Clinical Description
Individuals with Lynch syndrome are at increased risk for colorectal cancer (CRC) and other cancers including cancers of the endometrium, ovary, stomach, small intestine, hepatobiliary tract, upper urinary tract, brain, and skin (Table 3).
Table 3.
Cancer Risks by Gene in Individuals with Lynch Syndrome Age ≤70 Years Compared to the General Population
View in own window
Cancer Type | General Population Risk | MLH1 and MSH2 | MSH6 | PMS2 |
---|
Risk | Mean Age of Onset | Risk | Risk |
---|
Colorectal | 5.5% | M: 27%-74% F: 22%-53% | 27-46 yrs | M: 22% F: 10% | M: 20% F: 15% |
Endometrial | 2.7% | 14%-54% | 48-62 yrs | 16%-26% | 15% |
Gastric | <1% | 0.2%-13% | 49-55 yrs | M: 6% F: 22% | 6% |
Ovarian | 1.6% | 4%-20% | 43-45 yrs |
Small bowel | <1% | 4%-12% | 49 yrs |
Hepatobiliary tract | <1% | 0.2%-4% | 54-57 |
Urinary tract | <1% | 0.2%-25% | 52-60 yrs |
Brain | <1% | 1%-4% | ~50 yrs |
Sebaceous neoplasms | <1% | 1%-9% | Not reported | Unknown | Unknown |
Pancreas | 1.5% | 0.4%-4% | 63-65 yrs | Unknown | Unknown |
Prostate | 16.2% | 9%-30% | 59-60 yrs | Unknown | Unknown |
Breast | 12.4% | 5%-18% | 52 yrs | Unknown | Unknown |
Colorectal cancer. The risk of colorectal cancer (CRC) associated with MLH1 and MHS2 pathogenic variants is significantly higher than the risk associated with MSH6 or PMS2 pathogenic variants. The mean ages at onset for CRC in individuals with MSH6 and PMS2 pathogenic variants are also older than for onset for CRC associated with MLH1 and MSH2 pathogenic variants, 54-63 years and 47-66 years respectively. However, up to 8% of CRCs in PMS2 heterozygotes occur before age 30. Therefore, individuals with MLH6 or PMS2 pathogenic variants should follow the same CRC screening recommendations as MLH1 and MSH2 heterozygotes. A 2009 study of a Finnish cohort with high compliance with screening found no increase in mortality for individuals with Lynch syndrome compared to relatives without the Lynch syndrome-related pathogenic variant, indicating that annual colonoscopy is effective for the prevention and detection of CRC [Järvinen et al 2009].
Data on cancer risks for those with an EPCAM deletion is still limited, but Kempers et al [2011] reported on findings from 194 individuals with an EPCAM deletion. From this cohort they estimated a 75% (95% CI 65-86) cumulative incidence of CRC by age 70 years. The risk for CRC did not differ between those who had only a 3' deletion of EPCAM and those individuals with a deletion that encompassed EPCAM and MSH2. However, only those with an EPCAM deletion that includes MSH2 are at an increased risk for extracolonic cancers [Kempers et al 2011, Tutlewska et al 2013].
Tumors that are MSI-H, either due to sporadic causes or an underlying germline MMR pathogenic variant, tend to have a better prognosis than MSS tumors [Kawakami et al 2015].
Endometrial cancer. The risk for women with EPCAM deletions that encompass MSH2 is similar to that for individuals with MSH2 pathogenic variants.
Among females with Lynch syndrome who develop both CRC and endometrial cancer, approximately 50% present first with endometrial cancer [Lu et al 2005]. The risk for subsequent endometrial cancer to females with Lynch syndrome presenting first with CRC has been estimated at 26% within ten years of the initial CRC diagnosis [Obermair et al 2010].
Overall, a survival advantage similar to that in Lynch syndrome-related CRC has been reported in Lynch syndrome-related endometrial cancers [Maxwell et al 2001].
Gastric cancer. Intestinal-type adenocarcinoma, the most commonly reported pathology of Lynch syndrome-related gastric cancers [Aarnio et al 1997], differs histologically from the diffuse gastric cancer that is most commonly seen in hereditary diffuse gastric cancer caused by pathogenic variants in CDH1 [Guilford et al 1999]. However, Capelle et al [2010] reported that up to 20% of Lynch syndrome-related gastric cancers may be the diffuse type. The risk for gastric cancer is higher in individuals with Lynch syndrome who reside in countries with a high incidence of H pylori infection [Park et al 2000].
Ovarian cancer risk to females with a germline MLH1 or MSH2 pathogenic variant has been found to be 4%-20%. The mean age of diagnosis of Lynch syndrome-associated ovarian cancer has been reported at between age 43 and 45 years, although diagnosis at very young ages has been reported. Approximately 30% of Lynch syndrome-associated ovarian cancers are diagnosed before age 35 years [Watson et al 2008].
The distribution of pathology types is similar to that seen in sporadic ovarian cancers. Borderline ovarian tumors do not appear to be associated with Lynch syndrome [Watson et al 2001].
Small bowel cancer. The duodenum and jejunum are the most common sites for small bowel cancers, with approximately 50% in reach of upper endoscopy [Schulmann et al 2005]. The majority of small bowel cancers are adenocarcinomas [Rodriguez-Bigas et al 1998, Schulmann et al 2005].
Urinary tract cancers. The urinary tract cancers most commonly associated with Lynch syndrome are transitional carcinomas of the ureter and renal pelvis.
Bladder cancer risk is also likely increased in individuals with Lynch syndrome. A study of Dutch individuals with Lynch syndrome demonstrated a relative risk for bladder cancer of 4.4 for males and 2.2 for females; the majority of tumor tissue available for testing was MSI-H and/or had loss of protein expression by IHC [van der Post et al 2010].
Individuals with Lynch syndrome and a prior diagnosis of CRC were also at increased risk for subsequent bladder cancer (7.22, 95% CI=4.08-10.99) and other urinary tract cancers (kidney, renal pelvis, and ureter) (12.54, 95% CI 7.97-17.94) [Win et al 2013b].
Risk estimates for urinary tract cancers vary significantly based on gender and the gene involved.
Brain tumors. The most common type of central nervous system tumor is glioblastoma [Hamilton et al 1995, Wimmer & Etzler 2008]. The brain tumors associated with pathogenic variants in an MMR gene are typically MSI-H [Hamilton et al 1995, Suzui et al 1998].
Sebaceous neoplasms described in individuals with Lynch syndrome include: sebaceous adenomas, sebaceous epitheliomas, sebaceous carcinomas, and keratoacanthomas. Sebaceous neoplasms associated with Lynch syndrome are typically MSI-H [Entius et al 2000, Machin et al 2002]. Data on the frequency of sebaceous neoplasms in individuals with Lynch syndrome are limited. Studies have found that between 1% and 9% of individuals with a germline pathogenic variant in an MMR gene have a sebaceous neoplasm [Ponti et al 2006, South et al 2008].
Other Cancers
Other Lynch syndrome-related cancers that have characteristic features have been reported.
Pancreatic cancer. A study by Kastrinos et al [2009] based on reported family history found an 8.6-fold increased risk up to age 70 years for pancreatic cancer [Kastrinos et al 2009]. A prospective study that followed 446 individuals with an MMR pathogenic variant and 1,029 relatives for a median of five years found an increased risk for pancreatic cancer (SIR, 10.68; 95% CI 2.68-47.70), and no increased risk for individuals without a pathogenic variant [Win et al 2012]. However, other studies have not demonstrated an increased risk [Barrow et al 2009]. Lynch syndrome has been found to be a rare cause of familial pancreatic cancer [Gargiulo et al 2009].
Prostate cancer. Several studies have demonstrated an association with prostate cancer, with the increase in risk ranging from two- to fivefold [Raymond et al 2013b, Haraldsdottir et al 2014, Ryan et al 2014]. Raymond et al [2013b] found that the risk for prostate cancer was increased for males with an MMR pathogenic variant prior to age 60 [Raymond et al 2013b], whereas an analysis by Haraldsdottir et al [2014] did not find earlier age of onset or a more aggressive phenotype in the prostate cancers occurring in individuals with an MMR pathogenic variant. Pritchard et al [2016] identified a pathogenic variant in an MMR gene in four (0.5%) of 692 men with metastatic prostate cancer.
Breast cancer. The relationship between breast cancer and Lynch syndrome is unresolved. A systematic review evaluated 21 studies; 13 did not demonstrate an increased risk for breast cancer in individuals with Lynch syndrome and eight showed an increased risk [Win et al 2013b]. To date, breast cancer risk has only been evaluated in one prospective study. Individuals with an MMR pathogenic variant were found to have a standard incidence ratio for breast cancer of 3.95 (95% CI 1.59-8.13), and the median age of breast cancer diagnosis was 56 years. Tumor tissue IHC in 51% of breast cancers in individuals with an MMR pathogenic variant demonstrated loss of expression for the MMR gene with the germline pathogenic variant. Due to the high frequency of breast cancer in the general population, the presence of sporadic breast cancers complicates analysis of the association with Lynch syndrome.
Additional cancer risks. Several other cancer types have been reported to occur in individuals with Lynch syndrome. In some cases, MSI and/or IHC testing of tumor tissue demonstrated concordance between the extracolonic cancer and the molecular diagnosis of the affected individual. While such findings suggest that the underlying presence of a pathogenic variant in an MMR gene contributed to the development of the cancer, data are not sufficient to demonstrate that the risk of developing these cancers is increased in individuals with Lynch syndrome.
Several types of sarcomas have been reported in individuals with an MMR
pathogenic variant, including fibrous histiocytomas, rhabdomyosarcomas, leiomyosarcoma, and liposarcoma [
Sijmons et al 2000,
den Bakker et al 2003,
Nilbert et al 2009].
Nilbert et al [2009] determined that six of eight sarcomas in individuals with Lynch syndrome exhibited defective MMR, suggesting that sarcomas may also be part of the spectrum of Lynch syndrome tumors. Due to the rarity of sarcomas it may be difficult to determine the magnitude of risk associated with Lynch syndrome.
Adrenocortical carcinoma (ACC) has also been reported in families with Lynch syndrome. The most extensive study of this association, performed through a hereditary cancer clinic at the University of Michigan, found that two (1.7%) of 114 individuals presenting with ACC had a family history consistent with Lynch syndrome and had an MMR
pathogenic variant identified. This association was further evaluated by case review of 135 individuals with pathogenic MMR variants, which identified two (1.4%) individuals who also had ACC [
Raymond et al 2013a].
Lynch Syndrome Variants
Muir-Torre syndrome is a historical term used to describe individuals presenting with the combination of sebaceous neoplasms of the skin and one or more internal malignancies, commonly those seen in Lynch syndrome. The types of sebaceous skin neoplasms described include: sebaceous adenomas, sebaceous epitheliomas, sebaceous carcinomas, and keratoacanthomas [Misago & Narisawa 2000].
Turcot syndrome is a historical term used to describe individuals presenting with CRC or colorectal adenomas in addition to tumors of the central nervous system. The clinical presentation varies from numerous colonic polyps to a single polyp or colorectal cancer. Turcot syndrome is usually caused by either a pathogenic variant in one of the MMR genes associated with Lynch syndrome or an APC pathogenic variant (see Differential Diagnosis and APC-Associated Polyposis Conditions).
Constitutional MMR deficiency (CMMRD). Rare individuals who are homozygous for pathogenic variants in MLH1, MSH2, MSH6, and PMS2 have been reported. Affected individuals often have onset of colon or small bowel cancer prior to the second decade of life. One third of children with biallelic pathogenic variants in an MMR gene have been reported to have more than ten polyps. Hematologic cancer, brain tumors, and café au lait macules have also been reported [Wimmer & Etzler 2008, Durno et al 2010, Bakry et al 2014]. The cutaneous phenotype in affected individuals may be remarkably similar to that seen in neurofibromatosis type I as nearly all will have café au lait macules [Wimmer 2012, Bakry et al 2014]. Features in the family history that increase suspicion of CMMRD:
However, this diagnosis should not be excluded if the family history is negative, as a significant number of children with a confirmed diagnosis of CMMRD will not have a family history consistent with Lynch syndrome. Bakry et al [2014] reported that a history of Lynch syndrome cancers was rarely noted in the family members of children with CMMRD.
Phenotype Correlations by Gene
Cancer risks vary among the genes associated with Lynch syndrome.
MSH2. Heterozygosity for an MSH2 pathogenic variant is associated with the greatest risk for extracolonic cancers.
MSH2 pathogenic variants have been reported more commonly than a pathogenic variant in the other three MMR genes in individuals with the Muir-Torre variant of Lynch syndrome [South et al 2008].
MSH6. Heterozygosity for a pathogenic variant in MSH6 is associated with MSI-low tumors. The colorectal cancers in families with an MSH6 pathogenic variant may be later in onset and more distally located than the cancers in families with Lynch syndrome resulting from a pathogenic variant in one of the other MMR genes; endometrial cancer is commonly observed in females with an MSH6 pathogenic variant [Wu et al 1999, Berends et al 2002]. Slightly lower risks for colorectal cancer and higher risks for endometrial cancer have been reported in families with an MSH6 pathogenic variant than in families with an MLH1 or MSH2 pathogenic variant [Berends et al 2002, Baglietto et al 2010].
PMS2. Heterozygosity for a PMS2 pathogenic variant is associated with the lowest risk (25%-32% risk) for any Lynch syndrome-related cancer [Senter et al 2008]. However, while the overall risk of CRC is lower, age of onset may still be early. A review of 234 PMS2 pathogenic variant carriers found that 8% were diagnosed before age 30 [Goodenberger et al 2016].
EPCAM. Deletions of EPCAM that result in epigenetic silencing of MSH2 are associated with a significantly increased risk for colorectal cancer. Kempers et al [2011] reported a low risk for endometrial cancer in those with deletions of EPCAM compared to pathogenic variants in an MMR gene.