Study: Cancer risk associated with inherited mutations in Lynch syndrome genes

IN-DEPTH REVIEW OF RESEARCH

Study background:

Lynch syndrome is the most common inherited cause of cancer affecting 1 in 300 people. It is most  commonly associated with colorectal cancer. Lynch syndrome is a cancer syndrome that results from pathogenic (deleterious) mutations in one of several genes.

The Lynch syndrome genes (except for EPCAM) encode proteins that work as a repair system to fix errors in DNA called mismatch repair (MMR). When these genes are mutated, cancers can occur in various organs including the GI tract (colon, rectum, small bowel, stomach), bile ducts of liver, pancreas, endometrium of uterus, ovaries, and upper urinary tract.

There have been several studies done to determine risk of cancer associated with mutations in the Lynch syndrome genes. However, many of these studies were retrospective, and involved only cases in which participants had cancer. This study is a large prospective study of participants with known mutations in a Lynch syndrome gene to determine how frequently cancer occurs, in what organs, at what age.

Researchers of this study wanted to know:

What is the risk of cancer associated with mutations in each Lynch syndrome gene for each organ. They also wanted to know how that cancer risk varies by age and gender.

Populations looked at in this study:

6350 participants age 25 -75 years were enrolled in the Prospective Lynch Syndrome Database (PLSD) and were followed for a total 51,646 years, an average of 7.3 years for each participant. These include two groups: an original group of 2823 patients from Germany and a newly enrolled group of 3527 patients from sites in the U.S., Canada, Australia, New Zealand and Europe. Among the combined group, there were 3480 women and 2870 men with a mean age of 46.8 years. These included 2607 MLH1 carriers, 2495 MSH2 carriers, 841 MSH6 carriers and 407 PMS2 carriers. EPCAM mutations were not tested.

Only patients with mutations that were classified as clinically actionable (class 4 or 5 by the International Society of Gastrointestinal Hereditary Database, InSiGHT) were included as participants.

Study design:

This study is an international multi-center, prospective, observational study. Data was collected from a new group of patients with pathogenic mutations in a Lynch syndrome gene. Researchers reanalyzed data from a prior study of Lynch syndrome patients in the same way as among the new group of patients. They evaluated the two groups statistically and determined that data from the two groups of patients could be validly combined.

Participants had an initial colonoscopy to look for colorectal cancer to enrollment. Any cancer present at the time of enrollment or before was called a prior cancer. Participants were asked about cancer each year after enrollment. Researchers collected data about mutation status, age at enrollment, age at update, age at any cancer, type of cancer and age at death. Cancer diagnoses were confirmed with clinical and pathological reports.

The annual rate of cancer was calculated in 5 year age groups from 25-75 years. Survival after cancer was estimated as survival from age at diagnosis until last observation or death.
Study findings:  

6350 participants had 1808 cancers during the study time period.

• The most common cancers seen were colon, skin, endometrial and rectal cancers:
  • 580 (31%) were colon cancer, average age 53.
  • 215 (12%) were skin cancer, average age 61.
  • 173 (10%) were endometrial cancer, average age 51.
  • 127 (7%) were rectal cancer, average age 56.

*Note: Different clinical sites collected data on skin cancers differently and not systematically so these are not included in the table below. Overall incidence of skin cancer at 12% is higher than lifetime average of 2.3% in the general population. However, because this data was not systematically collected, skin cancer was left out of the remainder of the analysis.

• The average age at diagnosis of cancer at each organ was reported
  • brain cancer: average age 56
  • endometrial cancer: average age 51
  • colon cancer:  average age 53
  • ovarian cancer: average age 47
  • pancreatic cancer: average age 61
  • prostate cancer: average age 62
  • rectal cancer: average age 56
  • skin cancer : average age 61
  • stomach cancer: average age 62
• Researchers estimated the incidence of cancer in patients with each mutation, by age group. For comparison, we show the percentage of people with a given Lynch syndrome mutation with cancer at a particular organ by age 75.

Table: Estimated incidence of cancer by organ at age 75 in Lynch syndrome carriers

*Note: Experts believe that EPCAM mutations linked to Lynch syndrome confer similar risks to MSH2.
**Note: There was no control groups for the study data; participants from this study are from multiple countries and the rates of cancer in this combined population without a Lynch syndrome mutation was not determined. Cancer in general U.S. population shown in the table is from NCI /SEER statistics for 2014-2016 at lifetime risk to provide context.

MLH1 and MSH2 had higher associated rates of cancer than MSH6 or PMS2.

• MLH 1 patients were estimated to have 71-81% risk of a cancer by age 75.
• MSH 2 patients were estimated to have 75-84% risk of a cancer by age 75.
• MSH 6 patients were estimated to have 42-62% risk of a cancer by age 75.
• PMS 2 patients were estimated to have 34 % risk of a cancer by age 75.
• People in the general population are estimated by 2014-2016 SEER data to have a lifetime risk of cancer of 39%. These rates include individuals older than 75.
• People with PMS2 mutations had no increase in colorectal, endometrial, ovarian or urinary tract cancers before age 50. However, researchers note that there is less data about PMS2 carriers’ follow up which may be a limitation.

There were some differences in risk of cancer that depended on the gender of the person with the mutations:

• Men were more likely than women to have colon, stomach, small bowel, gallbladder and pancreatic cancer.
• Rates of cancer in women with a MSH6 mutations was higher than in men with MSH6 mutations.
  • Both men and women with MSH6 mutations had a moderately increased risk of colorectal cancer (18-20% by age 75) as compared to the historic lifetime rates in the general population of just over 4%.
  • Women with MSH6 mutations had a significantly increased risk of ovarian cancer (11%) and uterine/endometrial cancer (41%). In U.S., ovarian and endometrial cancer occur in 3% of women in the general population.
  • Researchers note that with only 20% risk of colorectal cancer for people with MSH6 mutations, families may go unrecognized as having Lynch syndrome despite the much higher risk (41%) of endometrial cancer for women with MSH6 mutations.

Survival after cancer diagnosis and before 65 years was determined for participants with MLH1, MSH2 and MSH6 mutations.

• The majority of patients with Lynch syndrome mutations survived cancers 10 or more years including cancers at the following organ sites: colon (88%), rectum (70%), endometrium (89%), ovary (84%), prostate (70%), breast (82%), upper urinary tract (67%) and bladder (68%).
• The majority of patients with pancreatic (29%), bile duct (42%) and brain cancer (15%) survived less than 10 years after diagnosis.

There was only a slight increase in breast cancer risk associated with participants in this study who had Lynch syndrome mutations.

• Women with MSH2, MSH6 or PMS2 mutations had a 14-15% breast cancer risk by age 75. This is a modestly elevated risk as compared to the general population (13% lifetime risk).
• Mean age of diagnosis for breast cancer was age 57.
• This contrasts with an earlier study by Roberts and colleagues which found higher breast cancer risks for MSH6 and PMS2 (reviewed by XRAY [https://www.facingourrisk.org/XRAYS/lynch-syndrome-breast-cancer]).
• All participants in this study had mutations in Lynch syndrome genes that were associated with significantly increased risk for colorectal cancer based on classification as class 4 or 5 by the International Society of Gastrointestinal Hereditary Database, InSiGHT).  The participants in the Roberts et al study were breast cancer patients who had mutations that were considered classified as pathogenic by the American College of Medical Genetics and Genomics/Association for Molecular Pathology. It is possible that different mutations in Lynch syndrome genes may have different cancer risks.  A comprehensive study [https://doi.org/10.1200/PO.19.0027] of patients with a family history of breast or ovarian cancer published by Jessica Stoll and colleagues in February 2020 showed no increase in breast cancer risk when comparing people with various Lynch syndrome mutations to those screened in the same way who do not have Lynch syndrome mutations. This suggests that the risk of breast cancer for people with Lynch syndrome mutations is similar to the risk of breast cancer in the general population.

Different cancer risks are associated with each of the 4 Lynch syndrome genes. These researchers suggest that recommendations for genetic counseling, surveillance and clinical management may need to differ for patients depending on which gene is affected.

• In particular they suggest that because only a few males with MSH6 mutations will develop cancers, it may appear that cancer risk "skips" a generation in MSH6 families and they will not be identified by current clinical criteria that assumes similar cancer rates in both genders.

Limitations:

• A major limitation of this study is the way in which the participants were selected. Mutations included were only those that have a strong gastrointestinal cancer association which may skew the results. Participants were selected based on diagnosis of Lynch syndrome typically from colorectal cancer patients in the family. There may be a bias if different mutations within these genes affect  other organs differently  (i.e., less colorectal cancer and more incidence at another site).
  • All participants in this study had a mutation that was  known to be associated with a high risk of colorectal cancer. Only mutations in these genes (class 4 or 5 mutations) were tested for association of cancer in other organs.
  • Mutations in  MLH1, MSH2, MSH6 or PMS2 that are associated with low or modest risk of colorectal cancer (so called "low penetrance" mutations) were not included.  Therefore, participants with mutations that increase colorectal cancer only modestly were not included in enrollment. The risk of other types of cancer associated with mutations at modest risk for colorectal cancer was not determined.
  • Mutations in MLH1, MSH2, MSH6 or PMS2 that are not associated with colorectal cancer were not tested. If mutations exist in the Lynch syndrome genes that cause primarily non-GI cancers (e.g. ovarian or breast cancer), the risk of these cancers may be underestimated.
• A comprehensive study of patients with Lynch syndrome mutations selected based on breast cancer risk criteria showed no increase in breast cancer risk. This suggests that the risk of breast cancer for people with Lynch syndrome mutations is similar to the risk of breast cancer in the general population.
• Another potential limitation of this study how frequency of cancers were estimated.
  • The cumulative incidence is estimated based on the participants whose average age is 46 and by counting until the first cancer diagnosed. While these participants were followed for an average of 7 years, cancer incidence at older ages (with fewer participants in this group) may underestimate cancer rates for some cancer types. A better study design would be a true prospective study that follows all patients until death. This is not typically practical for researchers.
• Researchers note that it is possible that geographic differences between care or follow care in different countries might affect cancer rates. The majority of the participants were in Europe in the initial group. However, they note that there were no significant differences between the group recruited from Germany and the second group recruited from the Americas, Australia, Asia and Germany or between prior studies, suggesting geographic differences have little effect on these results.
• There was a limited number of years that patients were followed after enrollment (average 7 years). There were statistical calculations of risk based on the cancers that were observed. It remains possible that actual risks of some cancers particularly in older ages may be underestimated.
  • Researchers particularly note that there was less data about PMS2 carriers’ follow up years which may be a limitation in understanding what the risks are for these participants.

Conclusions:

Different cancer risks are associated with each of the 4 Lynch syndrome genes. There are significant risks of cancers in addition to colorectal cancer that warrant clinical attention. Recommendations for genetic counseling, surveillance and clinical management may need to be tailored to reflect which Lynch syndrome gene is affected.

Share your thoughts on this XRAYS article by taking our brief survey. 

Posted 2/21/20