Prevalence and Prognosis of Lynch Syndrome and Sporadic Mismatch Repair Deficiency in Endometrial Cancer

Abstract Background Standard screening of endometrial cancer (EC) for Lynch syndrome (LS) is gaining traction; however, the prognostic impact of an underlying hereditary etiology is unknown. We established the prevalence, prognosis, and subsequent primary cancer incidence of patients with LS-associated EC in relation to sporadic mismatch repair deficient (MMRd)-EC in the large combined Post Operative Radiation Therapy in Endometrial Carcinoma-1, -2, and -3 trial cohort. Methods After MMR-immunohistochemistry, MLH1-promoter methylation testing, and next-generation sequencing, tumors were classified into 3 groups according to the molecular cause of their MMRd-EC. Kaplan-Meier method, log-rank test, and Cox model were used for survival analysis. Competing risk analysis was used to estimate the subsequent cancer probability. All statistical tests were 2-sided. Results Among the 1336 ECs, 410 (30.7%) were MMRd. A total of 380 (92.7%) were fully triaged: 275 (72.4%) were MLH1-hypermethylated MMRd-ECs; 36 (9.5%) LS MMRd-ECs, and 69 (18.2%) MMRd-ECs due to other causes. Limiting screening of EC patients to 60 years or younger or to 70 years or younger would have resulted in missing 18 (50.0%) and 6 (16.7%) LS diagnoses, respectively. Five-year recurrence-free survival was 91.7% (95% confidence interval [CI] = 83.1% to 100%; hazard ratio = 0.45, 95% CI = 0.16 to 1.24, P = .12) for LS, 95.5% (95% CI = 90.7% to 100%; hazard ratio = 0.17, 95% CI = 0.05 to 0.55, P = .003) for “other” vs 78.6% (95% CI = 73.8% to 83.7%) for MLH1-hypermethylated MMRd-EC. The probability of subsequent LS-associated cancer at 10 years was 11.6% (95% CI = 0.0% to 24.7%), 1.5% (95% CI = 0.0% to 4.3%), and 7.0% (95% CI = 3.0% to 10.9%) within the LS, “other,” and MLH1-hypermethylated MMRd-EC groups, respectively. Conclusions The LS prevalence in the Post Operative Radiation Therapy in Endometrial Carcinoma trial population was 2.8% and among MMRd-ECs was 9.5%. Patients with LS-associated ECs showed a trend towards better recurrence-free survival and higher risk for second cancers compared with patients with MLH1-hypermethylated MMRd-EC.

malignancy affecting women with LS (3), and their risk of metachronous cancer is approximately 24% at 10 years (4).
LS-associated cancers arise following MMR deficiency (MMRd) due to the somatic inactivation of the remaining wildtype MMR allele. MMRd leads to the accumulation of mismatches, insertions, and deletions in repeated sequences also known as microsatellite instability (MSI). MMRd is not an exclusive feature of LS; the vast majority (about 70%) of MMRd-ECs present with somatic inactivation of the MLH1 gene via hypermethylation of the promoter region (5,6). Most of the cases that are neither MLH1 hypermethylated nor harbor a MMR germline variant are considered sporadic due to biallelic somatic MMR gene inactivation; few are caused by an undetectable hereditary syndrome (frequently referred to as Lynch-like syndrome) (7)(8)(9). MMRd-ECs are known to have an intermediate prognosis within the molecular classification with a good response to immunotherapy (10)(11)(12)(13). The diagnosis of LS may allow clinicians to tailor treatment and patient information; LS-associated tumors may have a more favorable outcome (14), although there are no previous studies available on the prognostic impact of LS among MMRd-ECs.
Tumor triage by MMR-immunohistochemistry (IHC) and/or MSI analysis in combination with targeted MLH1-methylation testing can identify patients with LS. The Proportion of Endometrial Tumours Associated Lynch Syndrome study showed that IHC-based triage is most accurate, whereas clinical selection based on age and family history were imprecise predictors (15). Overall, an estimated 3% of EC cases are associated with LS (15)(16)(17), which is similar in colorectal cancer (CRC) (18). However, these estimations were mostly based on small trials with methodological heterogeneity, often selecting their test population by age and/or family history, and incomplete testing (16).
Given its relative rarity, the prevalence and prognosis of LS should be investigated in a large population, such as the welldocumented combined cohort of the Post Operative Radiation Therapy in Endometrial Carcinoma (PORTEC)-1, -2, and -3 trials. These randomized controlled trials have had a major impact on guidelines for treatment in ECs (19)(20)(21). Together they included 1336 evaluable patients comprising all risk groups with long and complete follow-up information and collected tumor blocks. The aim of our study was to investigate the prevalence and prognosis of LS-associated EC in relation to MLH1 hypermethylated MMRd-EC. Secondary objectives were to evaluate currently used age criteria for IHC-based tumor triage and the probability of developing a subsequent primary LS-associated cancer.

Study Population
In total, 1336 of 1801 ECs from the PORTEC-1, -2, and -3 clinical trials were eligible for analysis based on availability of formalinfixed paraffin-embedded (FFPE) slides. In the PORTEC-1 trial (1990)(1991)(1992)(1993)(1994)(1995)(1996)(1997) (20). In the international PORTEC-3 trial (2006-2013), 660 EC patients with high-risk features were randomly assigned to receive pelvic radiotherapy or chemoradiotherapy (21). In all trials, patients with a history of invasive cancer (for PORTEC-3 within the last 10 years), except for nonmelanoma skin cancer, were excluded. Full details and results of these trials have been published previously (19)(20)(21). The study protocols were approved by the Dutch Cancer Society and the medical ethics committees at participating centers. All patients provided informed consent for participation in the trial, and for use of their tumor block for subsequent translational research. Clinicopathological data including p53-IHC and POLE-mutation status were obtained from the trial databases. Specific ethics approval was obtained for variant analysis on normal tissue among those suspected of LS. Cases from PORTEC-1 and -2 were analyzed anonymized in view of the long interval since recruitment. Cases from PORTEC-3 who were found to have LS were informed by their own physicians if LS had not been already diagnosed clinically. PORTEC-1 was conducted before time of trial registries. PORTEC-2 is registered with ISRCTN number ISRCTN16228756, and ClinicalTrials.gov number NCT00376844. PORTEC-3 is registered with ISRCTN number ISRCTN14387080, and ClinicalTrials.gov number NCT00411138.

IHC, MSI, Methylation Analysis, and Next-Generation Sequencing (NGS)
Patients were included in the current analysis if they showed loss of expression of at least 1 of the 4 MMR proteins with positive internal control (including subclonal loss defined as abrupt and complete regional loss with intervening stromal positivity) or MSI-high status when MMR-IHC failed. Details on MMR-IHC and MSI testing and scoring were described previously (5,11,12,22). Cases with MMRd phenotype are referred to as MMRd-EC in this study irrespective of POLE mutation status.
MLH1 methylation testing was performed on MLH1-deficient and/or MSI-high tumors as described previously (23). All cases with loss of MLH1 or MSI-high status without MLH1 hypermethylation; loss of MSH2 and/or MSH6; or isolated loss of PMS2 were triaged as potential LS-associated MMRd-EC. DNA isolated from matched normal/tumor FFPE tissues of these cases was amplified using long-range polymerase chain reaction followed by targeted NGS for variants in the exonic regions of MLH1, MSH2, MSH6, PMS2, POLE, and POLD1 using the Ion Proton System or Ion S5 System (Thermo Fisher Scientific) (24,25). Variants were annotated according to the following GenBank

Statistical Analysis
Following complete triage, cases were classified into 3 groups according to the molecular cause of their MMRd-EC: LS, methylated (including cases with MLH1 hypermethylation and subclonal MLH1 loss), and other causes (a mixed group having alternative causes of MMRd; see the Supplementary Methods and Supplementary Figure 1 for full definitions, available online). v 2 Statistics or Fisher's exact test for categorical variables and 1-way analysis of variance or Kruskal-Wallis test for continuous variables were used to compare characteristics.
The sample size ensured sufficient power to detect an LS prevalence of 3.0% with a precision of 0.009 (95% confidence interval [CI] ¼ 2.1% to 3.9%) within the whole population and a prevalence of 12.0% with a precision of 0.03 (95% CI ¼ 9.0% to 15.0%) within the MMRd group (26).
Recurrence-free survival (RFS) was defined as time from random assignment to date of first relapse or death of any cause, whichever occurred first. Overall survival (OS) was defined as time from random assignment to date of death of any cause. Patients without an RFS or OS event were censored at the date of last contact. Five-year survival rates were estimated using the Kaplan-Meier method and compared with log-rank test. Cox proportional hazard models were used to estimate hazard ratios (HRs) over time; for adjusted analysis, age was included as covariate. The proportional hazard assumption was verified using Schoenfeld residuals. A competing-risk model with death as a competing event was used to estimate the cumulative incidence of developing a LS-associated second primary cancer (ie, colorectal, gallbladder, kidney, pancreas, small intestine, stomach, urinary bladder, and ureter cancer) in the different groups. A cause-specific Cox proportional hazard model was used to assess the statistical difference between the estimated probabilities. Time at risk started at random assignment and ended at date of occurrence of the first second cancer, death, or last date of study follow-up. P values less than .05 (2-tailed) were considered statistically significant. Statistical analyses were performed using R version 3.6.1.

MMR Causes and Variant Analysis
Complete triage was accomplished for 380 (92.7%) of the MMRd-ECs ( Figure 1; insufficient material in 27 cases for MLH1 methylation assay and 3 for NGS). Thirty-six path_MMR variant carriers were identified, giving a 2.8% LS prevalence in the overall population and a 9.5% LS prevalence within the MMRd group. There were 18 path_MSH6, 10 path_PMS2, 6 path_MSH2, and 2 path_MLH1 variant carriers. An overview of the LS cases is displayed in Table 2. In total, 275 (72.4%) cases were classified as methylated. The remaining 69 (18.2%) MMRd cases were neither LS nor MLH1 hypermethylated and were therefore classified as "other." LS patients were younger, with a median age of 60 years (interquartile range ¼ 54-67 years) and more often had p53 aberrant staining (20.0%) and serous (13.9%) or clear cell (8.3%) histology compared with the patients with methylated MMRd-EC (Table 3). Limiting screening of EC patients to age 50 years or younger, 60 years or younger, and 70 years or younger would have missed 31 (86.1%), 18 (50.0%), and 6 (16.7%) LS diagnoses, respectively. Figure 2 displays the distribution of the involved MMR proteins; all LS cases identified by the 4-panel approach would also have been identified by a 2-panel approach including only PMS2-and MSH6-IHC. No germline POLE/POLD1 variants affecting function were identified. LS patients with path_MSH6 and path_PMS2 variants were older than those with path_MLH1 and path_MSH2 variants (median age ¼ 63, 62, 50, and 50 years, respectively, P ¼ .01; Supplementary Table 1, available online).

Discussion
After complete IHC-based tumor triage, we found a 2.8% prevalence of LS in 1 of the largest EC trial populations worldwide, comprising all risk groups with long and complete follow-up. The prevalence of LS among patients with MMRd-EC was 9.5%. Patients with LS were relatively young, but restricted testing to    (10)(11)(12). Our survival analysis showed that EC patients with LS tend to have a better RFS than patients with methylated MMRd-EC (HR ¼ 0.45, P ¼ .12), whereas LS had no statistically significant prognostic value for OS after adjustment for age (age-adjusted HR ¼ 0.73, P ¼ .40). The favorable prognosis has been assumed to be induced by the active local immune response (14,27). Comparable survival analysis in CRC has been published. One study showing a better OS for 85 CRC patients with LS compared with 67 sporadic MMRd patients after adjustment for age, stage, and BRAF status (HR ¼ 0.29, 95% CI ¼ 0.09 to 0.95, P ¼ .04) (28). The other study also showing better OS in 37 CRC patients with LS compared with 106 methylated MMRd patients, although the difference was minimal after adjusting for age and stage (29).
The cumulative incidence for developing a second LSassociated cancer at 10 years was 11.6% (95% CI ¼ 0.0% to 24.7%) for patients with LS vs 7.0% (95% CI ¼ 3.0% to 10.9%) for patients with methylated MMRd-EC (HR ¼ 1.90, 95% CI ¼ 0.63 to 5.7, P ¼ .26). Our analysis was underpowered due to the small number of events in the LS group. Nevertheless, the elevated risk strengthens previous reports on subsequent cancers in EC or non-CRC LS patients (15%-24%) (4,30) and is of importance for surveillance strategies.
The 2.8% prevalence of LS-EC is consistent with previous publications in which prevalences of 2.8%-3.2% were reported (15)(16)(17). This prevalence is likely a slight underestimation. Firstly, our NGS panel did not include EPCAM and could not detect large rearrangements. To detect large rearrangement in EPCAM or the MMR genes, Multiplex Ligation-dependent Probe Amplification is most commonly used but performs poorly on FFPE tissue. Secondly, the patient selection in our trial design may have affected the prevalence. Patients younger than 60 years with stage I ECs were excluded from the PORTEC-2 trial. Nevertheless, the total PORTEC population deviates minimally from the general EC population as suggested by the similar age in the Proportion of Endometrial Tumours Associated Lynch Syndrome study, an unselected, prospective, cross-sectional study in the United Kingdom among 500 EC patients (15). Moreover, patients with a history of cancer were excluded from the PORTEC trials. The PORTEC population represents women with EC as their sentinel LS-associated malignancy, which is the case in more than one-half of those women with LS who develop cancer (3). Although this selection has potentially led to a slight underestimation of the prevalence of LS in EC, it does represent the patients in which LS could be detected by IHC-based tumor triage. The recently published meta-analysis by Ryan et al. (16) included mostly small trials with methodological heterogeneity, often selecting their test population by age and/or family history, and incomplete testing; only 1 publication included over 1000 ECs, but germline testing was limited to the minority of the triaged potential LS cases (6). Our study is the first with LS testing in an EC population consisting of more than 1000 women with almost complete MMR-IHC, targeted MLH1 methylation testing, and MMR germline testing, making our estimates more reliable. The path_MSH6 carrier rate of 50.0% among the PORTEC patients with LS is consistent with LS testing results in other unselected EC populations (15,17), but it is remarkably high compared with LS registry data. Only 13% of path_MMR carriers in the clinically selected Prospective Lynch Syndrome Database bear path_MSH6 [1]. As mentioned above, our cohort represents patients with EC as their sentinel cancer likely to induce a lower frequency of path_MLH1 and path_MSH2. Moreover, it must be considered that most of our participants were Dutch, and the path_MSH6 rate of 30% among the Dutch LS registry patients was relatively high compared with the overall Prospective Lynch Syndrome Database (31). Lastly, path_MSH6 families are not identified efficiently by current clinical criteria for LS (32) due to the later age of onset of CRC, incomplete penetrance, and a higher risk and later age of onset of EC (1,(33)(34)(35). The same applies to path_PMS2 carriers with a substantially lower cancer risk (1,2,15,16). Correspondingly, the path_MSH6 and path_PMS2 carriers were older than the path_MLH1 and path_MSH2 carriers in our population.
Triage of incident ECs based on IHC with targeted MLH1 methylation testing, as has been adopted widely for CRC, may be a more effective strategy to identify these LS families than age-and family history-based triage. An upper age screening limit would not be recommended, because limiting screening to EC patients who are aged 70 years or younger would have missed 6 (16.7%) LS diagnoses. We confirmed that a 2-antibody panel including MSH6-and PMS2-IHC, with MSH2-or MLH1-IHC only in case of inconclusive staining, is as sensitive as the full panel to detect LS, so this could be a reliable alternative to improve cost-effectiveness (5,36).
A limitation of our study was the lack of germline LS sequencing on the whole study population. Therefore, sensitivity of the IHC-based triage to identify LS patients could not be assessed. Some patients with LS might have been diagnosed before entering the trial, although many were diagnosed after inclusion and had no prior knowledge of the germline mutation.
The diagnosis of LS in EC is crucial for counseling and cancer surveillance even though these patients might be older than those presenting with CRC (18). Moreover, LS screening in incident ECs will have consequences for the patient's family. Cascade testing of at-risk relatives can identify path_MMR carriers who can benefit from cancer surveillance and riskreducing treatment (37,38). The clinical impact depends on the gene-specific cancer risk and is substantially lower for path_PMS2 carriers (1,2). Finally, LS identification may have consequences by allowing clinicians to better estimate and explain prognosis and to potentially tailor treatment in the upcoming immunotherapy era (14,27,39).
Further research into the causes of the 63 cases with neither MLH1 hypermethylation nor a MMR germline variant is ongoing. It is hypothesized that the majority will be explained by a sporadic origin through biallelic somatic MMR inactivation (15,40). The determination of a sporadic explanation excludes potential undetectable LS (or Lynch-like syndrome) and will avoid a clinical management dilemma in those cases.
In conclusion, LS was identified using MMR-IHC with targeted MLH1 methylation-based triage in 2.8% of 1336 patients with EC from the combined PORTEC-1, -2, and -3 trials, corresponding to 9.5% of the MMRd tumors. LS was mainly caused by germline variants in the MSH6 and PMS2 genes. Patients with LS-associated ECs showed a trend towards better RFS and higher risk for second primary cancers compared with patients with ECs caused by MLH1 hypermethylation. Besides a prognostic impact, screening all incident ECs without an upper age limit to identify LS using tumor-based triage may benefit counseling, affect treatment decisions, and facilitate prevention strategies for current and future patients and their families.

Notes
Role of the funder: The Dutch Cancer society had no role in the design and conduct of the study; collection, managements, analysis, and interpretation of the data; or decision to submit the manuscript for publication.