Polyposis in Lynch syndrome: A retrospective study

Abstract

BACKGROUND

National Comprehensive Cancer Network guidelines recommend patients with ≥ 10 lifetime adenomas undergo genetic evaluation for polyposis syndromes. Lynch syndrome (LS) has not historically been considered a polyposis syndrome. We utilized our hereditary gastrointestinal tumor registry to determine the number of lifetime adenomas in LS patients, and to assess differences in demographic and clinical factors between those with < 10 adenomas and those with ≥ 10.

AIM

To describe a cohort of patients with LS who have developed polyposis that may be used to improve endoscopic surveillance guidelines in this population.

METHODS

Medical records from LS patients enrolled in our registry since 2005 (n = 260) were reviewed for colonoscopy outcomes, colorectal cancer diagnosis, and other clinical factors. Groups of interest were compared using Fisher’s exact test and Wilcoxon-Mann-Whitney test.

RESULTS

Three patients were excluded, leaving 257 individuals for data analysis. Most patients were female and white. Mean follow-up time was 6.8 years. Eleven (4.3%) patients had ≥ 10 adenomas. Number of colonoscopies and follow-up time were significantly higher in patients with ≥ 10 adenomas compared with those diagnosed with < 10. For assessment of differences between the two groups, we limited our analysis to include those with ≥ 6 lifetime colonoscopies (n = 92). No significant differences in the evaluated characteristics were observed between those with ≥ 10 adenomas and those with < 10. MSH6 and MSH2 were the most commonly affected genes in the ≥ 10 adenomas group, accounting for 45.5% each, together representing 90.9% (10/11) of the polyposis cohort, compared to 24.7% and 35.8%, respectively, in the < 10 adenomas group.

CONCLUSION

Polyposis occurs in LS; possibly, more frequently among those with MSH6 or MSH2 pathogenic variants. LS should be included as a differential for attenuated polyposis.

Key Words: Colorectal cancer; Polyposis; Lynch syndrome; Hereditary colon cancer; Genetics

Core Tip: The purpose of this study was to describe and investigate polyposis (≥ 10 lifetime adenomas) in patients with a diagnosis of Lynch syndrome (LS). Previously known as hereditary nonpolyposis syndrome, more data have emerged describing polyposis in these patients. We used a registry of patients with LS at our institution to find possible associations and commonalities among patients who had ≥ 10 lifetime adenomas to help determine which of these high-risk patients may be at higher risk for developing adenomas. We discovered that patients with polyposis more often had pathogenic variants in the MSH2 and MSH6 genes.

INTRODUCTION

Lynch syndrome (LS) is the most common cause of inherited colorectal cancer (CRC) and is responsible for 3% of all CRC cases in the USA. It is an autosomal dominant condition defined by the presence of a germline pathogenic variant (PV) in a DNA mismatch repair gene (MLH1, MSH2, MSH6, and PMS2) or a deletion that involves the 3’ end of EPCAM[1-3]. LS was previously known as hereditary nonpolyposis colorectal cancer (HNPCC) due to the low number of adenomas found in these patients, and historical diagnosis criteria for LS, the Amsterdam I and II criteria, required the exclusion of familial adenomatous polyposis[4]. Since the term HNPCC was applied based on clinical diagnostic criteria, this nomenclature has been replaced by LS for families who have a germline PV identified through genetic testing[5,6].

In recent years, reports of individuals with LS who have > 10 lifetime adenomas have emerged challenging the prior label of HNPCC. Kalady et al[7] found 4% of patients with LS had ≥ 10 lifetime adenomas. Jain et al[8] reported multiple colorectal adenomas in patients with LS and concluded that regardless of genotype, multiple colorectal adenomas is a high-risk phenotype for the development of CRC among those with LS. Roberts et al[9] described a cohort with a higher proportion (16.8%) of LS patients with ≥ 10 adenomatous polyps as compared to previous studies. These findings challenge the existing notion that all patients with LS only develop a few or no polyps in their lifetime.

Although additional mechanisms of CRC development may be at play for individuals with LS, a higher polyp burden could in theory result in a higher CRC risk. Determining factors associated with polyp formation in the subset of patients with polyposis may provide valuable information when determining appropriate risk management, including colonoscopy surveillance intervals.

In this study, we utilized colonoscopy and pathology records from patients with LS included in our hereditary colorectal and associated tumor study to determine how many patients developed ≥ 10 lifetime adenomatous polyps and to describe characteristics of those patients to better understand factors associated with a polyposis phenotype in patients with LS.

MATERIALS AND METHODS

Study design

This was a single site retrospective study of number of lifetime adenomas in LS patients. The study was conducted at the University of Pittsburgh Medical Center through the Department of Gastroenterology and Hepatology, specifically within the Hereditary Gastrointestinal Tumor Program. Patients with LS due to a known germline PV in MLH1, MSH2, MSH6, or PMS2, or a deletion in EPCAM who were enrolled into our Hereditary Colorectal and Associated Tumor Study (MOD19050027-026) between January 1, 2005, and December 31, 2022, were identified. Patients were included in the analyses if they had undergone at least one colonoscopy.

Information was gathered via chart review of the electronic medical record (via EPIC, PowerChart, Care Everywhere, and outside scanned reports from other hospitals). Colonoscopy data in the past and during the study period were abstracted from individual colonoscopy and pathology reports. We collected data on the following variables: Date of colonoscopy; age at the time of colonoscopy; number of polyps identified; size of polyps; location of polyps; and histological type of polyps [hyperplastic polyp, sessile serrated adenoma (SSA), tubular adenoma (TA), tubulovillous adenoma (TVA), or adenocarcinoma]. We chose to include SSAs in our total polyp count as we intended to include all polyps with malignant potential. Hyperplastic polyps were not included in polyp count. Demographic data, other relevant medical history, including personal and family history of CRC, and follow-up/survival data were abstracted. For this study, the term adenoma referred to SSA, TA, TVA, and adenocarcinoma. Polyposis was defined as ≥ 10 lifetime adenomas.

Statistical analysis

A descriptive analysis was performed to highlight demographic and clinical characteristics of the patients in the cohort. We reported frequency and percentage as categorical variables, and mean ± SD and range for continuous variables. We compared groups of interest using Fisher’s exact test and Wilcoxon-Mann-Whitney test. P < 0.05 was considered significant. Statistical analyses were performed using SAS (version 9.4, SAS Institute Inc., Cary, NC, USA).

RESULTS

Cohort characteristics

Of the 260 LS patients in the hereditary colorectal and associated tumor study, three were excluded due to lack of colonoscopies. Characteristics of the remaining 257 patients are provided in Table 1. Most patients were female (65.4%) and white (34.6%). Forty patients (15.6%) were smokers at the time of enrollment. Most patients reported having a first- and/or second-degree relative with CRC, and the most common causative gene in our overall study cohort was MSH2 followed by MSH6 (35.0% and 24.9%, respectively). Eighty-eight (34.2%) patients had a diagnosis of CRC; 84 prior to enrollment in our study and four after. The number of colonoscopies ranged from one to 23 (mean: 4.7), and number of adenomas detected ranged from 0 to 44 (mean 2.4). Mean follow-up time was 6.8 years. Eleven (4.3%) of the patients had been diagnosed with ≥ 10 adenomas.

Table 1 Characteristics of the study population.

	Total cohort (n = 257)
Age at first colonoscopy, year, mean ± SD (range)	45.5 ± 14.3 (19-87)
Sex
Male	89 (34.6)
Female	168 (65.4)
Race
White	245 (95.3)
Black	7 (2.7)
Hispanic	1 (0.4)
Asian	3 (1.2)
Native American	1 (0.4)
Current smoker	40 (15.6)
BMI in kg/m2, mean ± SD (range)	28.8 ± 6.4 (17.4-53.9)
First-degree relative with CRC, yes	134 (52.1)
Second-degree relative with CRC, yes	148 (58.0)
Medication use
Aspirin use (yes)	50 (19.5)
Statin use (yes)	32 (12.5)
NSAID use (yes)	43 (16.7)
Affected gene
EPCAM	2 (0.8)
MLH1	51 (19.8)
MSH2	90 (35.0)
MSH6	64 (24.9)
PMS2	50 (19.5)
Diagnosis of CRC, yes	88 (34.2)
Age at CRC diagnosis in years, mean ± SD (range)	49.6 ± 13.5 (24-87)
Colectomy
Partial	50 (19.5)
Sub-total/total	28 (10.9)
None	179 (69.6)
Age at first polyp in years, mean ± SD (range)	49.2 ± 13.4 (22-87)
Number of colonoscopies, mean ± SD (range)	4.7 ± 3.5 (1-23)
Number of adenomas, mean ± SD (range)	2.4 ± 4.6 (0-44)
Diagnosed with ≥ 10 adenomas, yes	11 (4.3)
Time followed in years, mean ± SD (range)	6.8 ± 6.7 (0-36)

BMI: Body mass index; CRC: Colorectal cancer; NSAID: Non-steroidal anti-inflammatory drug.

Characteristics of those with ≥ 10 adenomas

Detailed characteristics of the 11 patients with ≥ 10 adenomas are provided in Table 2. Five patients had a CRC diagnosis. Patient 1 was diagnosed with CRC at age 50 years, prompting total colectomy. He had 44 lifetime adenomas that occurred prior to this surgery with no further adenomas detected following colectomy. Patient 2 had 34 TAs found on screening colonoscopies following CRC diagnosis and surgery. Not all records prior to her enrollment in the hereditary colorectal and associated tumor study were obtainable, but she had a minimum of 39 lifetime adenomas. Patient 6 had eight TAs found on colonoscopies prior to diagnosis of adenocarcinoma. However, colonoscopy 1 year prior to diagnosis of CRC was normal. The cancer was completely removed through polypectomy, so the patient did not have surgery. One year after this colonoscopy, he had two 1-mm TAs and two additional TAs discovered on subsequent annual colonoscopies. Patient 10 had four TAs and adenocarcinoma on initial screening colonoscopy. He underwent proctectomy at that time. During subsequent colonoscopies, he was found to have five additional TAs. Patient 11 had two TAs and adenocarcinoma during colonoscopy. He underwent transverse colectomy and had seven additional TAs found on subsequent colonoscopies.

Table 2 Characteristics of patients with ≥ 10 adenomas.

Patient	Age at first colonoscopy (year)	Age of LS diagnosis	Sex	Race	PV	Total lifetime adenomas	Diagnosis of CRC, age (year)	Surgery, age (year)	First-degree relative with CRC	Second-degree relative with CRC
1	26	31	Male	White	MSH2	44	Yes, 50	Total colectomy, 50	Yes	Yes
2	45	54	Female	White	MSH6	39	Yes, 41	Sigmoidectomy, 41	No	No
3	64	66	Female	White	MSH6	25	No	No	No	Yes
4	59	61	Female	White	MSH6	19	No	No	No	No
5	40	40	Male	White	MSH6	15	No	No	No	No
6	47	57	Male	White	MSH2	13	Yes, 55	No	Yes	Yes
7	44	50	Male	White	MSH6	10	No	No	Yes	Yes
8	51	53	Female	White	MSH2	10	No	No	Yes	Yes
9	50	62	Female	White	MSH2	10	No	No	Yes	Yes
10	34	34	Male	White	MSH2	10	Yes, 34	Proctectomy, 34	Yes	No
11	74	88	Male	Black	PMS2	10	Yes, 80	Transverse colectomy, 80	Yes	No

CRC: Colorectal cancer; PV: Pathogenic variant.

Comparison of individuals with < 10 and ≥ 10 adenomas

Number of colonoscopies and follow-up time were significantly higher in the ≥ 10 adenomas group compared to those diagnosed with < 10 adenomas (P < 0.0001 and P = 0.0003, respectively). Among those diagnosed with ≥ 10 adenomas, the minimum number of colonoscopies was six and the minimum follow-up was 4.8 years. Therefore, we limited our < 10 adenomas cohort to those with ≥ 6 lifetime colonoscopies for assessment of differences. This resulted in a dataset with 92 patients: 11 with ≥ 10 adenomas and 81 with < 10 adenomas. Except for higher numbers of colonoscopies and number of adenomas detected (P = 0.01 and P < 0.0001), no significant differences in the evaluated characteristics were observed between those with ≥ 10 adenomas and those with < 10 (Table 3).

Table 3 Participants with ≥ 6 colonoscopies, stratified by number of adenomas (< 10, ≥ 10).

	Participants with ≥ 6 colonoscopies (n = 92)
	< 10 lifetime adenomas (n = 81)	≥ 10 lifetime adenomas (n = 11)	P value
Age at first colonoscopy, year, mean ± SD (range)	47.1 ± 12.5 (20-72)	48.9 ± 13.6 (26-75)	0.51
Sex			0.33
Male	25 (30.9)	6 (54.5)
Female	56 (69.1)	5 (45.5)
Race			0.70
White	73 (90.1)	10 (90.9)
Black	5 (6.2)	1 (9.1)
Hispanic	1 (1.2)	0 (0.0)
Asian	1 (1.2)	0 (0.0)
Native American	1 (1.2)	0 (0.0)
Current smoker	9 (11.3)	1 (9.1)	1.00
BMI in kg/m2, mean ± SD (range)	27.9 ± 5.9 (18-50)	29.7 ± 6.2 (18.7-38.4)	0.38
First-degree relative with CRC, yes	45 (55.6)	7 (63.6)	0.75
Second-degree relative with CRC, yes	47 (58.8)	6 (54.6)	1.00
Medication use
Aspirin use (yes)	18 (22.2)	3 (27.3)	0.71
Statin use (yes)	7 (8.6)	2 (18.2)	0.29
NSAID use (yes)	11 (13.6)	2 (18.2)	0.65
Affected gene			0.28
EPCAM	1 (1.2)	0 (0.0)
MLH1	18 (22.2)	0 (0.0)
MSH2	29 (35.8)	5 (45.5)
MSH6	20 (24.7)	5 (45.5)
PMS2	13 (16.1)	1 (9.1)
Diagnosis of CRC, yes	43 (53.1)	5 (45.5)	0.75
Age at CRC diagnosis in years, mean ± SD (range)	48.6 ± 12.0 (24-70)	52.6 ± 17.5 (35-80)	0.80
Colectomy			1.00
Partial	25 (30.9)	2 (18.2)
Sub-total/total	12 (14.8)	1 (9.1)
None	44 (54.3)	8 (72.7)
Age at first polyp in years, mean ± SD (range)	49.7 ± 11.9 (22-70)	50.7 ± 14.9 (27-80)	0.87
Number of colonoscopies, mean ± SD (range)	8.1 ± 2.3 (6-16)	11.2 ± 4.7 (6-23)	0.01
Number of adenomas, mean ± SD (range)	2.7 ± 2.2 (0-9)	18.7 ± 12.3 (10-44)	< 0.0001
Time followed in years, mean ± SD (range)	12.2 ± 6.7 (4-36)	13.1 ± 5.4 (5-26)	0.23

CRC: Colorectal cancer; NSAID: Non-steroidal anti-inflammatory drug.

DISCUSSION

LS has traditionally been defined as HNPCC with the development of < 10 lifetime adenomas. It is typically differentiated from other hereditary CRC syndromes such as familial adenomatous polyposis/attenuated familial adenomatous polyposis based on a low adenoma burden. The definition of HNPCC is being challenged with reports of LS patients with ≥ 10 lifetime adenomas. Compared with other studies on polyposis in LS, 4.3% of the patients in our LS cohort had polyposis, with ≥ 10 lifetime adenomas[7,8]. Three patients had > 20 lifetime adenomas, with 44 adenomas as the highest count.

To identify common characteristics of LS patients with polyposis, we reviewed demographic and other background information in our cohort (Table 1). There were no significant differences in sex, race, medications (aspirin, statin, nonsteroidal anti-inflammatory drugs), smoking history, and family history between the polyposis and nonpolyposis groups (Table 3). The demographics and findings are consistent with similar previous studies[8]. None of the 11 patients with polyposis were related to each other. With no obvious risk factors related to polyposis, it is difficult to identify which patients may be at higher risk for the development of polyposis, and this likely cannot be identified until completion of multiple screening colonoscopies. The only notable difference between the two groups was the affected LS gene. Prior studies indicated that patients with a PV in MLH1 or MSH2 were at increased risk of CRC[10]. It has also been widely concluded that MSH6 and PMS2 PVs are associated with markedly lower LS-related cancer risk than other germline PVs are[10]. Although not significant, it is noteworthy that 10 of 11 patients in the polyposis cohort had a MSH2 or MSH6 PV. This suggests that one could consider CRC surveillance closer to the 1-year interval in patients with MSH2 or MSH6 mutations, while also taking into consideration personal polyp and family history.

A potential explanation for the observed association of MSH6 and MSH2 with the development of polyposis may be found at the molecular level. Prior literature indicates that PVs in the APC gene, either germline or acquired as the result of mismatch repair (MMR) deficiency, are associated with development of adenomatous polyps[11,12]. It has recently been shown that substitution or insertion/deletion mutations at repetitive sequences in the APC gene are more frequent in LS-associated CRC than in microsatellite-stable CRC[12,13]. These simple mismatches are usually recognized by MutSα, an MSH2 and MSH6 heterodimer, as part of the MMR machinery, so the error can be corrected via recruitment of MutLα, subsequent removal of erroneous DNA strand, and synthesis of the correct one[14]. By disrupting the interface between their mismatch binding domains, PVs in MSH2 or MSH6 reduce the DNA binding affinity of MutSα at the APC mismatch error site, preventing its ability to recognize the error and help correct it[14]. Resulting APC-driven polyp formation in these individuals could potentially help explain our observations. However, further studies are required to establish the validity of this hypothesis.

There were limitations to our study. First, given that we are a tertiary referral center, there may be missing colonoscopies from outside facilities that were not obtained. This was mitigated by having our team request past records that were uploaded into our system to capture all colonoscopy reports. Second, not all patients received multigene panel testing. Many of our patients received only single-site or single gene testing based on known family history or directed by immunohistochemistry results, so we cannot exclude the possibility that some patients in the polyposis group had PVs in genes associated with a polyposis phenotype. This study was subject to ascertainment bias as genetic testing results were used as an inclusion criterion. The size of the polyposis group (n = 11) was small and may have compromised statistical power. Lastly, this was a single site study and only describes the characteristics of patients in this region, however, our site does also serve as a large tertiary center that cares for patients across multiple states. A multicenter study with a larger sample size would improve and validate the findings of our study.

CONCLUSION

This study described a cohort of patients with LS; 11 of whom had ≥ 10 lifetime polyps. No clear risk factors, including smoking status, medication use, sex, race, or family history of CRC, could be identified in the polyposis group. However, MSH2 and MSH6 PV were more likely to be associated with polyposis in LS than MLH1 or PMS2, although this did not reach statistical significance (P = 0.28). This suggests that there are no clear modifiable risk factors that can be identified to prevent polyposis, and that the development of polyposis is driven by PV. We propose a mechanism to explain this finding that lies within the heterodimer nature of MSH2 and MSH6. Our findings suggest closer surveillance for MSH2 and MSH6 carriers initially to establish polyp burden, rather than initially starting at a longer interval (3 years). Future research will be aimed at further examining the molecular mechanism for the development of CRC and polyposis in individuals with MSH2 and MSH6 PVs.