Abstract
Background & Aims
Lynch syndrome is the most common inherited cause of gastrointestinal cancer and increases risk for a variety of malignancies, including gastric cancer. We aimed to identify clinical factors associated with gastric cancer in carriers of germline variants causing Lynch syndrome.
Methods
We collected data from 52,758 consecutive individuals tested for genetic variants associated with Lynch syndrome from June 2006 through July 2013 at a commercial laboratory. We obtained clinical and demographic data, as well as information on personal and family histories of cancer (first- and second-degree relatives) from forms completed by ordering providers. We performed multivariate logistic regression to identify clinical factors associated with gastric cancer in carriers of mutations that cause Lynch syndrome (pathogenic mutations).
Results
After we excluded individuals with missing clinical data (n=1664) or with multiple pathogenic mutations (n=8), we analyzed data from 51,086 individuals. Of these, 3828 persons carried pathogenic mutations (1346 with mutations in MLH1, 1639 with mutations in MSH2, 670 with mutations in MSH6, 145 with mutations in PMS2, and 28 with mutations in EPCAM). Of the 3828 carriers of pathogenic mutations, 41 (1.1%) had a previous gastric cancer and 350 (9.1%) had 1 or more first- or second-degree relatives with gastric cancer. In multivariate analysis, male sex (odds ratio [OR], 2.82; 95% CI, 1.48–5.38), older age (OR, 2.07 per 10 years; 95% CI, 1.64–2.61), mutations in MLH1 (OR, 6.53; 95% CI, 1.50–28.42) or MSH2 (OR, 5.23 compared to mutations in MSH6, PMS2, or EPCAM; 95% CI, 1.21–22.71), and first-degree relative with gastric cancer (OR, 2.52; 95% CI, 1.42–4.45), but not second-degree relatives (OR, 1.12; 95% CI, 0.40–3.18) were independently associated with gastric among carriers of pathogenic mutations.
Conclusions
In an analysis of data from almost 4000 carriers of Lynch syndrome-associated mutations, we found history of gastric cancer to be independently associated with male sex, older age, mutations in MLH1 or MSH2, and with having a first-degree relative with gastric cancer. These findings suggest that personalized, risk-stratified approaches to gastric cancer surveillance may be appropriate for individuals with Lynch syndrome-associated mutations.
Keywords: HNPCC, stomach, familial, screening
Introduction
Lynch syndrome (LS) is the most common inherited gastrointestinal cancer predisposition syndrome with an estimated 1:279 prevalence in the general population.1 LS is caused by autosomal dominant inheritance of pathogenic germline mutations in the DNA mismatch repair (MMR) genes (MLH1, MSH2, MSH6, and PMS2) or EPCAM, and LS carriers are at risk of developing a wide variety of malignancies, most commonly colorectal (CRC), endometrial, and ovarian cancers.2 For LS carriers, colonoscopies,3 hysterectomy with salpingo-oophorectomy,4 and aspirin5 are all proven effective interventions that can reduce the risk of these three most common malignancies. Comparably little is known about factors that predispose carriers to other LS spectrum malignancies. Addressing this knowledge gap is of particular importance with respect to gastric cancer (GC), since it is a particularly lethal cancer with a <30% 5-year overall survival in the United States across all disease stages, and only minimal improvements in early detection rates have been appreciated over recent years.6 Prior registry data have suggested that LS carriers have a nearly ten-fold increased risk of GC compared to non-carriers,7 yet >90% of individuals with LS will never develop GC2. Thus, a better understanding is needed regarding which individuals with LS could most benefit from GC screening.
In spite of these uncertainties, guidelines from numerous professional societies (Table 1)8–14 recommend consideration of some form of GC screening and risk-reduction with esophagogastroduodenoscopy (EGD) and/or testing for Helicobacter pylori (H. pylori) as an easily treatable modifiable risk factor for some or all individuals with LS. Due to a paucity of data, however, there is no concordance amongst guidelines on which individuals with LS are most likely to benefit from screening, though various guidelines speculate that LS carriers with family histories of GC, of Asian ancestry, or with other as-yet unidentified risk factors may be particularly strong candidates for surveillance. The primary hypothesis of this study was that specific clinical factors predispose LS carriers to GC, and our goal was to study the role of potential risk factors by analyzing >51,000 individuals who underwent LS genetic testing due to a personal/family history of cancer.
Table 1:
Lynch Syndrome Gastric Cancer Screening Recommendations and Guidelines from Various Professional Societies
| Professional Society | Recommendations |
|---|---|
| American College of Gastroenterology (ACG)8 | - Consider baseline EGD with biopsy for H. pylori for all Lynch syndrome carriers at age 30–35 |
| - Consider ongoing surveillance every 3–5 years for Lynch syndrome carriers with a family history of gastric or duodenal cancer | |
| United States Multi-Society Task Force9 | - Consider baseline EGD with biopsy for H. pylori for all Lynch syndrome carriers at age 30–35 |
| - Consider ongoing surveillance every 2–3 years based on individual patient risk factors | |
| National Comprehensive Cancer Network (NCCN)10 | - Consider EGD surveillance with H. pylori testing every 3–5 years for all Lynch syndrome carriers, beginning at age 40 |
| - Lynch syndrome carriers with a family history of gastric/small bowel cancer and those of Asian descent may benefit from surveillance | |
| European Society for Medical Oncology (ESMO)11 | - Consider EGD surveillance every 1–3 years in high risk subsets of Lynch syndrome carriers |
| - Test all Lynch syndrome carriers for H. pylori | |
| American Society of Clinical Oncology (ASCO)12 | - Consider EGD surveillance every 1–3 years in high risk subsets of Lynch syndrome carriers |
| - Test all Lynch syndrome carriers for H. pylori | |
| European Hereditary Tumour Group (EHTG; formerly the Mallorca Group)13 | - Consider EGD surveillance in Lynch syndrome carriers from countries with a high background incidence of gastric cancer (Korea and Japan) |
| - Test all Lynch syndrome carriers for H. pylori | |
| European Society of Digestive Oncology (ESDO)14 | - Perform EGD surveillance every 1–2 years in all Lynch syndrome carriers, beginning no later than age 30 |
| - Test all Lynch syndrome carriers for H. pylori |
Abbreviations: EGD, esophagogastroduodenoscopy
Materials and Methods
The study cohort consisted of 52,758 consecutively-ascertained individuals unknown to be related to one another who underwent clinical germline testing of ≥2 LS genes (MLH1, MSH2, MSH6, PMS2, and EPCAM) at a commercial laboratory (Myriad Genetic Laboratories, Inc., Salt Lake City, UT) between 6/2006–7/2013; individuals undergoing only single-site germline testing for a specific LS variant were not included. All germline testing was ordered as syndrome-specific testing for LS genes only, rather than multisyndromic panel testing. Germline testing methodologies and variant classification are as previously described.15
Clinical data were obtained from the test requisition forms completed by the healthcare provider ordering germline testing as previously described.15 Collected data included sex, race, age, personal history of cancer (including age at diagnosis), and family history of cancer in first- (parents, siblings, and children) and second-degree (grandparents, aunts, uncles, and grandchildren) relatives (FDRs and SDRs, respectively); family history data analyzed were limited to the side of the family suspected to be affected by LS, as previously described.15 These clinical data along with germline testing results were provided to investigators as a de-identified dataset by Myriad Genetic Laboratories, Inc. Individuals with pathogenic or likely pathogenic germline mutations in any of the five aforementioned genes were collectively defined as LS carriers. Individuals lacking such mutations (including those found to have germline variants of uncertain significance) were collectively defined as non-carriers.
For this analysis, subject age was defined as the age at LS genetic testing. Family history of GC was analyzed as both a categorical (yes/no) and a continuous variable (number of FDRs and SDRs with GC). Pearson’s chi-square test or Fisher’s exact test were used to determine differences between categorical groups. Continuous variables were compared using the Mann–Whitney U test. Analysis of clinical factors associated with GC were performed using univariate and multivariate logistic regression models. Selected variables found to have a significant association with personal history of GC on univariate analysis and other clinical factors previously described as being linked to risk of LS-associated GC (e.g. specific MMR gene altered) were included in the multivariate analysis. For multivariate analysis, LS carriers with MSH6, PMS2, and EPCAM mutations were analyzed jointly due to small sample size. Clinical factors with possible association with LS-associated GC were assessed via their odds ratios with 95% confidence intervals. P-values were two-sided and considered statistically significant at <0.05. All statistical analyses were performed using R software v3.4.3 (R Foundation for Statistical Computing, Vienna, Austria). The study was approved by the Dana-Farber/Harvard Cancer Center Institutional Review Board.
Results
Of the 52,758 individuals in the study cohort, 1672 were excluded from analysis due to missing clinical data (n=1664) or identification of multiple pathogenic LS mutations (n=8). Of the final study population of 51,086 individuals (Table 2), 10,599 (20.7%) were male, 29,880 (58.5%) were Caucasian, 1251 (2.4%) were Asian, 26,721 (52.3%) had a personal history of any LS spectrum cancer, 266 (0.5%) had a personal history of GC, 4633 (9.1%) had a family history of GC in FDRs/SDRs, and median age at LS testing was 49.0 years (IQR 41.0–58.0 years). 3828/51,086 (7.5%) individuals were LS carriers (1346 MLH1, 1639 MSH2, 670 MSH6, 145 PMS2, 28 EPCAM). LS carriers were significantly more likely than non-carriers to be male (36.5% versus 19.5%; P<0.001) and to have a personal history of any LS spectrum cancer (83.2% versus 49.8%; P<0.001), including GC (1.1% versus 0.5%; P<0.001). LS carriers were significantly more likely than non-carriers to have a family history of CRC, endometrial, ovarian, small bowel, and urinary tract cancers, but there was no significant difference in the number of LS carriers versus non-carriers with a family history of GC (9.1% versus 9.1%; P=0.87).
Table 2:
Clinical characteristics of 51,086 consecutively ascertained individuals undergoing germline testing for Lynch syndrome
| Total cohort (N=51,086) | Lynch syndrome carriers (N=3828) | Non-carriers (N=47,258) | P value | |
|---|---|---|---|---|
| N (%) | N (%) | N (%) | ||
| Male | 10599 (20.7) | 1397 (36.5) | 9202 (19.5) | <0.001 |
| Female | 40487 (79.3) | 2431 (63.5) | 38056 (80.5) | |
| Median age at germline testing, years (IQR) | 49.0 (41.0, 58.0) | 50.0 (41.0, 59.0) | 49.0 (41.0, 58.0) | 0.001 |
| Race | 0.001 | |||
| Caucasian/White | 29880 (58.5) | 2271 (59.3) | 27609 (58.4) | |
| African American/Black | 2349 (4.6) | 185 (4.8) | 2164 (4.6) | |
| Asian | 1251 (2.4) | 111 (2.9) | 1140 (2.4) | |
| Other/Multiple | 6591 (12.9) | 528 (13.8) | 6063 (12.8) | |
| Missing/No Answer | 11015 (21.6) | 733 (19.1) | 10282 (21.8) | |
| Personal history of Lynch syndrome spectrum cancer | ||||
| Any Lynch syndrome spectrum cancer | 26721 (52.3) | 3186 (83.2) | 23535 (49.8) | <0.001 |
| Gastric cancer | 266 (0.5) | 41 (1.1) | 225 (0.5) | <0.001 |
| Colorectal cancer | 19866 (38.9) | 2496 (65.2) | 17370 (36.8) | <0.001 |
| Endometrial cancer* | 6135 (15.2)* | 908 (37.4)* | 5227 (13.7)* | <0.001 |
| Ovarian cancer* | 2161 (5.3)* | 194 (8.0)* | 1967 (5.2)* | <0.001 |
| Pancreatic cancer | 266 (0.5) | 23 (0.6) | 243 (0.5) | 0.47 |
| Small bowel cancer | 192 (0.4) | 56 (1.5) | 136 (0.3) | <0.001 |
| Urinary tract cancer | 717 (1.4) | 158 (4.1) | 559 (1.2) | <0.001 |
| Other Lynch syndrome spectrum cancer*** | 494 (1.0) | 166 (4.3) | 328 (0.7) | <0.001 |
| Multiple Lynch syndrome spectrum cancers | 3885 (7.6) | 878 (22.9) | 3007 (6.4) | <0.001 |
| Family history of Lynch syndrome spectrum cancer** | ||||
| Gastric cancer | 4633 (9.1) | 350 (9.1) | 4283 (9.1) | 0.87 |
| Colorectal cancer | 33952 (66.5) | 2997 (78.3) | 30955 (65.5) | <0.001 |
| Endometrial cancer | 9771 (19.1) | 966 (25.2) | 8805 (18.6) | <0.001 |
| Ovarian cancer | 7603 (14.9) | 411 (10.7) | 7192 (15.2) | <0.001 |
| Pancreatic cancer | 3619 (7.1) | 253 (6.6) | 3366 (7.1) | 0.23 |
| Small bowel cancer | 357 (0.7) | 70 (1.8) | 287 (0.6) | <0.001 |
| Urinary tract cancer | 3417 (6.7) | 369 (9.6) | 3048 (6.4) | <0.001 |
| Other Lynch syndrome spectrum cancer*** | 3949 (7.7) | 310 (8.1) | 3639 (7.7) | 0.38 |
Denominator is for females only
First and second-degree relatives
Includes hepatobiliary cancers, brain cancers/gliomas, and sebaceous neoplasms of the skin
The 41 LS carriers with prior GC were diagnosed with GC at a median age of 56.5 years (IQR 49.25–65.5 years; age at GC diagnosis missing for 3 LS carriers; Table 3), whereas the 225 non-carriers with GC were diagnosed at a median age of 52.0 years (IQR 40.0–62.0 years; age at GC diagnosis missing for 33 non-carriers; P=0.087). LS carriers with GC underwent LS genetic testing at a median of 1.5 years (IQR 0.25–7.5 years) after their GC diagnosis. 16/41 (39.0%) LS carriers with GC were diagnosed with GC as their first LS spectrum cancer (4 had no other LS spectrum cancer, 5 had GC and later developed CRC, and 7 developed GC concurrent with another LS spectrum cancer [6 CRC, 1 urinary tract cancer]) and 4/41 (9.8%) were diagnosed with GC and another LS spectrum cancer, though the sequence of these diagnoses was unknown. Conversely, 2¼1 (51.2%) LS carriers with GC were diagnosed with ≥1 LS spectrum cancer prior to their GC (20 with ≥1 CRC, 4 endometrial cancer, 2 urinary tract cancer, and 4 other LS spectrum cancers); among these 21 LS carriers, median age at first LS spectrum cancer was 46 years (IQR 40–50 years) and median time from first LS spectrum cancer to GC diagnosis was 16 years (IQR 9–24 years).
Table 3:
Characteristics of individuals with Lynch syndrome
| Carriers with a personal history of gastric cancer (N=41) | Carriers without a personal history of gastric cancer (N=3787) | P value | |
|---|---|---|---|
| N (%) | N (%) | ||
| Male | 25 (61.0) | 1372 (36.2) | 0.002 |
| Female | 16 (39.0) | 2415 (63.8) | |
| Median age at germline testing, yrs (IQR) | 61.0 (56.0, 73.0) | 50.0 (41.0, 58.0) | <0.001 |
| Median age at gastric cancer diagnosis, yrs (IQR)* | 56.5 (49.25, 65.5) | - | - |
| Race | |||
| Caucasian/White | 22(53.7) | 2249 (59.4) | 0.560 |
| African American/Black | 2 (4.9) | 183 (4.8) | >0.99 |
| Asian | 2 (4.9) | 109 (2.9) | 0.771 |
| Other/Multiple | 8 (19.5) | 520 (13.7) | 0.401 |
| Missing/No Answer | 7 (17.1) | 726 (19.2) | 0.889 |
| MMR gene mutation | 0.115 | ||
| MLH1 | 19 (46.3) | 1327 (35.0) | 0.179 |
| MSH2 | 20 (48.8) | 1619 (42.8) | 0.537 |
| MSH6 | 2 (4.9) | 668 (17.6) | 0.036 |
| PMS2 | 0 (0.0) | 145 (3.8) | 0.406 |
| EPCAM | 0 (0.0) | 28 (0.7) | >0.99 |
| Personal history of Lynch syndrome spectrum cancer | |||
| Any Lynch syndrome spectrum cancer (except gastric) | 37 (90.2) | 3145 (83.0) | 0.29 |
| Colorectal cancer | 35 (85.4) | 2461 (65.0) | 0.005 |
| Endometrial cancer** | 5/16 (31.3) | 903/2415 (37.4) | 0.080 |
| Ovarian cancer** | 0/16 (0.0) | 194/2415 (8.0) | 0.63 |
| Pancreatic cancer | 0 (0.0) | 23 (0.6) | 1.00 |
| Small bowel cancer | 0 (0.0) | 56 (1.5) | 1.00 |
| Urinary tract cancer | 5 (12.2) | 153 (4.0) | 0.025 |
| Other Lynch syndrome spectrum cancer*** | 6 (14.6) | 160 (4.2) | 0.008 |
| Family history of Lynch syndrome spectrum cancer**** | |||
| Gastric cancer | 11 (26.8) | 339 (9.0) | <0.001 |
| Colorectal cancer | 28 (68.3) | 2969 (78.4) | 0.170 |
| Endometrial cancer | 12 (29.3) | 954 (25.2) | 0.677 |
| Ovarian cancer | 5 (12.2) | 406 (10.7) | 0.798 |
| Pancreatic cancer | 7 (17.1) | 246 (6.5) | 0.016 |
| Small bowel cancer | 1 (2.4) | 69 (1.8) | 0.533 |
| Urinary tract cancer | 6 (14.6) | 363 (9.6) | 0.281 |
Age at gastric cancer diagnosis missing for 3 Lynch syndrome carriers
Denominator is for female carriers only
Includes hepatobiliary cancers, brain cancers/gliomas, and sebaceous neoplasms of the skin
First- and second-degree relatives
On univariate analysis, LS carriers with GC were significantly more likely than those without GC to be male (61.0% versus 36.2%; P=0.002), older (P<0.001), and have a family history of GC (26.8% versus 9.0%; P<0.001) or pancreatic cancer (17.1% versus 6.5%; P=0.016) in FDRs/SDRs. LS carriers with GC were significantly less likely to harbor MSH6 mutations versus LS carriers without GC (4.9% versus 17.6%; P=0.036). There was no significant difference in the proportion of LS carriers with and without GC who were of Asian ancestry (4.9% versus 2.9%; P=0.771), though such individuals only accounted for 2.9% of all LS carriers.
On multivariate logistic regression analysis, age (OR 2.07 per 10 years; 95% CI 1.64–2.61), male sex (OR 2.82; 95% CI 1.48–5.38), pathogenic germline MLH1 (OR 6.53; 95% CI 1.50–28.42) and MSH2 mutations (OR 5.23; 95% CI 1.21–22.71; reference: MSH6/PMS2/EPCAM mutations), and number of FDRs (OR 2.52; 95% CI 1.42–4.45), but not number of SDRs with GC (OR 1.12; 95% CI 0.40–3.18), were each significantly associated with prior GC among LS carriers (Table 4). Stratifying by type of FDR, multivariate analysis showed that family history of GC both in parents and siblings was significantly associated with GC in LS carriers (there were too few individuals with children with GC for analysis; Supplemental Table).
Table 4:
Multivariate logistic regression analysis for association between various clinical factors and personal history of GC among Lynch syndrome carriers
| Odds ratio | 95% CI | P value | ||
|---|---|---|---|---|
| Age | 2.07 (per 10 years) |
1.64 | 2.61 | <0.001 |
| Male sex | 2.82 | 1.48 | 5.38 | 0.002 |
| Female sex | 1.0 | Reference | - | |
| MMR gene mutation | ||||
| MLH1 | 6.53 | 1.50 | 28.42 | 0.013 |
| MSH2 | 5.23 | 1.21 | 22.71 | 0.027 |
| MSH6/PMS2/EPCAM | 1.0 | Reference | - | |
| Number of 1st degree relatives with GC | 2.52 | 1.42 | 4.45 | 0.002 |
| Number of 2nd degree relatives with GC | 1.12 | 0.40 | 3.18 | 0.826 |
Abbreviations: GC, gastric cancer; MMR, mismatch repair
Discussion
The link between LS and GC risk has been recognized since the original seminal descriptions of “Family G” by Warthin and Lynch in 1913 and 1971, respectively.16, 17 In spite of this longstanding association, however, there have been minimal data to date regarding patient-specific factors that predict for GC among LS carriers. This has limited efforts at improving early detection. In this study of >50,000 individuals clinically suspected to have LS and >3800 confirmed LS carriers, we identified male sex, increasing age, family history of GC in FDRs, and MLH1 or MSH2 mutations as factors being independently associated with GC among LS carriers, suggesting that a particular subset of LS patients may be most likely to benefit from GC screening strategies.
Recent data2 from the multinational Prospective Lynch Syndrome Database have similarly suggested that older LS carriers with pathogenic MLH1 or MSH2 mutations are at particular risk for upper GI cancers, with such individuals having a 7.1% and 7.7% cumulative risk of GC by age 75, respectively, with a 61% 5-year survival rate. Other prior data18–20 have demonstrated that MLH1 and MSH2 carriers, especially males, are at higher risk for GC than female LS carriers and those with germline mutations in other LS genes.
Our data confirm these prior observations and additionally suggest that these risks may be further pronounced in LS carriers with a strong family history of GC, with more than a doubling in the strength of association with personal history of GC for each affected FDR. A recent Dutch study21 demonstrated that H. pylori infection rates are no higher among LS patients with a FDR with GC versus those without such family history, suggesting that the etiology of this familial risk is due to other as-yet unidentified shared environmental and/or genetic factors (e.g. dietary patterns, obesity). Another Dutch registry study18 concluded that there was no evidence of familial clustering of GC within LS families since only 10 of 32 (31.3%) GC patients in their study had a family history of GC. Interestingly, this rate is comparable to the 26.8% of LS GC patients in our cohort with a family history of GC, yet our multivariate analysis nonetheless confirmed a significant incremental association between family history of GC and personal history of GC among LS carriers after controlling for multiple other clinical factors, including underlying specific MMR gene altered.
Virtually all current professional society guidelines (Table 1)8–12, 14, 22 acknowledge the paucity of data regarding GC screening in LS carriers and, at most, suggest that such screening be “considered.” One notable exception are the recently published European Society of Digestive Oncology recommendations14 which do formally recommend GC screening in all LS carriers with H. pylori testing and EGD every 1–2 years, beginning no later than age 30. In contrast, guidelines from the United States Multi-Society Task Force,9 the European Society for Medical Oncology,11 and the American Society of Clinical Oncology12 all recommend that ongoing GC screening be limited to LS carriers with high risk for GC, but provide no guidance on how to characterize such carriers as high risk. Likewise, guidelines from the American College of Gastroenterology8 and the National Comprehensive Cancer Network10 suggest that LS carriers with family histories of GC should be prioritized for GC screening, though data to support this hypothesis have been lacking.
Our findings thus make important contributions to address these gaps, as they provide support for the notion that male sex, pathogenic MLH1 and MSH2 mutations, and number of FDRs with GC are each independently associated with likelihood of GC in LS carriers, suggesting that these factors should be used to guide personalized risk-stratified approaches to GC screening. While we emphasize that our data are insufficient to justify a specific approach to GC screening, we propose that LS carriers with ≥1 of these factors be considered for ongoing GC surveillance with EGD and biopsy at least every 3 years, with consideration of 1–2 year intervals for individuals with ≥2 factors and/or other identifiable risk factors (e.g. H. pylori, intestinal metaplasia). Prospective studies will be critical to defining the optimal approach for personalized risk-stratified GC surveillance in LS carriers. Notably, recommendations from the NCCN10 and the European Hereditary Tumour Group22 also suggest that LS carriers of Asian ancestry may benefit most from GC screening, although our data are insufficient to evaluate this hypothesis.
Although professional society guidelines8–14 suggest at least consideration of EGD-based GC screening in LS carriers, there are scant data about the efficacy of such screening. A small prospective study23 of one-time EGD screening in 73 Finnish LS carriers identified 26% with H. pylori, 32% with inflammatory changes, 14% with atrophic gastritis, and 14% with intestinal metaplasia, though no GCs were detected. Another small retrospective single-institution study24 of 66 LS carriers (21 of whom underwent EGD surveillance) found that 19% of screened individuals had potential precursor findings on EGD (e.g. H. pylori, intestinal metaplasia), but none had GC. More recently, observational data on 44 GC patients with known/presumed LS in the German Consortium for Familial Intestinal Cancer Registry25 demonstrated that 78% of GCs diagnosed within 30 months of prior EGD were stage I (none were stage IV) whereas only 23% of GCs diagnosed ≥30 months from last EGD were stage I, suggesting that regular EGDs may facilitate downstaging of GCs in LS carriers.
The limited data on the efficacy of screening is a real-world clinical dilemma for all LS-associated cancers (aside from CRC3), not just GC.8 While prophylactic surgery is effective at reducing the incidence of LS-associated endometrial and ovarian cancers,4 it remains highly uncertain whether there is any meaningful benefit to screening for other associated cancers including GC, small bowel, pancreatic, biliary, or urinary tract cancers, even though LS carriers are clearly at significantly increased risk for these potentially lethal malignancies.7 Rather than screening all LS carriers for every possible LS spectrum cancer, however, one common (albeit not evidence-based) real-world practice is to selectively screen for a specific non-colorectal non-gynecologic LS spectrum cancer only if there is a known family history of that particular cancer.8 Our finding that family history of GC in FDRs is independently and incrementally associated with LS carriers’ likelihood of GC now supports this approach to family history-based risk stratification while also identifying other factors that may help guide personalized screening. Further study is needed to determine whether family history of other LS spectrum cancers can be similarly used to identify carriers at particular risk for such malignancies.
Strengths of our study include the large size and consecutively ascertained nature of the cohort. By ascertaining patients through a large commercial laboratory, we were able to obtain a cohort that was more geographically diverse than would be expected from a single-institution registry, and likely representative of the overall U.S. population of patients undergoing LS genetic testing (though predominantly Caucasian and female26). Unlike other studies that have analyzed GC in LS, we were able to examine family history of GC as both a continuous and categorical variable, allowing for a more refined understanding of its contribution to GC risk. Furthermore, by collecting clinical data at the time of germline testing through a commercial laboratory, this ensured that cases and controls (LS carriers with and without prior GC, respectively) were ascertained through comparable processes. Our study also included LS carriers with mutations in all five associated genes, whereas most prior studies have focused only on individuals with MLH1, MSH2, and MSH6 mutations.
There are several important limitations to this study. Since clinical data were obtained from test request forms, we were unable to confirm the accuracy or completeness of reported personal/family cancer history data. Importantly, however, one recent study analyzed the completeness and accuracy of clinical data from such commercial genetic testing laboratory test request forms and found >99% accuracy for cancer diagnoses reported in test subjects, FDRs, and SDRs, though varying completeness of family history data and particularly incomplete data for third- and fourth-degree relatives.27
Additional limitations include this being a cross-sectional rather than longitudinal analysis, and we were unable to prospectively study the impact of various clinical factors on GC risk or survival, and were unable to estimate the cumulative risk of GC by age. Subjects were specifically referred for germline LS testing (rather than multisyndromic panel testing), but the degree to which their personal and/or family history of GC contributed to clinical suspicion for LS likely varied, as evidenced by the heterogeneous patterns and timing of prior LS spectrum cancers among carriers, which introduces the potential for ascertainment bias. We acknowledge the possibility of survival bias, since some GC patients may not have survived long enough after a GC diagnosis to undergo LS genetic testing and thus would not have been included in the study cohort. We lacked data on modifiable risk factors such as H. pylori status, diet, tobacco use, obesity, or GC screening behaviors. We also lacked data on age at GC diagnosis for family members with GC. This was a predominantly Caucasian United States cohort, and was underpowered to assess whether LS patients of Asian ancestry are particularly predisposed to GC, as has been speculated.10, 13 Additionally, since all patients in the cohort were clinically referred for germline LS testing, we acknowledge that the non-carriers in our study likely had a higher prevalence of GC and other LS spectrum cancers than the general population, and thus our data cannot definitively quantify the risk of GC among LS carriers versus the general population. Lastly, PMS2 carriers were underrepresented in our study cohort compared to their estimated prevalence in the general population,1 though recent registry data have suggested that PMS2 carriers may have no meaningfully increased risk of GC compared to the general population.28
In conclusion, our data provide important additions to the existing literature on clinical factors associated with GC in LS carriers. In addition to confirming prior data demonstrating that sex, age, and specific MMR gene influence GC risk, we demonstrate that family history of GC in FDRs is independently and incrementally associated with GC in LS carriers. Interestingly, male sex, age, and family history of GC are all known risk factors for sporadic, non-LS GC as well, raising the hypothesis that other traditional GC risk factors may likewise compound the inherited GC risk seen in LS carriers.29 Together, such data can facilitate the development and assessment of personalized risk-stratified early detection strategies in LS carriers at highest risk for GC (e.g. EGDs every 3 years for those with ≥1 such clinical factor and more frequently for those with increasing number of such factors), while also suggesting that carriers lacking these factors might be able to safely forego low-yield and potentially unnecessary EGD surveillance. Furthermore, the observation that family history of GC is independently associated with LS-associated GC, raises important hypotheses for other uncommon but potentially deadly LS spectrum malignancies (e.g. small bowel cancer, pancreaticobiliary cancer, urinary tract cancer, gliomas) that should be further studied to guide personalized early detection strategies.
Supplementary Material
WHAT YOU NEED TO KNOW.
Background
There are limited data on patient-specific risk factors for extracolonic cancers, including gastric cancer, in individuals with Lynch syndrome. Knowledge about such factors could allow for personalized approaches to gastric cancer screening and risk-reduction in Lynch syndrome carriers.
Findings
This was a cross-sectional analysis of cohort of 52,758 consecutively ascertained individuals undergoing clinical evaluation for Lynch syndrome, including 3828 confirmed carriers of mutations causing Lynch syndrome. Male sex, age, MLH1 and MSH2 mutations, and number of first-degree relatives with gastric cancer were independently associated with a personal history of gastric cancer among Lynch syndrome carriers.
Implications for Patient Care
Personalized, risk-stratified approaches to gastric cancer surveillance may be appropriate for individuals with Lynch syndrome-associated mutations (e.g. more frequent EGD screening in those with increasing number of risk factors, and foregoing of repeated EGDs in those without such factors).
Acknowledgments
Funding Information
This work was supported by the National Institutes of Health (National Cancer Institute) K24CA113433 (Dr. Syngal), R01CA132829 (Dr. Syngal), K07CA151769 (Dr. Kastrinos), 4P30CA006516-51 (Drs. Parmigiani and Syngal), and The Pussycat Foundation Helen Gurley Brown Presidential Initiative (Dr. Ukaegbu). The study sponsors had no role in the design, collection, analysis, or interpretation of data, and had no tole in the writing of the manuscript.
Abbreviations
- LS
Lynch syndrome
- GC
gastric cancer
- EGD
esophagogastroduodenoscopy
- FDR
first-degree relative
- SDR
second-degree relative
- OR
odds ratio
- MMR
mismatch repair
Footnotes
Disclosures
Dr. Syngal is a consultant for Myriad Genetic Laboratories, Inc. and has rights to an inventor portion of licensing revenues from PREMM5. The remaining authors report no other conflicts of interest.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
References
- 1.Win AK, Jenkins MA, Dowty JG, et al. Prevalence and Penetrance of Major Genes and Polygenes for Colorectal Cancer. Cancer Epidemiol Biomarkers Prev 2017;26:404–412. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Møller P, Seppala TT, Bernstein I, et al. Cancer risk and survival in path_MMR carriers by gene and gender up to 75 years of age: a report from the Prospective Lynch Syndrome Database. Gut 2018;67:1306–1316. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Järvinen HJ, Aarnio M, Mustonen H, et al. Controlled 15-year trial on screening for colorectal cancer in families with hereditary nonpolyposis colorectal cancer. Gastroenterology 2000;118:829–34. [DOI] [PubMed] [Google Scholar]
- 4.Schmeler KM, Lynch HT, Chen LM, et al. Prophylactic surgery to reduce the risk of gynecologic cancers in the Lynch syndrome. N Engl J Med 2006;354:261–9. [DOI] [PubMed] [Google Scholar]
- 5.Burn J, Gerdes AM, Macrae F, et al. Long-term effect of aspirin on cancer risk in carriers of hereditary colorectal cancer: an analysis from the CAPP2 randomised controlled trial. Lancet 2011;378:2081–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Jim MA, Pinheiro PS, Carreira H, et al. Stomach cancer survival in the United States by race and stage (2001–2009): Findings from the CONCORD-2 study. Cancer 2017;123 Suppl 24:4994–5013. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Win AK, Young JP, Lindor NM, et al. Colorectal and other cancer risks for carriers and noncarriers from families with a DNA mismatch repair gene mutation: a prospective cohort study. J Clin Oncol 2012;30:958–64. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Syngal S, Brand RE, Church JM, et al. ACG clinical guideline: Genetic testing and management of hereditary gastrointestinal cancer syndromes. Am J Gastroenterol 2015;110:223–62; quiz 263. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Giardiello FM, Allen JI, Axilbund JE, et al. Guidelines on genetic evaluation and management of Lynch syndrome: a consensus statement by the US Multi-Society Task Force on colorectal cancer. Gastroenterology 2014;147:502–26. [DOI] [PubMed] [Google Scholar]
- 10.NCCN Clinical Practice Guidelines in Oncology. Genetic/Familial High-Risk Assessment: Colorectal. Version 1.2018. http://www.nccn.org/professionals/physician_gls/pdf/genetics_colon.pdf. Accessed February 7, 2019.
- 11.Balmaña J, Balaguer F, Cervantes A, et al. Familial risk-colorectal cancer: ESMO Clinical Practice Guidelines. Ann Oncol 2013;24 Suppl 6:vi73–80. [DOI] [PubMed] [Google Scholar]
- 12.Stoffel EM, Mangu PB, Gruber SB, et al. Hereditary colorectal cancer syndromes: American Society of Clinical Oncology Clinical Practice Guideline endorsement of the familial risk-colorectal cancer: European Society for Medical Oncology Clinical Practice Guidelines. J Clin Oncol 2015;33:209–17. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Vasen HF, Blanco I, Aktan-Collan K, et al. Revised guidelines for the clinical management of Lynch syndrome (HNPCC): recommendations by a group of European experts. Gut 2013;62:812–23. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Vangala DB, Cauchin E, Balmana J, et al. Screening and surveillance in hereditary gastrointestinal cancers: Recommendations from the European Society of Digestive Oncology (ESDO) expert discussion at the 20th European Society for Medical Oncology (ESMO)/World Congress on Gastrointestinal Cancer, Barcelona, June 2018. Eur J Cancer 2018;104:91–103. [DOI] [PubMed] [Google Scholar]
- 15.Kastrinos F, Uno H, Ukaegbu C, et al. Development and Validation of the PREMM5 Model for Comprehensive Risk Assessment of Lynch Syndrome. J Clin Oncol 2017;35:2165–2172. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Warthin AS. Heredity with reference to carcinoma. Arch Intern Med 1913;12:546–55. [Google Scholar]
- 17.Lynch HT, Krush AJ. Cancer family “G” revisited: 1895–1970. Cancer 1971;27:1505–11. [DOI] [PubMed] [Google Scholar]
- 18.Capelle LG, Van Grieken NC, Lingsma HF, et al. Risk and epidemiological time trends of gastric cancer in Lynch syndrome carriers in the Netherlands. Gastroenterology 2010;138:487–92. [DOI] [PubMed] [Google Scholar]
- 19.Engel C, Loeffler M, Steinke V, et al. Risks of less common cancers in proven mutation carriers with lynch syndrome. J Clin Oncol 2012;30:4409–15. [DOI] [PubMed] [Google Scholar]
- 20.Bonadona V, Bonaiti B, 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] [PubMed] [Google Scholar]
- 21.Soer EC, Leicher LW, Langers AM, et al. Equivalent Helicobacter pylori infection rates in Lynch syndrome mutation carriers with and without a first-degree relative with gastric cancer. Int J Colorectal Dis 2016;31:693–7. [DOI] [PubMed] [Google Scholar]
- 22.Vasen HFA, Blanco I, Aktan-Collan K, et al. Revised guidelines for the clinical management of Lynch syndrome (HNPCC): recommendations by a group of European experts. Gut 2013;62:812–823. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Renkonen-Sinisalo L, Sipponen P, Aarnio M, et al. No support for endoscopic surveillance for gastric cancer in hereditary non-polyposis colorectal cancer. Scand J Gastroenterol 2002;37:574–7. [DOI] [PubMed] [Google Scholar]
- 24.Galiatsatos P, Labos C, Jeanjean M, et al. Low yield of gastroscopy in patients with Lynch syndrome. Turk J Gastroenterol 2017;28:434–438. [DOI] [PubMed] [Google Scholar]
- 25.Ladigan S, Vangala DB, Kuhlkamp J, et al. Value of EGD for gastric cancer surveillance in patients with hereditary non-polyposis colorectal cancer (HNPCC) or Lynch syndrome (LS). J Clin Oncol 2018;36: (suppl; abstr 1522). [Google Scholar]
- 26.Childers KK, Maggard-Gibbons M, Macinko J, et al. National Distribution of Cancer Genetic Testing in the United States: Evidence for a Gender Disparity in Hereditary Breast and Ovarian Cancer. JAMA Oncol 2018;4:876–879. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.LaDuca H, McFarland R, Gutierrez S, et al. Quality of Clinician-Reported Cancer History When Ordering Genetic Testing. JCO Clin Cancer Inform 2018;2:1–11. [DOI] [PubMed] [Google Scholar]
- 28.ten Broeke SW, van der Klift HM, Tops CMJ, et al. Cancer Risks for PMS2-Associated Lynch Syndrome. J Clin Oncol 2018;36:2961–2968. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Karimi P, Islami F, Anandasabapathy S, et al. Gastric cancer: descriptive epidemiology, risk factors, screening, and prevention. Cancer Epidemiol Biomarkers Prev 2014;23:700–13. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.