. The presence of MMR pathogenic variants was determined by Sanger sequencing and multiplex ligation dependent probe amplification assay. The sequence variant nomenclature followed the Human Genome Variation Society guidelines,9 and the final classification was performed according to the International Society for Gastrointestinal Hereditary Tumors (InSiGHT) criteria for the interpretation of MMR gene variants.10 Only carriers of pathogenic or likely pathogenic variants (ie, class 5 and 4 according to the InSiGHT classification) were included in the study.10 Data were obtained from the computerized surgical database of the two institutions. This database is meticulously updated by trained residents according to the American College of Surgeons' National Quality Improvement Program platform.11 Individual patients’ records were reviewed in order to identify baseline patient and disease characteristics. Some patients included in the study were already included in other studies on Lynch syndrome.8 12 13 During the entire study period, surgical staging was not standardized. Hysterectomy plus bilateral salpingo-oophorectomy was performed in all cases, while lymphadenectomy has been performed for staging purposes in apparent uterine-confined disease since 1988. Cytoreduction was performed in cases of peritoneal involvement. Post-surgical adjuvant strategies included the use of systemic chemotherapy with or without radiotherapy. Over the study period, dedicated gynecologic pathologists performed pathologic reviews of all specimens at both institutions. Stage and architectural grade were reported in accordance with the 2009 International Federation of Obstetrics and Gynecologists (FIGO).4 14 WHO taxonomy was used to classify histological sub-types.14 Non-endometrioid endometrial cancer included (a) serous, (b) clear cell, or (c) undiffentiated cancers.4
Data were summarized using standard descriptive statistics. This is a retrospective comparison between two groups (patients with proven Lynch syndrome versus patients with sporadic cancers). In order to provide a comparison between the groups with the aim to reduce possible biases, we performed a propensity-matched analysis. Propensity-matched comparison attempts to estimate the effect of a treatment by accounting for possible factors (eg, disease characteristics) that predict favorable/unfavorable outcomes, thus reducing possible allocation biases. Propensity-matched comparison aims to reduce biases arising from different covariates. The propensity score was developed using a multivariable logistic regression model. Depth of myometrial invasion, histology, nodal status, and stage of disease were included in the model. Patients with Lynch syndrome were matched 1:2 with patients selected from a cohort of patients with non-endometrioid endometrial cancer considered to be affected by sporadic cancers, using a caliper width ≤0.1 standard deviations of the logit odds of the estimated propensity score. Cases were considered sporadic because of a lack of a family history suggestive of Lynch syndrome and the absence of a deficit of MMR protein expression at immunohistochemical analysis. A detailed description of propensity matching is described elsewhere.14 15 Basic descriptive statistics were used to describe the two populations (Lynch syndrome vs sporadic cancers). Differences in categorical variables were analyzed using the Fisher exact test. OR and 95% confidence intervals (95% CI) were calculated for each comparison. A t-test and Mann–Whitney test were used to compare continuous variables as appropriate. Survival outcomes (disease-free and overall survivals) were estimated using the Kaplan–Meier model. The log-rank test was used to compare the risk of developing recurrence and the risk of death between the two groups over time. P values <0.05 were considered statistically significant. Statistical analysis was performed with GraphPad Prism Version 6.0 (GraphPad Software, San Diego, California, USA) and IBM-Microsoft SPSS Version 20.0 (SPSS Statistics, IBM 2013, Armonk, USA) for Mac.
Overall, 137 patients with endometrial cancer with Lynch syndrome were identified. The study population included 110 (80%) and 27 (20%) patients with endometrioid and non-endometrioid endometrial cancer, respectively. The prevalence of non-endometrioid endometrial cancer was similar in both institutions (16/69 (23.1%) at Fondazione IRCCS Istituto Nazionale Tumori and 11/68 (16.2%) at University of Insubria; p=0.39). Figure 1 shows the study design.
Mean patient age was 49.2 (10.9) years. Genes involved in the Lynch syndrome included MLH1, MSH2, and MSH6 in 43%, 39%, and 18% of cases, respectively. Table 1 shows the baseline characteristics of the study population. For the population with non-endometrioid endometrial cancer (n=27), mean patient age was 48 (11.1) years. Patients with non-endometrioid endometrial cancer had a mutation in MLH1, MSH2, and MSH6 in 28%, 50%, and 22% of cases, respectively. Non-endometrioid endometrial cancer included serous, clear cell, or undiffentiated histology in 16 (60%), 9 (33%), and 2 (7%) patients, respectively. Location of the tumor in the uterus included isthmus, corpus, and fundus in 4 (15%), 21 (78%), and 2 (7%) cases, respectively.
Patients diagnosed with Lynch syndrome who developed non-endometrioid endometrial cancer (n=27) were compared 1:2 with patients with sporadic cancers using a propensity matching algorithm. Overall, 54 (23%) patients were selected from a cohort of 229 patients with sporadic non-endometrioid endometrial cancer. Table 2 shows the baseline characteristics of the two study groups. As the result of propensity matching, the groups are balanced in terms of baseline and histological features. After a median follow-up of 134 months (range 1–295), two of the 27 patients (7.4%) in the study group developed recurrent disease (at 3 and 36 months) and subsequently died of disease (at 7 and 91 months). Patients diagnosed with Lynch syndrome experienced better disease-free survival than patients in the control group (median disease-free survival was not reached and 101 months for patients with Lynch syndrome and controls, respectively; p=0.0016, log-rank (Mantel–Cox) test). Similarly, patients diagnosed with Lynch syndrome experienced better overall survival than patients in the control group (median overall survival was not reached and 105 months for patients with Lynch syndrome and controls, respectively; p=0.0149, log-rank (Mantel–Cox) test). Figure 2 shows survival outcomes. These findings were confirmed using a Cox proportional analysis: after adjusting for age, stage of disease, and adjuvant treatment, diagnosis of Lynch syndrome improved both disease-free (HR 7.86 (95% CI 1.79 to 34.5); p=0.006) and overall survival (HR 5.33 (95% CI 1.18 to 23.9); p=0.029).
This study showed that, although uncommon, non-endometrioid endometrial cancer accounted for approximately 20% of patients with proven Lynch syndrome. Second, patients with Lynch syndrome-related non-endometrioid endometrial cancer experienced better survival outcomes than patients without Lynch syndrome, thus underlying different molecular mechanisms between the two entities. Third, our data showed that, in the context of Lynch syndrome, patients with non-endometrioid endometrial cancer are more likely to have tumor involving the uterine corpus and the fundus, while only 15% of patients had tumor located in the isthmic region.
Historically, Lynch syndrome-related endometrial cancer is characterized by endometrioid histology.16 17 To date, few studies have reported outcomes of patients affected by Lynch syndrome-related endometrial cancer.16 17 Boks et al reported data from the registry of the Netherlands Foundation for Hereditary Tumors, including 50 patients with Lynch syndrome-related endometrial cancer from 46 families (not all harboring a germline mutation). They observed that endometrioid endometrial cancers represented 92% and 88% among Lynch syndrome and sporadic cases, respectively.16 Rossi et al reported data from 49 endometrial cancer patients with Lynch syndrome treated in five centers in France from 1977 to 2013.17 They observed that endometrial cancer in the context of Lynch syndrome is characterized by young age at onset, high prevalence of lower uterine segment involvement, and synchronous ovarian cancers.17 The prognosis of these cancers does not appear different from sporadic tumors. More recently, our study group reported data on 68 patients with Lynch syndrome-related endometrial cancer including 16 (23.5%) with non-endometrioid endometrial cancer. In this latter study our group observed that gene-specific germline pathogenic variants impact the clinical presentation of endometrial cancer in Lynch syndrome. Patients with MLH1 and MSH2 pathogenic variants are at a higher risk of early onset of endometrial cancer and are characterized by more aggressive disease than patients with a MSH6 pathogenic variant.
One of the recent main developments in endometrial cancer research is the molecular/genetic classification of endometrial cancer.18–20 Through molecular and genetic profiling, endometrial cancer can be classified into four categories: microsatellite instability hypermutated, copy-number low, copy-number high, and POLE ultramutated.21 The Proactive Molecular Risk Classifier for Endometrial Cancer (ProMisE) identified four molecular sub-types that are similar to those proposed by the Cancer Genome Atlas (TCGA).18 19 Patients with Lynch syndrome are generally classified in the microsatellite instability group (ie, mismatch repair deficient (MMR-D) according to the ProMisE classification) and are characterized by an overall good prognosis, while patients with non-endometrioid endometrial cancer are generally classified in the copy-number high group (ie, p53 abnormal (p53abn) according to the ProMisE classification) and are characterized by a poor prognosis.18–21 Theoretically, these findings might explain the results achieved in the present study on non-endometrioid endometrial cancer. Although in our study molecular data are missing, we can speculate that patients with Lynch syndrome-related non-endometrioid endometrial cancer and controls represent two different diseases from a biological and molecular point of view. Hence, even in the setting of non-endometrioid endometrial cancer, patients harboring MMR deficiency are characterized by a better prognosis than patients with p53 abnormalities.
The main weaknesses of our study include the inherent biases of the retrospective study design and the long-term study period. Other limitations include the non-standardization of surgical and post-operative adjuvant strategies that could impact on survival outcomes. Additionally, we have to point out that survival curves could be highly influenced by the small number of events in the study population group. The main strength of the study is the large sample size of patients with proven Lynch syndrome (even if we focus only on patients with non-endometrioid endometrial cancer).
The present research investigated disease presentation and outcomes of Lynch syndrome-related non-endometrioid endometrial cancer and found that the occurrence of non-endometrioid endometrial cancer is not rare (about one patient in five) and that patients with Lynch syndrome might experience better oncologic outcomes than sporadic cases. Further molecular/genetic studies are warranted in order to assess prognostic factors in patients with non-endometrioid endometrial cancer.
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Competing interests None declared.
Patient consent for publication Not required.
Provenance and peer review Not commissioned; externally peer reviewed.
Data availability statement All data relevant to the study are included in the article or uploaded as supplementary information.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.