You are here

Article Text

Article menu
Cancer genetics
Original research
Estimating cancer risk in carriers of Lynch syndrome variants in UK Biobank
  1. Eilidh Fummey1,
  2. Pau Navarro1,2,
  3. John-Paul Plazzer3,
  4. Ian M Frayling4,5,
  5. Sara Knott6,
  6. Albert Tenesa1,2
  1. 1 MRC Human Genetics Unit, Institute of Genetics and Cancer, University of Edinburgh, Edinburgh, UK
  2. 2 The Roslin Institute, University of Edinburgh, Roslin, Midlothian, UK
  3. 3 Colorectal Medicine and Genetics, The Royal Melbourne Hospital, Parkville, Victoria, Australia
  4. 4 The Centre for Familial Intestinal Cancer, St Mark's the National Bowel Hospital and Academic Institute, London, UK
  5. 5 Institute of Cancer & Genetics, Cardiff University, Cardiff, UK
  6. 6 Institute of Ecology and Evolution, University of Edinburgh, Edinburgh, UK
  1. Correspondence to Dr Albert Tenesa; albert.tenesa{at}ed.ac.uk

Background

Lynch syndrome (LS) is an inherited cancer predisposition syndrome caused by genetic variants affecting DNA mismatch repair (MMR) genes MLH1, MSH2, MSH6 and PMS2. Cancer risk in LS is estimated from cohorts of individuals ascertained by individual or family history of cancer, which may upwardly bias estimates.

Methods 830 carriers of pathogenic or likely pathogenic (path_MMR) MMR gene variants classified by InSiGHT were identified in 454 756 UK Biobank (UKB) participants using whole-exome sequence. Nelson-Aalen survival analysis was used to estimate cumulative incidence of colorectal, endometrial and breast cancer (BC).

Results Cumulative incidence of colorectal and endometrial cancer (EC) by age 70 years was elevated in path_MMR carriers compared with non-carriers (colorectal: 11.8% (95% confidence interval (CI): 9.5% to 14.6%) vs 1.7% (95% CI: 1.6% to 1.7%), endometrial: 13.4% (95% CI: 10.2% to 17.6%) vs 1.0% (95% CI: 0.9% to 1.0%)), but the magnitude of this increase differed between genes. Cumulative BC incidence by age 70 years was not elevated in path_MMR carriers compared with non-carriers (8.9% (95% CI: 6.3% to 12.4%) vs 7.5% (95% CI: 7.4% to 7.6%)). Cumulative cancer incidence estimates in UKB were similar to estimates from the Prospective Lynch Syndrome Database for all genes and cancers, except there was no evidence for elevated EC risk in carriers of pathogenic PMS2 variants in UKB.

Conclusion These results support offering incidentally identified carriers of any path_MMR surveillance to manage colorectal cancer risk. Incidentally identified carriers of pathogenic variants in MLH1, MSH2 and MSH6 would also benefit from interventions to reduce EC risk. The results suggest that BC is not an LS-related cancer.

  • Genetic Predisposition to Disease
  • Human Genetics
  • Neoplasms
  • Whole Exome Sequencing
  • Germ-Line Mutation

Data availability statement

Data may be obtained from a third party and are not publicly available. Data used can be obtained through application to UK Biobank: https://www.ukbiobank.ac.uk/enable-your-research/apply-for-access.

https://creativecommons.org/licenses/by/4.0/

This is an open access article distributed in accordance with the Creative Commons Attribution 4.0 Unported (CC BY 4.0) license, which permits others to copy, redistribute, remix, transform and build upon this work for any purpose, provided the original work is properly cited, a link to the licence is given, and indication of whether changes were made. See: https://creativecommons.org/licenses/by/4.0/.

https://doi.org/10.1136/jmg-2023-109791

Statistics from Altmetric.com

    Request Permissions

    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.

    WHAT IS ALREADY KNOWN ON THIS TOPIC

    • Lynch syndrome (LS) is one of the most common genetic tumour syndromes and is associated with elevated colorectal and endometrial cancer (EC) risk that can be managed with increased surveillance and prophylactic surgery.

    • An individual may receive a diagnosis of LS as an incidental finding if sequenced for diagnostic or research purposes.

    WHAT THIS STUDY ADDS

    • This study uses a large population-based cohort, which is not ascertained through individual or family history like disease cohorts, to characterise the cancer risk associated with LS gene variants.

    • Risk estimates for colorectal, endometrial and breast cancer obtained from pathogenic LS gene variant carriers in this population-based cohort were concordant with published estimates from cohorts of individuals with an LS diagnosis undergoing cancer surveillance, with the exception of EC risk in carriers of pathogenic PMS2 variants, which was not elevated relative to non-carriers.

    HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

    • This suggests that individuals who receive an incidental diagnosis of LS would benefit from interventions to manage their colorectal and EC risk; however, this may not be required for EC risk in carriers of pathogenic PMS2 variants.

    Introduction

    Lynch syndrome (LS) is an autosomal dominant cancer predisposition syndrome caused by variants affecting one of four DNA mismatch repair (MMR) genes: MLH1, MSH2, MSH6 or PMS2.1 Individuals with LS generally have increased risk of developing early-onset colorectal cancer (CRC) and some extracolonic cancers, such as endometrial cancer (EC). Understanding the cancer risk associated with likely pathogenic or pathogenic variants affecting MMR gene function (jointly referred to as path_MMR) is important for informing management plans for prevention and early detection of LS-associated cancers.

    The majority of studies that estimate the cancer risk associated with path_MMR use cohorts of individuals with a diagnosis of LS. These cohorts are subject to ascertainment bias because only individuals who have been diagnosed, or have a close family member who was diagnosed, with an LS-related cancer are likely to receive mutational screening. In other hereditary cancer predisposition syndromes, this has led to overestimation of cancer risk.2 3

    If the cancer risk associated with path_MMR is overestimated, this poses a problem for counselling individuals with incidentally identified path_MMR. The American College of Medical Genetics and Genomics and Genomics England recommend reporting incidentally discovered path_MMR to patients even if they are asymptomatic and have no family history suggestive of LS.4 5 Incidental identification of pathogenic variants is becoming more common because of large-scale biobanks that return genetic results to patients, such as the National Institute of Health’s All of Us research programme, and the increasing use of sequencing as a diagnostic tool in clinical genetics.6 7

    Large population-based studies with sequencing data present an opportunity to estimate the penetrance of monogenic disease-causing variants without the bias introduced by selecting individuals based on disease status or family history.8 However, these studies are potentially subject to healthy volunteer bias and survivor bias, depending on the disease and study design, and thus may underestimate penetrance. Hence, such cohorts may provide a lower bound penetrance estimate, whereas classical studies of individuals diagnosed with rare genetic disease provide an upper bound.

    The International Society for Gastrointestinal Hereditary Tumours (InSiGHT) maintains the Colon Cancer Gene Variant Database.9 This database was created to bring together new and existing data on the clinical significance of germline variants for CRC risk. An expert panel reviews this evidence and classifies variants as pathogenic or benign following the International Agency for Research on Cancer’s five-tier classification system.10

    Here, carriers of MMR gene variants listed in the InSiGHT Colon Cancer Gene Variant Database (referred to as the InSiGHT database) are identified in a large population-based study, UK Biobank (UKB), using whole-exome sequence (WES) from 454 756 individuals. Carriers of path_MMR are used to generate cumulative incidence estimates for colorectal and EC, the most commonly diagnosed cancers in individuals with LS,11 and breast cancer (BC), given the mixed evidence surrounding whether BC risk is elevated in LS.12–17 The effect of sex and polygenic risk on CRC risk was investigated. Additionally, cancer risk associated with variants of uncertain significance (VUS) is examined.

    Methods

    Phenotypic data

    UKB consists of ~500 000 individuals living in the UK, recruited between ages 40 and 69 years.18 At recruitment, individuals answered a health and lifestyle questionnaire, provided biological samples and consented to this being linked to their electronic health records, including cancer and death registries.

    The cancer registry data come from two sources: NHS Digital for England and Wales and the National Records of Scotland NHS Central Register. For England and Wales, the data included recorded cancer cases from 1971 onwards, and for Scotland, cases were recorded from 1957 onwards.

    Only cancers that were diagnosed prior to the relevant cancer registry’s censor date were considered. The censor dates (as calculated by UKB19) were 29 February 2020 for England and Wales and 31 January 2021 for Scotland. Death registry records were used to identify the appropriate end of observation for each participant.

    The following International Classification of Diseases (ICD) codes were used to identify relevant cancer diagnoses: ICD-9 153-154 and ICD-10 C18-20 for CRC; ICD-9 182 and ICD-10 C54 for EC; and ICD-9 174 and ICD-10 C50 for BC.

    Family history of CRC was determined from recruitment questionnaire data. Participants were asked ‘has/did your (father/mother/brothers or sisters) ever suffered from any of the following illnesses?’, for which one of the responses was ‘bowel cancer’.

    Identification of carriers of InSiGHT classified variants

    Exome sequencing protocols in UKB have been described previously.20 Genotypes supported by less than 10 reads or with a genotyping quality of less than 20 were set to missing. Sites were removed if more than 10% of genotypes were missing after these filters were applied, or if the site quality was less than 20. Variants located in exons 11–15 of PMS2 were not considered due to this region’s homology with pseudogene PMS2CL.

    Chromosome, base pair position, reference and alternative alleles for 1932 InSiGHT classified variants were obtained from ClinVar21 and used to identify carriers from the UKB exome sequence. Variants are described relative to coding reference sequences on GRCh38.p14 (NC_000007.14): NM_000249.4(MLH1), NM_000251.2(MSH2), NM_000179.2(MSH6) and NM_000535.6(PMS2).

    Estimation of cumulative cancer incidence

    Cumulative cancer incidence was estimated using a modified form of Nelson-Aalen survival analysis as described in Dominguez-Valentin et al.22 Incidence rates were estimated for 5-year intervals from ages 25 to 74, assuming no cancers occurred prior to the age of 25. The incidence rate within an interval is the number of events observed divided by the total number of years individuals were observed for within the interval, multiplied by five (the interval length). The cumulative incidences are obtained from incidence rates using the Nelson-Aalen estimator. Differences between survival curves were tested using a Wilcoxon test where the proportional hazards assumption was not met and a log rank test otherwise.

    Polygenic Risk Score calculation

    Polygenic risk scores (PRS) for CRC were calculated for 694 path_MMR carriers who both self-identified as white British and had genetic principal components indicating European ancestry. The score was calculated as the sum of alleles at 76 variants as in Schmit et al,23 weighted by the effect estimates of the variants. This PRS was chosen as its weights were calculated in samples independent of UKB. Genotypes used in PRS calculation were obtained from the genotyping array data available for UKB participants, described in Bycroft et al.18 Cox regression was used to test the effect of the PRS on CRC risk, with sex included as a covariate. Each gene was analysed individually.

    Results

    465 variants classified by InSiGHT present in UKB

    Among 454 756 UKB participants, 465 variants classified by InSiGHT had at least one alternative allele carrier. Of these 465 variants, 160 were classified as benign or likely benign; 186 were VUS and 119 were pathogenic or likely pathogenic (path_MMR).

    The 119 different path_MMR variants were carried by 830 individuals (~1 in 550 individuals). No individual carried more than one path_MMR and all individuals were heterozygous. Stratified by gene, 431 (51.9%) carried a pathogenic or likely pathogenic variant in PMS2 (path_PMS2), 290 (34.9%) in MSH6 (path_MSH6); 62 in MSH1 (path_MLH1) (7.5%) and 47 in MSH2 (path_MSH2) (5.7%).

    Of the 119 path_MMR represented, 52 (43.7%) were observed in a single UKB participant (figure 1A). All but one path_MMR (PMS2:c.137G>T, n=368) was observed in 50 or fewer participants. An additional 7849 UKB participants carried a VUS (approximately 1 in 60 individuals). The number of carriers per VUS ranged from 1 to 644 individuals (figure 1B).

    Figure 1
    Figure 1

    Histograms showing numbers of carriers in UK Biobank for mismatch repair (MMR) gene variants classified in the InSiGHT database as (A) pathogenic/likely pathogenic (path_MMR) and (B) of uncertain significance (VUS).

    Elevated risk of colorectal and endometrial, but not breast, cancer in path_MMR carriers

    In total, there were 7854 participants with diagnoses of CRC, 84 (1.1%) of which occurred in path_MMR carriers, and 2310 participants with diagnoses of EC, 54 (2.3%) of which occurred in path_MMR carriers. Nelson-Aalen curves for CRC and EC were fit for path_MMR carriers, stratified by gene, and for non-carriers. Leave-one-variant-out curves were generated for variants that made up more than 10% of carriers for a gene. These curves were created using individuals who carried path_MMR for a gene, excluding one particular variant, and were compared with the curve generated using all carriers of path_MMR for that gene to examine whether the shape of a curve for a gene was driven by a single variant. The leave-one-variant-out curves did not differ significantly from the all-variant curve: PMS2:c.137G>T (n=368; Wilcoxon tests: colorectal: W=0.05, p=0.82; endometrial: W=0.68, p=0.41), MSH6:c.3226C>T (n=50; log rank tests: colorectal: χ2=0.30, df=1, p=0.57; endometrial: χ2=0.47, df=1, p=0.49) and MSH6:c.3261dup (n=46; Wilcoxon test, colorectal: W=0.04, p=0.84; log rank test, endometrial: χ2=0.25, df=1, p=0.62). This suggested that the results are not driven by a single variant.

    The CRC Nelson-Aalen curve for path_MMR carriers was significantly different to the curve for non-carriers (log rank test: χ2=398.45, df=1, p=1.20×10−88), with cumulative incidence of CRC by age 70 elevated in path_MMR carriers compared with non-carriers (11.8% (95% CI: 9.5% to 14.6%) vs 1.7% (95% CI: 1.6% to 1.7%)). Stratified by gene, path_MMR carriers had significantly different CRC curves from non-carriers at p<0.01 (log rank tests: path_MLH1: χ2=882.63, df=1, p=5.86×10−194; path_MSH2: χ2=236.81, df=1, p=1.92×10−53; path_MSH6: χ2=182.88, df=1, p=1.14×10−41; path_PMS2: χ2=7.05, df=1, p=0.008). In all 5-year intervals, cumulative incidence of CRC for path_MMR carriers in all genes was equal to or greater than that of non-carriers (figure 2). Among path_MMR carriers, the cumulative incidence of CRC was highest for path_MLH1 carriers and lowest for path_PMS2 carriers in all intervals. The median age of CRC diagnosis was 51.8 for path_MLH1 carriers; 52.3 for path_MSH2; 59.5 for path_MSH6; 67.2 for path_PMS2 and 65.2 for individuals who do not carry path_MMR.

    Figure 2
    Figure 2

    Nelson-Aalen curves of colorectal cancer for carriers of path_MLH1, path_MSH2, path_MSH6, path_PMS2 and individuals who do not carry path_MMR. Shaded areas represent 95% CIs. CRC, colorectal cancer.

    The EC Nelson-Aalen curve for path_MMR carriers was also significantly different to the curve for non-carriers (log rank test: χ2=660.85, df=1, p=9.76×10−146), with cumulative incidence of EC by age 70 elevated in path_MMR carriers compared with non-carriers (13.4% (95% CI: 10.2% to 17.6%) vs 1.0% (95% CI: 0.9% to 1.0%)). When stratified by gene, only path_MLH1, path_MSH2 and path_MSH6 carriers had significantly different EC survival curves from non-carriers at p<0.01 (log rank tests: path_MMR: χ2=660.85, df=1, p=9.76×10−146; path_MLH1: χ2=470.93, df=1, p=2.01×10−104; path_MSH2: χ2=95.6, df=1, p=1.14×10−22; path_MSH6: χ2=935.13, df=1, p=2.27×10−205). In all 5-year intervals, the cumulative EC incidence was greater for carriers of path_MMR variants in these genes than it was for non-carriers (figure 3). The median age of onset for path_MLH1, path_MSH2 and path_MSH6 carriers was 56.4, 47.3 and 59.3, respectively, compared with 62.5 for non-carriers. However, path_PMS2 carriers did not have a significantly different EC survival curve from non-carriers at p<0.01 (log rank test; χ2=3.58, df=1, p=0.059).

    Figure 3
    Figure 3

    Nelson-Aalen curves of endometrial cancer for carriers of path_MLH1, path_MSH2, path_MSH6, path_PMS2 and individuals who do not carry path_MMR. Shaded areas represent 95% CIs

    path_MMR carrier Nelson-Aalen curves for BC were not significantly different to the curve for non-carriers (Wilcoxon tests: path_MMR: W=1.22, p=0.27; path_MLH1: W=0.21, p=0.31; path_MSH2: W=1.02, p=0.31; path_MSH6: W=0.12, p=0.73; path_PMS2: W=0.95, p=0.33) (figure 4). Cumulative incidence of BC by age 70 in carriers of path_MMR was similar to that of non-carriers (8.9% (95% CI: 6.3% to 12.4%) vs 7.5% (95% CI: 7.4% to 7.6%)).

    Figure 4
    Figure 4

    Nelson-Aalen curves of breast cancer for carriers of path_MLH1, path_MSH2, path_MSH6, path_PMS2 and individuals who do not carry path_MMR. Shaded areas represent 95% CIs

    Sex differences in CRC risk observed for path_MLH1 but not other path_MMR

    For each gene, CRC Nelson-Aalen curves of male and female path_MMR carriers were compared with establish whether sex influenced CRC risk (figure 5). The curve for male path_MLH1 carriers was significantly different to the curve for female path_MLH1 carriers at p<0.05 (logrank test: χ2=4.29, df=1, p=0.038), although the 95% CIs of both curves do intersect within all age intervals. Men had a higher cumulative incidence of CRC at age 70 (men: 53.5%, (95% CI: 34.7% to 74.7%); women: 38.2% (95% CI: 23.3% to 58.4%)) and an earlier median age of onset (men: 47.6; women: 55.1) (figure 5A). Differences between curves for men and women were not significant for any other gene (Wilcoxon tests: path_MSH2: W=0.44, p=0.50; path_MSH6: W=0.15, p=0.70; path_PMS2: W=0.12, p=0.73).

    Figure 5
    Figure 5

    Nelson-Aalen curves of colorectal cancer for male and female carriers of (A) path_MLH1, (B) path_MSH2, (C) path_MSH6, and (D) path_PMS2. Shaded areas represent 95% CIs.

    path_MMR carriers report family history of CRC more frequently than non-carriers

    Higher proportions of path_MMR carriers reported a close family history (parent or sibling) of CRC than non-carriers (24.1% vs 10.4%), but the proportion differs significantly between MMR genes (χ2 test: χ2=77.97, df=3, p=8.38×10−17). Over half of path_MLH1 and path_MSH2 carriers report a close family history of CRC (path_MLH1: 56.5%, path_MSH2: 51.1%). Reports of close family history of CRC were less frequent in path_MSH6 carriers (27.2%) and even lower for path_PMS2 carriers (14.4%). However, the proportion of close family history of CRC reported by path_PMS2 carriers was still significantly greater than that of non-carriers (χ2 square test: χ2=6.76, df=1, p=0.0093).

    No evidence that polygenic risk modifies CRC risk in path_MMR carriers

    Cox regression was used to test for the effect of the PRS developed in Schmit et al 23 on CRC risk within path_MMR carriers. White British UKB participants in the top 1% of this PRS had a fourfold increased risk of CRC compared with individuals in the bottom 1%. For path_MSH2, path_MSH6 and path_PMS2, PRS was the sole predictor included in the Cox regression. For path_MLH1, the model included both sex and the PRS, as sex was a significant predictor at p<0.05 (Wald test: b=1.03, z=2.40, p=0.02). There is no evidence that the PRS modifies CRC risk for any gene (Wald test: path_MLH1: n=51, b=6.41, z=1.62, p=0.11; path_MSH2: n=35, b=1.47, z=0.28, p=0.78; path_MSH6: n=242, b=2.00, z=0.64, p=0.52; path_PMS2: n=366, b=3.98, z=0.95, p=0.34).

    Most VUS likely benign

    60 VUS were carried by 30 or more individuals. The proportion of CRC cases observed for carriers was greater than for non-carriers for 33 of these 60 VUS. None of the CRC Nelson-Aalen survival curves for these variants differed significantly from the non-carrier curve a p-value threshold, Bonferroni corrected for the number of VUS, of p<0.0008. One variant had a curve that was significant at a nominal significance threshold of p<0.05: MSH2:c.2400A>G (log rank test: χ2=8.80, df=1, p=0.003) (figure 6). The cumulative incidence of CRC by age 70 in carriers of MSH2:c.2400A>G was 7.6% (95% CI: 3.7% to 15.3%). This was lower than the cumulative incidence of CRC by age 70 for path_MSH2 (31.8% (95% CI: 18.8% to 50.3%)) although it was greater than path_PMS2 (4.3% (95% CI: 2.4% to 7.6%)). However, there was no evidence of increased family history of CRC in carriers of MSH2:c.2400A>G: 12.1% of carriers report a family history of CRC, compared with 10.5% of non-carriers (χ2 test; χ2=0.32, df=1, p=0.57). Moreover, MSH2:c.2400A>G is reported in ClinVar (accession VCV000090965.39) as a synonymous change (NP_000242.1:p.Leu800=) with no predicted impact on splicing and is annotated as likely benign by multiple other sources.

    Figure 6
    Figure 6

    Nelson-Aalen curves of colorectal cancer for carriers (TRUE) of the variant of uncertain significance, MSH2:c.2400A>G, compared to non-carriers (FALSE). Shaded areas represent 95% CIs. CRC, colorectal cancer.

    Discussion

    The analysis of carriers of InSiGHT classified path_MMR in UKB replicates known differences in the CRC and EC risk conferred by pathogenic variants in different MMR genes. EC risk is greater than CRC risk in path_MSH6 carriers.22 24 CRC risk is less and generally later onset in path_MSH6 carriers than path_MLH1 and path_MSH2 carriers.22 24 25 EC and CRC risk is notably less for path_PMS2 carriers compared with carriers of path_MMR in other genes.22

    In UKB, carriers of path_MLH1 and path_MSH2 are least frequently observed, many more path_MSH6 carriers are identified, but the most carriers are observed for path_PMS2. This cannot be explained by greater numbers of InSiGHT classified path_PMS2 variants compared with other genes, as there are more path_MLH1 (535, 45.07%) and path_MSH2 (454, 38.25%) variants than path_MSH6 (167, 14.07%) or path_PMS2 (31, 2.61%). Moreover, some regions of PMS2 were not investigated due to pseudogene homology. The frequency of InSiGHT classified path_MMR in each gene in UKB is the inverse of the penetrance of these variants for CRC. This is expected if more penetrant alleles are subject to stronger purifying selection.26 27

    In cohorts of individuals with an LS diagnosis, the opposite trend is typically observed.24 27 28 An example is Prospective Lynch Syndrome Database (PLSD), which collects data on carriers of path_MMR from the point of genetic diagnosis. At last publication, the database included 8500 carriers: path_MLH1 (36.8%) and path_MSH2 (37.3%) carriers are most common, followed by path_MSH6 (19.4%), and the fewest carriers are observed for path_PMS2 (6.5%).22

    There are a number of potential explanations for this disparity between UKB and LS cohorts. First, there are differences in how path_MMR were identified. The short read sequencing technology used to sequence UKB participants cannot accurately genotype structural variants, such as whole-exon deletions or duplications, and thus these were not considered, although they are frequently identified in LS.29 Additionally, only variants classified by InSiGHT as pathogenic or likely pathogenic were considered as path_MMR to ensure sufficient weight of evidence for pathogenicity.

    Cohort design may influence differences in path_MMR carrier frequency. If path_MLH1 and path_MSH2 carriers are more likely to be affected by cancer than path_MSH6 and path_PMS2 carriers prior to receiving an invitation to participate in UKB (which were sent out to individuals between the ages of 40 and 60), then they may be less likely to enrol.30 31 path_MSH6 and path_PMS2 carriers may be under-represented in LS cohorts because lower penetrance obscures the autosomal dominant inheritance pattern used to identify affected families.

    Contrastingly, the distribution of carriers among genes in UKB is similar to the distribution of biallelic path_MMR carriers enrolled in the International Replication Repair Deficiency Consortium (PMS2: 61%, MSH6: 29%, MSH2: 5.5%, MLH1: 4.5%).32 This consortium enrols individuals with suspected constitutional mismatch repair deficiency syndrome (CMMRD): a highly penetrant cancer predisposition syndrome characterised by multiple childhood-onset cancers.33 If path_PMS2 and path_MSH6 are more common in the population, as the distribution in UKB suggests, this may explain why they are more often observed as a cause of CMMRD.

    Despite differences in ascertainment between UKB and LS cohorts, CRC cumulative incidence estimates from UKB are similar to those in the literature.15 22 25 Using PLSD as an example, the cumulative CRC incidence by age 70 for male carriers of path_MLH1 is 53.5% (95% CI: 34.7% to 74.7%) in UKB and 52.8% (95% CI: 46.1% to 59.8%) in PLSD; for path_MSH2, 31.0% (95% CI: 14.9% to 57.32%) vs 51.0% (95% CI: 42.8% to 59.7%); for path_MSH6, 14.6% (95% CI: 9.1% to 23.0%) vs 13.5% (95% CI: 7.1% to 24.8%) and for path_PMS2 3.4% (95% CI: 1.3% to 9.0%) vs 10.5% (95% CI: 2.7% to 36.0%) (online supplemental figure 1). As in UKB, female carriers of path_MLH1 in PLSD have a lower cumulative CRC incidence by age 70 than male carriers: 42.1% (95% CI: 36.2% to 48.6%) in UKB vs 38.2% (95% CI: 23.25% to 58.39%) in PLSD. Values are also similar for female carriers of path_MMR in the other three genes: for path_MSH2, 32.8% (95% CI: 15.5% to 60.7%) in UKB vs 39.8% (95% CI: 33.5% to 46.7%) in PLSD; for path_MSH6 11.0% (95% CI: 6.6% to 18.1%) vs 17.3% (95% CI: 11.2% to 26.3%); and for path_PMS2 5.0% (95% CI: 2.5% to 9.9%) vs 8.5% (95% CI: 2.1% to 31.5%) (online supplemental figure 2). However, it is unclear if path_MMR carriers in UKB have been diagnosed with LS and are therefore receiving the CRC surveillance offered to individuals in PLSD.

    Supplemental material

    The cumulative EC incidence by age 70 for path_MLH1 is 43.7% (95% CI: 24.5% to 69.1%) in UKB vs 35.8% (95% CI: 29.9% to 42.5%) in PLSD; for path_MSH2 23.7% (95% CI: 8.95% to 54.1%) vs 42.1% (95% CI: 35.0% to 50.1%); for path_MSH6 32.4% (95% CI: 23.5% to 43.5%) vs 41.4% (95% CI: 32.3% to 52.0%); and for path_PMS2 2.6% (95% CI: 1.07% to 6.32%) vs 12.7% (95% CI: 5.5% to 27.9%) (figure 3). The only case where the 95% CI for UKB does not overlap the PLSD estimate and vice versa is for path_PMS2. In PLSD, carriers of path_PMS2 have elevated EC risk but this trend is not observed in UKB. Note that 85% of female path_PMS2 carriers in UKB carry c.137G>T, which has an allele frequency of 8.09×10−4 in UKB and 1.73×10−4 in gnomAD V.2.1.1.34 It is possible that c.137G>T is less penetrant than other path_PMS2, although the analysis of path_PMS2 carriers excluding this variant also failed to find elevated EC risk. However, as there was a low number of individuals carrying path_PMS2 that were not c.137G>T (n=39), there may have been insufficient power to detect an effect if penetrance is low. Further study is required to clarify the relationship between path_PMS2 and EC risk.

    Importantly, the similarity between UKB and PLSD does not suggest the existence of large subgroups of path_MMR carriers with low cancer risk that have been excluded from previous LS cancer risk estimates due to ascertainment bias. This is not the case in all monogenic diseases. One paper reported that the penetrance of diabetes in pathogenic HNF1A/HNF4A variant carriers was much lower in carriers identified through unselected cohorts (32%) vs cohorts of individuals with a molecular diagnosis (98%).35 However, there was no difference in diabetes penetrance between different ascertainment contexts for carriers of pathogenic GCK variants. The data presented here for LS suggest that variants in all four of the MMR genes discussed are more similar to GCK than HNF1A/HNF4A: the method of identification of variant carriers does not change their penetrance estimates to such an extent that it would warrant different clinical management of incidentally identified carriers. The only exception is that path_PMS2 carriers are observed to have lower EC risk in UKB than previously reported.

    Elevated incidence of BC has been reported in individuals carrying path_MSH6 and path_PMS2.12–14 However, this is not consistently observed across all studies.15–17 Elevated BC incidence is more commonly observed in studies where path_MMR carriers are identified in cohorts of individuals referred for non-specific hereditary cancer screening. It has been argued that because individuals with path_MMR are typically identified through screening of CRC patients, families with elevated risk of BC may be excluded, leading to inconsistent findings.13

    If this were the case, elevated BC incidence would be expected in path_MSH6 and path_PMS2 carriers in UKB. Rather, there is no evidence of increased BC risk in path_MMR carriers of any kind in UKB. Previous links between path_MMR and BC may be caused by higher rates of mammographic screening in individuals with a family history of LS-associated cancers, such as ovarian, leading to overdiagnosis.15 Additionally, due to their population frequencies, path_MSH6 and path_PMS2 may often be incidentally identified in individuals undergoing multigene hereditary cancer screening. The population undergoing this screening has a higher proportion of individuals with a BC diagnosis than the general population, so elevated BC risk in path_MMR carriers identified in this way could be attributed to ascertainment bias.16

    One potential reason why one path_MMR carrier may develop CRC while another does not is additional genetic risk conferred by common variants. Previously, a 96 variant PRS was reported to be associated with CRC in UKB path_MMR carriers36 37; however, UKB samples made up 20% of the sample used to calculate the weights for this PRS, so there is a risk of overfitting.38 Studies that have tested the association between the 95 variant PRS and CRC in individuals with LS outside of UKB have failed to observe a significant relationship.39 40 Here, a 76 variant PRS for CRC was not associated with CRC risk in UKB path_MMR carriers. Although the power of this analysis is limited by small sample size, it appears likely that the relationship between this PRS and CRC risk is not strong enough for useful risk stratification in LS.

    Around 1 in 60 UKB participants carries a variant classified as of uncertain significance in the InSiGHT database. VUS are commonly encountered when performing mutational testing of MMR genes, introducing uncertainty around diagnosis. For most VUS observed in UKB, there was little evidence of pathogenicity. Most VUS behaving like benign variants in UKB is consistent with the observation that VUS are more likely to be reclassified as benign or likely benign than upgraded to pathogenic or likely pathogenic.41–43 Cumulative incidences of CRC for VUS are provided in . This evidence can be incorporated into future discussions about reclassification of VUS in MMR genes.

    In summary, these results do not advocate for the reclassification of any VUS as pathogenic. However, variants classified as pathogenic or likely pathogenic by InSiGHT are associated with increased risk of CRC in UKB, similar to published estimates. The same observation was made for EC risk in carriers of path_MLH1, path_MSH2 and path_MSH6. This suggests that individuals incidentally found to carry these variants would benefit from interventions to manage these risks. However, the absence of a heightened risk of EC in carriers of path_PMS2 necessitates further investigation into the appropriateness of interventions to manage EC risk in this subgroup. Lastly, failure to observe elevated BC risk in path_MMR carriers in UKB supports the stance that BC should not be considered an LS-related cancer.

    Data availability statement

    Data may be obtained from a third party and are not publicly available. Data used can be obtained through application to UK Biobank: https://www.ukbiobank.ac.uk/enable-your-research/apply-for-access.

    Ethics statements

    Patient consent for publication

    Ethics approval

    This study involves human participants. National Research Ethics Service Committee North West-Haydock gave ethical approval for the UK Biobank project (REC reference: 21/NW/0157). Participants gave informed consent to participate in the study before taking part.

    Acknowledgments

    This research has been conducted using the UK Biobank Resource under Application Number 44986. This work uses data provided by patients and collected by the NHS as part of their care and support.

    References

    1. ↵
      2. Peltomäki P
      . Update on Lynch syndrome Genomics. Fam Cancer 2016;15:385–93. doi:10.1007/s10689-016-9882-8
      OpenUrlCrossRef
    2. ↵
      2. de Andrade KC ,
      3. Mirabello L ,
      4. Stewart DR , et al
      . Higher-than-expected population prevalence of potentially pathogenic Germline Tp53 variants in individuals Unselected for cancer history. Hum Mutat 2017;38:1723–30. doi:10.1002/humu.23320
      OpenUrlCrossRefPubMed
    3. ↵
      2. Antoniou A ,
      3. Pharoah PDP ,
      4. Narod S , et al
      . Average risks of breast and ovarian cancer associated with Brca1 or Brca2 mutations detected in case series Unselected for family history: A combined analysis of 22 studies. Am J Hum Genet 2003;72:1117–30. doi:10.1086/375033
      OpenUrlCrossRefPubMedWeb of Science
    4. ↵
      2. Miller DT ,
      3. Lee K ,
      4. Chung WK , et al
      . ACMG SF V3.0 list for reporting of secondary findings in clinical Exome and genome sequencing: a policy statement of the American college of medical Genetics and Genomics (ACMG). Genetics in Medicine 2021;23:1381–90. doi:10.1038/s41436-021-01172-3
      OpenUrlPubMed
    5. ↵
      1. Additional findings
      . Genomics England, Available: https://www.genomicsengland.co.uk/initiatives/100000-genomes-project/additional-findings [Accessed 22 Nov 2022].
    6. ↵
      2. Harrison SM ,
      3. Austin-Tse CA ,
      4. Kim S , et al
      . Harmonizing variant classification for return of results in the all of us research program. Hum Mutat 2022;43:1114–21. doi:10.1002/humu.24317
      OpenUrlCrossRef
    7. ↵
      2. Turro E ,
      3. Astle WJ ,
      4. Megy K , et al
      . Whole-genome sequencing of patients with rare diseases in a national health system. Nature 2020;583:96–102:96. doi:10.1038/s41586-020-2434-2
      OpenUrlCrossRefPubMed
    8. ↵
      2. Wright CF ,
      3. West B ,
      4. Tuke M , et al
      . Assessing the Pathogenicity, Penetrance, and Expressivity of putative disease-causing variants in a population setting. Am J Hum Genet 2019;104:275–86. doi:10.1016/j.ajhg.2018.12.015
      OpenUrlCrossRefPubMed
    9. ↵
      2. Thompson BA ,
      3. Spurdle AB ,
      4. Plazzer J-P , et al
      . Application of a 5-tiered scheme for standardized classification of 2,360 unique mismatch repair gene variants in the insight locus-specific database. Nat Genet 2014;46:107–15. doi:10.1038/ng.2854
      OpenUrlCrossRefPubMed
    10. ↵
      2. Plon SE ,
      3. Eccles DM ,
      4. Easton D , et al
      . Sequence variant classification and reporting: recommendations for improving the interpretation of cancer susceptibility genetic test results. Hum Mutat 2008;29:1282–91. doi:10.1002/humu.20880
      OpenUrlCrossRefPubMedWeb of Science
    11. ↵
      2. Dominguez-Valentin M ,
      3. Sampson JR ,
      4. Seppälä TT , et al
      . Cancer risks by gene, age, and gender in 6350 carriers of pathogenic mismatch repair variants: findings from the prospective Lynch syndrome database. Genetics in Medicine 2020;22:15–25. doi:10.1038/s41436-019-0596-9
      OpenUrlCrossRefPubMed
    12. ↵
      2. Espenschied CR ,
      3. LaDuca H ,
      4. Li S , et al
      . Multigene panel testing provides a new perspective on Lynch syndrome. J Clin Oncol 2017;35:2568–75. doi:10.1200/JCO.2016.71.9260
      OpenUrlCrossRefPubMed
    13. ↵
      2. Roberts ME ,
      3. Jackson SA ,
      4. Susswein LR , et al
      . Msh6 and Pms2 germ-line pathogenic variants implicated in Lynch syndrome are associated with breast cancer. Genetics in Medicine 2018;20:1167–74. doi:10.1038/gim.2017.254
      OpenUrlCrossRef
    14. ↵
      2. Sheehan M ,
      3. Heald B ,
      4. Yanda C , et al
      . Investigating the link between Lynch syndrome and breast cancer. Eur J Breast Health 2020;16:106–9. doi:10.5152/ejbh.2020.5198
      OpenUrl
    15. ↵
      2. Møller P
      . The prospective Lynch syndrome database reports enable evidence-based personal precision health care. Hered Cancer Clin Pract 2020;18:6. doi:10.1186/s13053-020-0138-0
    16. ↵
      2. Stoll J ,
      3. Rosenthal E ,
      4. Cummings S , et al
      . No evidence of increased risk of breast cancer in women with Lynch syndrome identified by Multigene panel testing. JCO Precision Oncology 2020;4:51–60. doi:10.1200/PO.19.00271
      OpenUrl
    17. ↵
      2. Win AK ,
      3. Lindor NM ,
      4. Jenkins MA
      . Risk of breast cancer in Lynch syndrome: A systematic review. Breast Cancer Res 2013;15:R27. doi:10.1186/bcr3405
    18. ↵
      2. Bycroft C ,
      3. Freeman C ,
      4. Petkova D , et al
      . The UK Biobank resource with deep Phenotyping and Genomic data. Nature 2018;562:203–9. doi:10.1038/s41586-018-0579-z
      OpenUrlCrossRefPubMed
    19. ↵
      1. UK Biobank
      . Data providers and dates of data availability, 2023. Available: https://biobank.ndph.ox.ac.uk/ukb/exinfo.cgi?src=Data_providers_and_dates [Accessed 18 Feb 2024].
    20. ↵
      2. Van Hout CV ,
      3. Tachmazidou I ,
      4. Backman JD , et al
      . Exome sequencing and characterization of 49,960 individuals in the UK Biobank. Nature 2020;586:749–56. doi:10.1038/s41586-020-2853-0
      OpenUrlCrossRefPubMed
    21. ↵
      2. Landrum MJ ,
      3. Lee JM ,
      4. Riley GR , et al
      . Clinvar: public Archive of relationships among sequence variation and human phenotype. Nucl Acids Res 2014;42:D980–5. doi:10.1093/nar/gkt1113
      OpenUrlCrossRefPubMedWeb of Science
    22. ↵
      2. Dominguez-Valentin M ,
      3. Haupt S ,
      4. Seppälä TT , et al
      . Mortality by age, gene and gender in carriers of pathogenic mismatch repair gene variants receiving surveillance for early cancer diagnosis and treatment: a report from the prospective Lynch syndrome database. EClinicalMedicine 2023;58:101909. doi:10.1016/j.eclinm.2023.101909
    23. ↵
      2. Schmit SL ,
      3. Edlund CK ,
      4. Schumacher FR , et al
      . Novel common genetic susceptibility Loci for colorectal cancer. J Natl Cancer Inst 2019;111:146–57. doi:10.1093/jnci/djy099
      OpenUrlPubMed
    24. ↵
      2. Bonadona V ,
      3. Bonaïti B ,
      4. Olschwang S , et al
      . Cancer risks associated with Germline mutations in Mlh1, Msh2, and Msh6 genes in Lynch syndrome. JAMA 2011;305:2304–10. doi:10.1001/jama.2011.743
      OpenUrlCrossRefPubMedWeb of Science
    25. ↵
      2. Wang C ,
      3. Wang Y ,
      4. Hughes KS , et al
      . Penetrance of colorectal cancer among mismatch repair gene Mutation carriers: A meta-analysis. JNCI Cancer Spectrum 2020;4. doi:10.1093/jncics/pkaa027
    26. ↵
      2. Charlesworth B ,
      3. Charlesworth D
      . Elements of Evolutionary Genetics. Colorado, USA: Roberts and Company, 2010.
    27. ↵
      2. Win AK ,
      3. Jenkins MA ,
      4. Dowty JG , et al
      . Prevalence and Penetrance of major genes and Polygenes for colorectal cancer. Cancer Epidemiol Biomarkers Prev 2017;26:404–12. doi:10.1158/1055-9965.EPI-16-0693
      OpenUrlAbstract/FREE Full Text
    28. ↵
      2. Moreira L ,
      3. Balaguer F ,
      4. Lindor N , et al
      . Identification of Lynch syndrome among patients with colorectal cancer. JAMA 2012;308:1555. doi:10.1001/jama.2012.13088
      OpenUrlCrossRefPubMedWeb of Science
    29. ↵
      2. Gylling A ,
      3. Ridanpää M ,
      4. Vierimaa O , et al
      . Large Genomic Rearrangements and Germline Epimutations in Lynch syndrome. Int J Cancer 2009;124:2333–40. doi:10.1002/ijc.24230
      OpenUrlCrossRefPubMed
    30. ↵
      2. Struijk EA ,
      3. May AM ,
      4. Beulens JWJ , et al
      . Mortality and cancer incidence in the EPIC-NL cohort: impact of the healthy volunteer effect. Eur J Public Health 2015;25:144–9. doi:10.1093/eurpub/cku045
      OpenUrlCrossRefPubMed
    31. ↵
      2. Fry A ,
      3. Littlejohns TJ ,
      4. Sudlow C , et al
      . Comparison of Sociodemographic and health-related characteristics of UK Biobank participants with those of the general population. Am J Epidemiol 2017;186:1026–34. doi:10.1093/aje/kwx246
      OpenUrlCrossRefPubMed
    32. ↵
      2. Durno C ,
      3. Ercan AB ,
      4. Bianchi V , et al
      . Survival benefit for individuals with constitutional mismatch repair deficiency undergoing surveillance. JCO 2021;39:2779–90. doi:10.1200/JCO.20.02636
      OpenUrlCrossRefPubMed
    33. ↵
      2. Bakry D ,
      3. Aronson M ,
      4. Durno C , et al
      . Genetic and clinical determinants of constitutional mismatch repair deficiency syndrome: report from the constitutional mismatch repair deficiency consortium. Eur J Cancer 2014;50:987–96. doi:10.1016/j.ejca.2013.12.005
      OpenUrlCrossRefPubMed
    34. ↵
      2. Karczewski KJ ,
      3. Francioli LC ,
      4. Tiao G , et al
      . The mutational constraint spectrum quantified from variation in 141,456 humans. Nature 2020;581:434–43. doi:10.1038/s41586-020-2308-7
      OpenUrlCrossRefPubMed
    35. ↵
      2. Mirshahi UL ,
      3. Colclough K ,
      4. Wright CF , et al
      . Reduced Penetrance of MODY-associated HNF1A/HNF4A variants but not GCK variants in clinically Unselected cohorts. Am J Hum Genet 2022;109:2018–28. doi:10.1016/j.ajhg.2022.09.014
      OpenUrlCrossRefPubMed
    36. ↵
      2. Hassanin E ,
      3. Spier I ,
      4. Bobbili DR , et al
      . Clinically relevant combined effect of Polygenic background, rare pathogenic Germline variants, and family history on colorectal cancer incidence. BMC Med Genomics 2023;16:1–12. doi:10.1186/s12920-023-01469-z
      OpenUrl
    37. ↵
      2. Fahed AC ,
      3. Wang M ,
      4. Homburger JR , et al
      . Polygenic background modifies Penetrance of Monogenic variants for tier 1 Genomic conditions. Nat Commun 2020;11:3635. doi:10.1038/s41467-020-17374-3
    38. ↵
      2. Huyghe JR ,
      3. Bien SA ,
      4. Harrison TA , et al
      . Discovery of common and rare genetic risk variants for colorectal cancer. Nat Genet 2019;51:76–87. doi:10.1038/s41588-018-0286-6
      OpenUrlCrossRefPubMed
    39. ↵
      2. Jenkins MA ,
      3. Buchanan DD ,
      4. Lai J , et al
      . Assessment of a Polygenic risk score for colorectal cancer to predict risk of Lynch syndrome colorectal cancer. JNCI Cancer Spectrum 2021;5. doi:10.1093/jncics/pkab022
    40. ↵
      2. Dueñas N ,
      3. Klinkhammer H ,
      4. Bonifaci N , et al
      . Ability of a Polygenic risk score to refine colorectal cancer risk in Lynch syndrome. J Med Genet 2023;60:1044–51. doi:10.1136/jmg-2023-109344
      OpenUrlAbstract/FREE Full Text
    41. ↵
      2. Shah N ,
      3. Hou Y-CC ,
      4. Yu H-C , et al
      . Identification of misclassified Clinvar variants via disease population prevalence. Am J Hum Genet 2018;102:609–19. doi:10.1016/j.ajhg.2018.02.019
      OpenUrlCrossRefPubMed
    42. ↵
      2. Chiang J ,
      3. Chia TH ,
      4. Yuen J , et al
      . Impact of variant reclassification in cancer predisposition genes on clinical care. JCO Precision Oncology 2021;5:577–84. doi:10.1200/PO.20.00399
      OpenUrl
    43. ↵
      2. Mersch J ,
      3. Brown N ,
      4. Pirzadeh-Miller S , et al
      . Prevalence of variant reclassification following hereditary cancer genetic testing. JAMA 2018;320:1266–74. doi:10.1001/jama.2018.13152
      OpenUrlCrossRefPubMed

    Supplementary materials

    • Supplementary Data

      This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

    Footnotes

    • X @Doctor_IMF

    • Correction notice The article has been corrected since it was published online. Statements in the Acknowledgment section have been added for clarity. To comply with UK Biobank's guidance, supplementary file 1 has been removed because all tables included values less than 5. The conclusions of the article have not been impacted by this change.

    • Contributors AT conceived the idea of the study. Analyses were planned by EF, PN, SK and AT. EF carried out the analysis. All authors contributed to the interpretation of analysis results. EF wrote the manuscript with input from all authors. AT acts as a guarantor.

    • Funding This work was funded under the awards from the Biotechnology and Biological Sciences Research Council (BB/S508032/1, BBS/E/D/10002070, BBS/E/D/30002275, BBS/E/RL/230001A), Health Data Research UK (HDR-9004) and the Medical Research Council (MC_PC_U127561128, MC_PC_U127592696, MC_UU_00007/10, MR/P015514/1).

    • Competing interests None declared.

    • Provenance and peer review Not commissioned; externally peer reviewed.

    • Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.

    Linked Articles