MSH2 Gene and EPCAM: Lynch Syndrome Cancer Risk Explained

01 / 11

What is the MSH2 gene?

The MSH2 gene, on chromosome 2 at 2p21-p16.3, makes a protein that spots copying mistakes in DNA. MSH2 pairs with MSH6 to form the mismatch-recognition complex MutS alpha. A harmful change in one copy of MSH2 causes Lynch syndrome, and MSH2 variants account for 20–40% of cases. ^[1,12]

MSH2 also pairs with MSH3 to form MutS beta. NCBI lists the aliases FCC1, COCA1, HNPCC1, LYNCH1, hMSH2 and MMRCS2. MedlinePlus describes MSH2 as causing up to 40% of Lynch syndrome. About 1 in 4,000 people in the United States carries an MSH2 variant, while Lynch syndrome overall affects about 1 in 280–600. ^[1,4,10,12]

02 / 11

How do MSH2 variants cause Lynch syndrome?

MSH2 variants cause Lynch syndrome by switching off one working copy of a DNA repair gene. The cell keeps one normal copy and repairs DNA until a later, second hit disables it; copying errors then accumulate in that tissue and tumours show microsatellite instability. MSH2 variants include missense, truncating, splice and large-deletion changes. ^[1]

This is a loss-of-function mechanism. ClinGen classified the MSH2–Lynch syndrome relationship as Definitive on June 20, 2022. Pathogenicity is assessed with ACMG/AMP principles and the ClinGen MSH2 Variant Curation Expert Panel specification. ^[1,8,13]

03 / 11

EPCAM deletions and MSH2

An EPCAM deletion can cause Lynch syndrome without any change in MSH2 itself. EPCAM sits next to MSH2 on chromosome 2; a deletion at the 3′ end of EPCAM lets transcription read through into MSH2, and the MSH2 promoter becomes methylated and silent in cells that make EPCAM. Only large deletions that include EPCAM's last exon do this. ^[1,7,11]

Deletion pattern	Reported risk	What is known
Extends into MSH2	Colorectal 33–52%; endometrial 21–57%; ovarian 8–38%	Risks and screening are treated like an MSH2 pathogenic variant
Does not extend into MSH2	Mainly colorectal: 75%; endometrial 12–25%	Very rare; NCCN gives no separate screening schedule

Early reports suggested low extracolonic risk, but later work showed that deletions extending into MSH2 can carry broader MSH2-like risk. EPCAM is not a mismatch-repair gene. GeneReviews attributes under 10% of Lynch syndrome to EPCAM deletions, while MedlinePlus says up to 3%; the sources use different estimates. ^[1,5,7,11]

EPCAM deletions are found by deletion/duplication analysis, not sequence analysis. GeneReviews reports 100% of known EPCAM Lynch-causing variants are found this way. Variants in both EPCAM copies cause the separate recessive condition congenital tufting enteropathy. ClinGen classifies the EPCAM–Lynch relationship as Definitive. Loss of EPCAM protein in tumour tissue can be a clue; read about tumour MSI and IHC testing. ^[1,9,11,15]

04 / 11

MSH2 cancer risk by cancer type, age and sex

People with a pathogenic MSH2 variant have a high risk of colorectal cancer and the widest spread of other Lynch cancers. NCCN Version 1.2026 gives 33–52% lifetime colorectal and 21–57% endometrial risk. In the Prospective Lynch Syndrome Database, colon cancer risk by age 65 was 30% in women and 41.5% in men. ^[1,2,4]

Cumulative risk through age 65

General population and MSH2 carrier cumulative cancer risk through age 65.
Cancer	General population F / M	MSH2 carriers F / M
Colon	0.45% / 0.6%	30% / 41.5%
Rectum	0.3% / 0.5%	8% / 13%
Endometrium	0.6%	38%
Ovary	0.5%	11%
Ureter, kidney	0.2% / 0.35%	10% / 11.5%
Prostate	1% (M)	11% (M)
Small bowel	Unknown	3% / 4.5%
Bladder	0.1% / 0.35%	5% / 6%
Gastric	0.3% / 0.6%	3% / 4%
Brain	0.2% / 0.25%	1% / 3%
Bile duct	Unknown	1% / 1%
Pancreas	0.2% / 0.25%	1% / 1%

Source: GeneReviews Table 3, updated September 17, 2026, using PLSD data through October 2022 from 8,500 carriers and 71,713 follow-up years. ^[1]

NCCN lifetime ranges

Cancer	MSH2 lifetime risk	General population	Average age: MSH2 / general
Colorectal	33–52%	4%	44 / 68–72
Endometrial	21–57%	3.1%	47–48 / 60
Ovarian	8–38%	1.1%	43 / 63
Kidney / ureter	2.2–28%	—	—
Bladder	4.4–12.8%	—	—
Gastric	Up to 9%	—	—
Small bowel	1.1–10%	—	—
Pancreatic	2–5%	—	—
Prostate	Up to 24%	12.8%	59–63 / 66
Brain	2.5–7.7%	—	—
Biliary	0.02–1.7%	—	—
Sarcoma	4.2%	0.1%	—

NCCN Versions 1.2026 and 1.2027 as reproduced by FORCE on September 24, 2026. ^[2,3,4]

Why the tables differ: NCCN ranges are lifetime estimates pooled from several studies. PLSD follows carriers prospectively only through age 65 while they receive surveillance. The figures therefore cannot be compared line by line. Not everyone with MSH2 develops cancer because penetrance is below 100%. ^[1,2,6]

PLSD reported median age at any-organ cancer of 53.6 years in men and 50.6 in women; median onset was 51.7 for endometrial, 47.4 for ovarian and 53.9 in men and 56.2 in women for colorectal cancer. Under colonoscopy surveillance, particularly for MSH2 carriers, more deaths followed non-colorectal Lynch cancers than colorectal cancers. ^[6]

Prostate estimates use different age cut-offs and designs: Myriad gives 13–17%, eviQ 15%, NCCN up to 24%, GeneReviews 11% by 65, and PLSD reported 39.7% cumulative incidence by 75 in MSH2 men. These are not interchangeable estimates. ^{[1,2,6,16,17]}

05 / 11

How MSH2 differs from the other Lynch genes

MSH2 variants carry the widest range of non-colorectal cancers among the four Lynch genes, including urinary tract, ovarian and prostate cancers. By age 65, PLSD estimates colon cancer at 30% in women and 41.5% in men for MSH2, 36% and 48% for MLH1, 10% and 13% for MSH6, and 3% and 9.5% for PMS2. ^[1]

Cancer through 65	MSH2	MLH1	MSH6	PMS2
Colon F / M	30% / 41.5%	36% / 48%	10% / 13%	3% / 9.5%
Endometrium	38%	32%	32%	13%
Ovary	11%	8%	3%	2.5%
Ureter / kidney F / M	10% / 11.5%	2% / 2.5%	3% / 1%	Low / uncertain

PMS2 estimates rest on only 549 carriers and are less certain. Compare the MLH1 gene, MSH6 gene and PMS2 gene. Muir-Torre syndrome, with sebaceous skin tumours, is an MSH2-associated presentation; NCCN-based guidance recommends a skin examination every 1–2 years. ^[1,4,10]

06 / 11

What does “MSH2 positive” mean?

On a genetic test report, “MSH2 positive” means a pathogenic or likely pathogenic variant was found in MSH2, which establishes Lynch syndrome. On a tumour pathology report, “MSH2 positive” usually means the MSH2 protein is present on staining, the normal result; loss of staining is abnormal. A variant of uncertain significance neither confirms nor rules out Lynch syndrome. ^[1,13]

Report context	What the result means	What happens next
Germline “positive”	A pathogenic or likely pathogenic MSH2 variant was identified	Clinical confirmation, genetic counselling, surveillance and family testing; it is not a cancer diagnosis
Tumour IHC “positive”	MSH2 protein is retained or present—the normal staining result	Interpret with MSH6 and the other mismatch-repair proteins
Tumour loss of MSH2/MSH6	Abnormal absent staining can suggest MSH2-pathway dysfunction	Germline testing is considered; tumour staining alone does not diagnose Lynch syndrome
Variant of uncertain significance	Evidence cannot show whether the variant is harmful	Do not use it alone to diagnose Lynch syndrome or make irreversible decisions

Pathogenic and likely pathogenic variants are managed alike under ACMG/AMP principles. Read the separate guide to tumour MSI and IHC testing. ^[1,13]

07 / 11

How MSH2 variants run in families

An MSH2 variant is inherited in an autosomal dominant pattern: each child of a carrier has a 50% chance of inheriting it, and it does not skip generations. Relatives can be tested for the one known family variant. A child who inherits an MSH2 variant from both parents has a separate, far more severe condition. ^[1,4]

50%

Each pregnancy is independent. Cascade testing starts with the exact family variant. EPCAM deletions that cause Lynch syndrome follow the same autosomal dominant inheritance pattern.

Two pathogenic mismatch-repair variants can cause constitutional mismatch repair deficiency (CMMRD). The first tumour usually occurs before 18; about half of affected children develop cancer by 10 and about 90% by 18. ClinGen classifies MSH2–CMMRD as Definitive and adult Lynch actionability as Definitive, but paediatric actionability for monoallelic Lynch syndrome has insufficient evidence. Children are therefore not routinely tested for adult-onset Lynch syndrome. ^[1,8,10]

08 / 11

Screening and prevention for MSH2 carriers

For people with an MSH2 variant, NCCN Version 1.2026 advises colonoscopy every 1–2 years starting at age 20–25, or 2–5 years before the youngest colorectal cancer in the family. GeneReviews gives the same schedule for MSH2, MLH1 and EPCAM. Plans are individual and set with a genetic counsellor or clinician. ^[1,2,4]

Area	When to discuss or start	Interval / note
Colonoscopy	Age 20–25, or 2–5 years before the youngest family diagnosis	Every 1–2 years
Endometrial biopsy	Discuss from age 30–35	Every 1–2 years
Transvaginal ultrasound	After menopause	Discuss; not a reliable stand-alone screen
Hysterectomy with salpingectomy	Discuss from age 40	Individual decision after childbearing
Ovarian risk-reducing surgery	After childbearing	Individual decision
Upper endoscopy	Age 30–40	Every 2–4 years
Urinary-tract screening	Discuss from age 30–35 if relatives are affected	Evidence is limited
PSA conversation	Age 40	Discuss benefits and harms; reassess at 75
Pancreatic MRCP / EUS	Age 50 only with relevant family history	Annually in an experienced centre
Skin examination	Individual review	Every 1–2 years
Aspirin	Discuss with a clinician	Dose, duration and bleeding risk require individual review

EPCAM carriers whose deletions involve MSH2 follow the MSH2 schedule. eviQ, last reviewed April 16, 2024 and overdue for review since December 31, 2025, starts colonoscopy at 25 and recommends no prostate screening; this differs from NCCN's PSA discussion. Read the detailed colonoscopy schedule and aspirin guide and endometrial and ovarian cancer in Lynch syndrome. ^[1,4,5,16]

09 / 11

How MSH2 variants are found, and what testing can miss

MSH2 variants are found by DNA sequencing plus a separate deletion and duplication analysis, usually within a multigene panel. Sequencing finds about 60–80% of MSH2 pathogenic variants and deletion/duplication analysis the other 20–40%. Rarer structural changes can slip through: a 2025 case report described an MSH2 exon 1–7 inversion missed by a Lynch-specific panel. A tumour test is a different test. ^[1,14]

MSH2 sequencing

60–80%

Share found by sequence analysis in GeneReviews Table 1.

MSH2 del / dup

20–40%

Requires copy-number analysis in addition to sequencing.

EPCAM deletions

100% by del / dup

Known Lynch-causing EPCAM variants are large deletions, not sequence variants.

In the Mayo Clinic Proceedings case, a 21-year-old with a known family MSH2 variant had no single-letter, small insertion/deletion or copy-number finding on the panel. The answer was a 10-megabase paracentric inversion involving exons 1–7. This illustrates a limitation; it does not establish the performance of any consumer genome service. Findings require clinical confirmation and counselling. Read how Lynch syndrome genetic testing works. ^[14]

10 / 11

MSH2 questions people ask

Q1What does MSH2 positive mean?

On a germline report, it means a pathogenic or likely pathogenic MSH2 variant was found and establishes Lynch syndrome. On tumour IHC, “positive” generally means MSH2 protein is present. A positive genetic result is not a cancer diagnosis. ^[1,13]

Q2What cancers are associated with MSH2?

MSH2 is associated with colorectal, endometrial, ovarian, urinary-tract, stomach, prostate, small-bowel, brain and sebaceous skin tumours. Risk is not the same for every cancer, and not every carrier develops cancer. ^[1,2,4]

Q3Can a change in the MSH2 protein cause Lynch syndrome?

Yes. A pathogenic variant in MSH2 can prevent its mismatch-repair protein from working and cause Lynch syndrome. A 3′ EPCAM deletion can also silence MSH2 without changing the MSH2 gene itself. ^[1,7]

Q4What is the life expectancy of someone with Lynch syndrome?

There is no single reliable figure. Outlook varies with the gene, cancer type and stage, surveillance and treatment. Regular gene-specific surveillance is intended to prevent cancer or find it earlier. ^[1,6] Read about Lynch syndrome life expectancy.

Q5Does MSH2 raise breast cancer risk?

ClinGen classified the claimed MSH2–hereditary breast carcinoma relationship as Refuted on March 14, 2023. Breast screening should instead follow personal and family history and other established risk factors. ^[8]

Q6Can I get life insurance with Lynch syndrome?

Insurance rules are outside what this medical page can answer and vary by place and policy. A genetic counsellor can help identify appropriate local resources; this is not legal or insurance advice. ^[1] Return to the Lynch syndrome overview.

11 / 11

Why whole-genome sequencing, and where it falls short

A whole-genome sequence reads MSH2 and EPCAM along with the rest of the genome in one $599 test, whereas a Lynch panel reads a fixed list of genes. HLI describes a cancer risk assessment covering BRCA1 and BRCA2 among others; its Lynch-gene report coverage is unconfirmed. A whole-genome result is a starting point: confirmatory clinical testing and a genetic counsellor follow, and a sequence cannot show an existing cancer or tumour MSI/IHC results.

Route	What it can add	Important limit
Clinical Lynch / multigene panel	Clinician-ordered after counselling; includes deletion/duplication analysis and ClinGen-based classification; may be insurance-covered when criteria are met	A fixed gene list, but the better route for a known family variant or strong history; WGS is not a substitute
Tumour-first pathway	MSI and IHC on colorectal or endometrial tumour tissue can identify mismatch-repair loss before germline testing	Not inherited-DNA testing; WGS does not perform tumour testing
23andMe	Standard Health Predisposition reports do not include Lynch genes; the clinician-ordered Total Health exome report lists MLH1, MSH2, MSH6 and PMS2	EPCAM is not listed; it is an exome, not a genome, and the reviewed page showed no price
HLI whole-genome service	One $599 test at approximately 30× average coverage; broad genome-wide data and a cancer risk assessment	Coverage is not guaranteed at every position. HLI has not published performance for EPCAM deletions or MSH2 inversions and must not be assumed to detect them

Short-read WGS may or may not call a particular EPCAM deletion or MSH2 inversion reliably. It does not diagnose Lynch syndrome, show cancer today, replace screening or promise genetic counselling. See what the $599 genome service includes, and compare established BRCA1 and BRCA2 risk. ^[19]

References and review notes

Sources and versions were reviewed October 5, 2026. A named genetic counsellor or clinical geneticist author and a named MD medical reviewer were not supplied, so no author, reviewer, sameAs or reviewedBy field has been invented. Pathology wording for retained MSH2/MSH6 staining requires clinical review before publication. HLI coverage of MLH1, MSH2, MSH6, PMS2 and EPCAM, structural-variant performance, confirmation workflow and counselling inclusion remain unverified. Recheck NCCN and GeneReviews every six months and whenever revised.