Strategies for Identifying Hereditary Nonpolyposis Colon Cancer D. Gareth Evans,a Sheila Walsh,b James Hill,c and Raymond T. McMahond,e Hereditary nonpolyposis colorectal cancer (HNPCC) is the term given to a predisposition syndrome caused by inherited mutations in one of at least five DNA mismatch repair (MMR) genes. Dominant mutations in these genes predispose individuals to a range of cancers in addition to the most frequent, colorectal cancer. Endometrial cancer is the most notable additional malignancy, followed by ovarian, gastric, upper urethelial, and biliary cancers, and gliomas. Recognition of HNPCC is important so that targeted screening can be effected that will reduce the incidence of the main cancers. While such clinical criteria as Amsterdam and modified Amsterdam are reasonably specific, they lack sensitivity. Thus, tumor-related features have been used to improve sensitivity for identifying patients who can be selected for the relatively expensive direct mutation analysis of the various genes. Microsatellite instability (MSI) and loss of antibody staining for the proteins have been widely vaunted but have their own drawbacks. No one approach has received universal acceptance, and therefore adoption of one of perhaps three strategies, including clinicaland laboratory-based approaches, is still appropriate until an easier, quicker, and cheaper approach can be developed. Semin Oncol 34:411-417 © 2007 Elsevier Inc. All rights reserved.
olorectal cancer is the second most common malignancy of both sexes in developed countries and the fourth commonest form of cancer occurring worldwide. Globally, most new colorectal cancers occur in economically developed countries, with the highest incidence in North America followed by Western Europe, and also in developing regions such as eastern Asia. The lowest rates are reported in Africa and Polynesia.1 In 2006 it is estimated there will be 148,610 new cases of colorectal cancer and 55,170 deaths in the United States.2 In the year 2000 there were 16,270 colorectal cancer–related deaths in the United Kingdom, representing 11% of all cancer mortality.3 In terms of trend, between 1971 and 1997 the total number of colorectal cancers rose by 42% in England and Wales,3 but this figure may be exaggerated by aAcademic
Unit of Medical Genetics, Regional Genetics Service, St Mary’s Hospital, Manchester, UK. bDepartment of Surgery, Trafford General Hospital, Manchester, UK. cDepartment of Surgery, Manchester Royal Infirmary, Manchester, UK. dDepartment of Histopathology, Manchester Royal Infirmary, Manchester, UK. eDivision of Regenerative Medicine, University of Manchester, Manchester, UK. Address correspondence to D. Gareth Evans, MD, FRCP, Professor in Medical Genetics, Department of Clinical Genetics, St. Mary’s Hospital (SM2), Hathersage Road, Manchester M13 OJH, UK. E-mail: Gareth. [email protected]
0093-7754/07/$-see front matter © 2007 Elsevier Inc. All rights reserved. doi:10.1053/j.seminoncol.2007.07.001
increased awareness, better diagnostic techniques, and improved case-recording by cancer registries. Nonetheless, the ever-aging population affects the cancer incidence, and colorectal cancer is the most common cancer among women aged over 70 years. The lifetime risk of large bowel malignancy in the UK population is one in 20 for men and around one in 25 for women, with a prevalence for colorectal cancer of 1.9 per 1,000.3 Lifetime risks in the United States are now quoted at one in 17 and one in 18, respectively.2 In recent years our understanding of the cellular and molecular events underlying the development of colorectal cancer has improved immeasurably, but despite this better understanding and advances in surgery, radiotherapy, and chemotherapy, the average 5-year survival rate remains approximately 40%.3 The human and financial costs of this disease have prompted considerable research into its prevention, screening, and treatment. Also, from our understanding of the natural history of the disease so far, polypectomies of adenomas found during endoscopy have been proven to prevent their progression into cancers.4 Considerable efforts have thus been concentrated on the ability of screening tests to detect cancer at an early curable stage, or even at the premalignant polyp stage. One of the most important fundamental concepts in colorectal cancer to emerge in recent years has been 411
Figure 1 Simulated results of MSI/IHC and molecular testing based on population screening studies.
the association between genetic alterations and development of neoplasia.
Sporadic Adenomas and Colorectal Carcinoma: Adenoma–Carcinoma Hypothesis Sporadic colorectal cancers comprise about 80% of all colorectal malignancies.5 In the past several decades, it has become clear that these sporadic cancers originate from benign adenomatous polyps. Adenomas, and eventually frank carcinomas, arise through a stepwise accumulation of somatic mutations, which activate oncogenes and inactivate tumorsuppressor genes within colonic epithelial cells. The net result of these mutations is to confer a growth advantage to the abnormal colonocytes, allowing them to outgrow and predominate over the surrounding cells. This classic multistep model for colorectal carcinogenesis has been termed the “adenoma– carcinoma sequence.” It is the total accumulation of these mutations, rather than the exact order in which they accumulate, that is the main determinant of carcinogenesis.5
Hereditary Nonpolyposis Colorectal Cancer From epidemiological studies, it is evident that approximately 4% to 6% of colorectal cancers are due to high-risk single genes, although about 30% are thought to have a “hereditary element” from twin studies.6 It is likely that much of the hereditary component is polygenic, and many genes conferring lower overall extra risks of colorectal cancer are likely to be identified in the coming years. Nevertheless, the testing for mutations in the known high-risk genes already has important utility in identifying those people who may benefit most from screening and other preventive strategies. The majority of families with a high lifetime risk of developing colorectal cancers will be due to inherited conditions such as familial adenomatous polyposis (FAP) or hereditary nonpolyposis colorectal cancer (HNPCC).6 The diagnosis of FAP is
G.D. Evans et al aided when there are distinguishing signs, such as the multiple adenomas found on endoscopy. Testing for mutations in APC and, if appropriate, MYH is well established in individuals presenting with large numbers of adenomatous polyps. A more difficult situation exists with the diagnosis of HNPCC as patients lack distinctive phenotypic signs but nevertheless have as high lifetime risk of developing cancers.7 The most reliable method of diagnosing HNPCC entails a combination of clinical assessment, mutation testing for mismatch repair (MMR) genes, and the assessment of tumor material for microsatellite instability (MSI) or loss of antibody staining for one of the MMR proteins. HNPCC families have a common autosomal dominant inheritance pattern, which was first recognized as early as 1895 by Warthin.8 Colorectal cancers found in families with many affected members are commonly located in the proximal rather than the distal colon. The age of onset is earlier than with sporadic cancers. The carcinogenesis pathway of colorectal cancer follows the accelerated adenoma– carcinoma sequence. The families with HNPCC also have elevated malignancy risks at other sites such as endometrial, gastric, small bowel, pancreatic, biliary tract, urinary tract, hematological, skin, and laryngeal cancers and brain gliomas. The international collaborative group on HNPCC defined the clinical criteria in 1991. These are called the Amsterdam criteria, which are used as a means of clinically categorizing families as high-risk for HNPCC (Table 1). Within our own population in the Manchester region of the United Kingdom (population 4 million) there are currently 105 families with known MMR gene mutations causing HNPCC with an additional 300 or so families that fulfill one of the Amsterdam criteria, ie, Amsterdam I/II (Table 1). Amsterdam II is an updated version of the initial guidelines, incorporating extra-colonic cancers that are associated with the syndrome.9 Individuals with one or more first-degree relatives affected by colorectal cancer have an increased risk of developing the disease themselves. About 0.4% of patients with colorectal cancer have two affected first-degree relatives.10,11 People from families with HNPCC represent the high end of the risk spectrum, with mutation carriers having up to an 80% lifetime risk of developing the disease.12,13 As the predisposition follows an autosomal dominant pattern, untested first-degree relatives of affected individuals in these families have a 40% lifetime risk of colorectal cancer. Colonoscopy and polypectomy confers a survival benefit in these families.14 Those who fulfill the HNPCC criteria should be referred to a geneticist who will verify records and family history, counsel patients, coordinate mutation screening, and recommend the frequency of screening for at risk individuals.15 While recognition of a need to refer is straightforward when the clinical criteria are met, identification of less typical HNPCC families is more difficult. As approximately 1% to 3% of colorectal cancers are due to inherited MMR mutations16 and only 0.3% to 0.4% of consecutive colorectal cancers fulfill HNPCC criteria,11,16 strategies other then family history have to be ap-
Identifying hereditary nonpolyposis colon cancer
Table 1 Diagnostic Criteria for HNPCC Selection I. Amsterdam Criteria (all criteria must be met) 1. One member diagnosed with colorectal cancer before age 50 years 2. Two affected generations 3. Three affected relatives, one of them a first-degree relative of the other two 4. FAP should be excluded 5. Tumors should be verified by pathologic examination II. Amsterdam Criteria II (all criteria must be met) 1. There should be at least 3 relatives with an HNPCC-associated cancer (colorectal cancer or cancer of the endometrium, small bowel, ureter, or renal pelvis) 2. One should be a first-degree relative of the other two 3. At least two successive generations should be affected 4. At least one should be diagnosed before age 50 years 5. FAP should be excluded in the colorectal cancer cases 6. Tumors should be verified by pathologic examination III. Bethesda Revised Guidelines (meeting features listed under any of the numbered criteria is sufficient) 1. Colorectal carcinoma
plied. These include clinical criteria (the Bethesda criteria) followed by tumor-based testing, using algorithms or computer-based models.
in MSH2 and MLH1, but increased testing of MSH6 is revealing a substantial number of inherited mutations.
Mismatch Repair Genes
Microsatellites and Microsatellite Instability
At least 50% to 60% of families fulfilling HNPCC clinical criteria who are referred to risk assessment units have inherited MMR mutations.17,18 However, this figure drops to as low as 22% to 42% when taken from population-based series (4/1816 and 16/3819). There continues to be debate as to whether the term HNPCC should reflect the clinical entity encompassed by Amsterdam criteria or only refer to families with MMR mutations. It may be that there is a better term to reflect the multisite predisposition from MMR mutations, including reverting to the old term “Lynch syndrome.” For the purposes of this review, HNPCC refers to the inherited predisposition caused by MMR mutations. The MMR genes express proteins, which recognize and repair DNA replication errors. A number of genes have been identified, and mutations in them have been associated with HNPCC. These genes are, in order of decreasing frequency in HNPCC, MLH1/MSH2, MSH6, PMS2, and MSH3. PMS1 may also be involved, but the initial report of germline mutations has not been supported by subsequent observations. The majority of mutations currently associated with HNPCC are
In the presence of mutations in one of the MMR genes, the most susceptible region in the DNA is microsatellites. Microsatellites are regions of DNA repeat sequences up to 40 base pairs in length. These sequences are often in the noncoding region of the DNA, and their function is unknown. As a result of MMR deficiency, the number of repeated sequences will alter (either upward or downward because of lack of repair) in the affected cell, which may develop into cancer. This microsatellite abnormality is termed microsatellite instability (MSI) and can be detected when the same genomic region is compared between tumor and normal DNA in the same person (either cells from normal-looking bowel lining or from blood lymphocytes). Several studies have reported that more than 90% of colorectal cancers in HNPCC patients demonstrate some degree of MSI.20 However, MSI also is present in 12% to 15% of all sporadic colorectal cancers. Most of this is due to nonheritable somatic involvement of MMR genes, especially MLHI promoter methylation and gene silencing. In accordance with the international guidelines for the evaluation of MSI in colorectal cancer, five markers are used:
G.D. Evans et al
414 Table 2 Results of Testing for Germline Mutations in Population-Based Studies Study
No. of Cases
Pinol24 Barnetson19 Evans25 Southey22 Salovaara26 Ponz de Leon27 Katballe28 Percesepe29 Hampel30
Spain/all ages*§ Scotland/<55 yrs England/all ages* Australia/<45 yrs¶ Finland/all ages† Italy/all ages† Denmark/All ages† Italy/all ages‡ USA/all ages¶
1,222 870 1,800 131 535 1,721 1,200 336 1,066
5 (0.4%) 16 (1.8%) 2 (0.1%) 4 (3%) 1 (0.1%) 2 (0.1%) 6 (0.5%) 1 (0.3%) 13 (1.3%)
3 (0.25%) 15 (1.7%) 4 (0.2%) 9 (7%) 17 (3%) 1 (0.05%) 4 (0.3%) 0 5 (0.5%)
Not tested 7 (0.8%) Not tested 4 (3%) Not tested Not tested Not tested Not tested 3 (0.3%)
8/1,222 (0.66%) 38/870 (4.4%) 6/1,808 (0.33%) 17/131 (13%) 18/535 (3.4%) 3/1721 (0.02%) 10/1,200 (0.8%) 1/336 (0.3%) 21/1,066 (2%)
NOTE. The Evans et al report25 has been updated to include analysis of a further 400 patients. *MMR testing was only carried out in Amsterdam-positive families and Bethesda-positive families with MSH, IHC-negative. †MMR testing was only carried out and suspected HNPCC families. ‡MMR testing was only carried out in Amsterdam-positive and cases with MSH or IHC-negative tumors. §Three mutations of unclear pathogenicity were excluded. ¶PMS2 mutations in these two studies are not included.
two mononucleotide (BAT25, BAT26) and three dinucleotide repeats (D2123, D5S346, D17S250). A tumor will be classified as MSI-high (MSH) if two or more markers are unstable, MSI-low (MSL) if one is unstable and microsatellite stable (MSS) if none are unstable. Tumors should be tested for MSI if they meet the modified Bethesda criteria (Table 1), which were designed to increase the detection rate of patients with HNPCC. In most published studies, only a small proportion of colorectal cancers in patients with germline MMR gene mutations will be MSS .21 However, the analysis depends on tumor sections being cut by a trained pathologist and at least 70% of the cells analyzed being tumor cells. In order to be sure that germline testing is unnecessary (⬍10% chance of identifying a mutation) in an Amsterdam criteria–positive family, a laboratory needs to ensure that it has a validated sensitivity of greater than 86% for detection of MSH in cancers from MMR mutation carriers.17 While most reports find figures above this level, a number of reports do not,17,22 and clinicians and laboratories should take care before abandoning the Amsterdam criteria on the basis of a single laboratory test in selecting an individual or family for MMR germline testing.
Immunohistochemistry An alternative or complementary test to MSI is that of antibody staining of microscope slides from relevant cancers (colorectal, endometrial, etc) for the main MMR proteins. Absence of staining is highly likely to mean that the gene has been mutated or lost in both copies or, in the case of MLH1, may have been silenced by somatic methylation. Immunohistochemistry (IHC) can be undertaken at the time of initial tumor diagnosis by the pathologist (or on archived tissue), and it informs the geneticist/scientist which gene to test for germline mutations. If MLH1 is lost, a test for promoter hypermethylation can be carried out, which may obviate the need for mutation screening. Like MSI, IHC has major advocates. Some studies boast 90% to 100% sensitivity, with specificity for loss of MSH2 being also close to 100%. However,
certain pathogenic mutations, such as missense mutations, may leave a stable protein product, which will give a false negative result (protein staining will persist). Thus, IHC needs to be performed by a trained pathologist or technician and is open to interpretation with some reports of “patchy” staining. Sensitivities as low as 72% have been reported for IHC in one study,23 and two of 18 patients with mutations in an Australian study were scored as having absent protein only after some debate.22 Therefore, neither lack of MSI nor the presence of protein by IHC alone reliably excludes (below 10%) a germline mutation in families fulfilling the Amsterdam criteria.
Population Testing for Colorectal Cancer It is relatively straightforward when a patient with colorectal or endometrial cancer presents to a cancer genetic/high-risk clinic to decide whether or not to initiate genetic analysis. Referral criteria to such clinics usually preselect on the basis of a strong family history. What is less clear is how patients should be selected on a wider basis for mutation testing. The results of population-based testing for colorectal cancer are presented in Table 2. To our knowledge, no large population study has tested all incident cases for MMR germline mutations regardless of age. Some studies have preselected patients based on the Bethesda guidelines for MSI or IHC or both,24-26 while others have selected on the basis of age and tested all incident cases.19,22 Still others have only tested HNPCC or suspected HNPCC cases.27,28 Detection rates for MMR mutations vary from 0.3%25,29 to 3% in population series in which all incident cases have received some selection other than HNPCC/suspected HNPCC. In the age-selected series with more comprehensive testing, this figure rises to 4% for age less than 55 years19 to 13% for age less than 45 years.22 Apart from the latter two comprehensive analyses, which also included MSH6, the other series were likely to have missed MMR germline patients for a number of reasons inherent in study design. The Finnish study was atypical in
Identifying hereditary nonpolyposis colon cancer
Table 3 Cost-Effectiveness Studies of Different Studies to Identify HNPCC Carriers
Cost of MSI MMR germline testing Testing relative for mutation Counseling costs Cost-effectiveness using Bethesda guidelines per QALY Cost-effectiveness testing all patients with MSI per QALY Cost-effectiveness testing all patients for MMR per QALY Cost per carrier using Amsterdam Cost per additional carrier using Bethesda (MSI) Cost per additional carrier using Amsterdam ⴙ Bethesda (MSI) Cost per additional carrier testing all patients with MSI
Kievet et al,31 2005
Ramsey et al,32 2001
Reyes et al,33 2002
529 Euro ($687) 1,308 Euro ($1,700) 286 Euros ($370) 54 Euros 3,801 Euro*
$120 (high end) $2,030 (high end) $78 $100 $7,556
$350 $1,950 $315 $295
Ramsey et al,43 2003
$11,865 $35,617 $267,548 $277 $6,832 $6,441 $51,151
*Using revised Bethesda guidelines.9
that 17 of 18 mutations detected were in MLH1 and 50% (nine of 18) mutations were a single founder deletion.26 Perhaps the best estimate in a non-founder population is from the United States, where 1,066 consecutive colorectal cancers were evaluated with MSI and IHC. Twenty-three mutations were identified (two in PMS2) at a rate of 2.1%.30 As such, it would appear that a likely estimate for germline MMR mutations in all incident cases of colorectal cancer outside founder populations would be 1% to 2.5%. If the revised Bethesda guidelines had been implemented in the Scottish study of 870 colorectal cancer cases diagnosed before age 55 years,19 two mutation carriers would have been missed among the 315 patients who did not fulfill the Bethesda criteria. However, neither of these was identified on the basis of MSI or IHC loss. Given the cost of full MMR analysis (⬃$2,000) it would have cost $315,000 per mutation carrier to detect these two cases. In the US study,30 four of the patients with mutations in the accepted high-risk genes (MSH2, MLH1, and MSH6) did not have a history fulfilling the Bethesda criteria. Unfortunately, the presence or absence of Bethesda criteria was not reported for the whole cohort. If we take a conservative estimate that 75% would not have met the Bethesda criteria,24 then only four of 800 (0.5%) cases not fulfilling the Bethesda criteria would have been missed. The results from three cost effectiveness analyses31-33 are presented in Table 3. It would appear that, using both the Bethesda and revised Bethesda guidelines, a very acceptable cost per life-year is obtained, when the effect on relatives is included. The extra cost of testing all incident cases appears prohibitive. It is notable that the mixed strategy of testing for MMR mutation in all cases fulfilling Amsterdam criteria was actually cheaper than the strategy of testing all Bethesda cases, including Amsterdam cases, for MSI first. Some authors have suggested combining the two prescreening strategies of MSI and IHC prior to MMR testing. If this is done sequentially, it will build in a significant delay, especially for
Amsterdam-positive cases with a higher a priori risk. It is not clear whether combining pre-screening tests is therefore justified. Some authors have suggested that the sensitivity of MSI is so high that all families should have MSI as a prescreen before mutation testing.23 Our own analysis and others have found that MSI does not reliably have a sensitivity above 86%, which would be necessary to negate the need for MMR testing using a 10% threshold.17 It also has been suggested that protein staining alone or in combination with MSI increases sensitivity, but the interpretation of absence of staining is subjective, and certain mutations are associated with a stable protein product. Given that sensitivities as low as 72% have been reported for antibody studies,23 neither MSI nor IHC alone reliably produces sensitivities sufficient to reduce the detection rate in Amsterdam families below 10%. Laboratories should be certain that their sensitivities for IHC or MSI are above 90% on a large sample set before abandonment of the Amsterdam criteria. Recently, three new models have been developed for the assessment of MMR risk in colorectal cancer.34-36 These models have not yet been independently validated and as such should not replace the Bethesda guidelines as a selection method until this has taken place.
Selection in Non-colorectal Cancer Patients Due to the known association of additional tumors with HNPCC, as well as colorectal, consideration should be given to HNPCC analysis in patients with cancers of the endometrium, stomach, ovarian, pancreas, ureter and renal pelvis, biliary tract, and brain (usually glioblastoma as seen in Turcot syndrome). The same is true of patients with sebaceous gland adenomas and keratoacanthomas in the Muir–Torre syndrome and carcinoma of the small bowel. Nevertheless,
G.D. Evans et al
416 Table 4 A Strategy to Identify HNPCC Patients Colorectal cancer patient Family history and pathology Amsterdam
?IHC first MMR testing
Offer family testing
No further testing
Colonoscopy 1–2 yearly
Nil, no further testing
5 Yearly colonoscopy
Relative not tested
Colonoscopy 1–2 yearly
Colonoscopy 1–2 yearly
although the detection may be similar to colorectal cancer, selection criteria have not been developed for patients with these findings. Even using the Bethesda criteria, most mutations would be missed.37 In one study, 118 of 543 (21.7%) unselected patients with endometrial cancer undergoing MSI testing were MSI-positive (98 MSH and 20 MSL). Only nine of 118 (8%) patients (one with MLH1, three with MSH2, and five with MSH6) had detectable mutations. In addition, one case with an MSS tumor had abnormal MSH6 staining by IHC and was subsequently found to have a mutation in MSH6. Although 1.8% of patients with newly diagnosed endometrial cancer had MMR mutations, only three (0.5%) met the Bethesda or Amsterdam criteria. In a population study of upper urothelial tumors, 11 of 216 (5%) patients showed loss of IHC expression, most commonly for MSH2 and MSH6. As such, these patients were highly likely to have germline mutations.38 A more pragmatic approach to selecting patients for testing who do not have colorectal cancer may be necessary. Clearly, any cancer occurring in the context of Amsterdam II criteria will qualify, but age selection criteria may not be as sensitive as Bethesda criteria in endometrial cancer.37
Mutation Testing for MMR Once patients have been selected for MMR gene testing, a sensitive technique or set of tests needs to be applied. Duplication or deletion of single or multiple exons constitutes 10% to 17% of detectable gene abnormalities in MSH2/MLH1.17 These will not be detected by gene sequencing. A test such as multiple ligation-dependent probe amplification (MLPA) needs to be undertaken in addition to exon scanning.17,39,40,41 In those families with Amsterdam criteria and MSH, or with loss of IHC (especially MSH2) in which a mutation is not detected, the members should be treated as having HNPCC
Screening depending on history
and screened with colonoscopies every 1 to 2 years (Table 4). A particular problem in patient selection is the high rate of somatic involvement of MLH1. Up to 15% of patients with colorectal cancer show a MMR phenotype, yet germline mutations are identified in less than 20% of these cases. MLH1 is inactivated largely by methylation and tests to identify this, particularly in elderly patients, may obviate the need for mutation screening. Finding a mutation in colorectal tumors in the BRAF gene (V600E) appears a useful additional screen in MSH patients, as this has not been found in patients with germline MMR mutations.42
Conclusion Selection of patients/families for germline MMR testing is a vital step to identifying individuals at high risk of developing the preventable cancers in the HNPCC spectrum. A number of selection strategies are being employed, and none has so far been proven to be superior over all the others. Given the cost of MMR testing, pre-screening of selected patients with MSI or IHC appears to be the best current approach, although very-high-risk families could forego this step unless IHC testing is established to guide the molecular geneticist to the correct gene.
References 1. Parkin DM, Pisani P, Ferlay J: Estimates of the worldwide incidence of eighteen major cancers in 1985. Int J Cancer 54:594-606, 1993 2. Jemal A, Siegel R, Ward E, et al: Cancer statistics. CA Cancer J Clin 56:106-130, 2006 3. Hayne D, Brown RSD, McCormack M, et al: Current trends in colorectal cancer: Site, incidence, mortality and survival in England and Wales. Clin Oncol (R Coll Radiol) 13:448-452, 2001 4. Jarvinen H, Mecklin JP, Sistonen P: Screening reduces colorectal cancer rate in families with hereditary nonpolyposis colorectal cancer. Gastroenterology 108:1405-1411, 1995
Identifying hereditary nonpolyposis colon cancer 5. Souza RF: A molecular rationale for the how, when and why of colorectal cancer of screening. Aliment Pharmacol Ther 15:451-462, 2001 6. Hemminki K, Lonnstedt I, Vaittinen P, et al: Estimation of genetic and environmental components in colorectal and lung cancer and melanoma. Genetic Epidemiol 20:107-116, 2001 7. Schoen RE: Families at risk for colorectal cancer. J Clin Gastroenterol 31:114-120, 2000 8. Warthin A: Hereditary with reference to carcinoma. Arch Intern Med 12:546-555, 1913 9. Vasen HF, Watson P, Mecklin JP, et al: New clinical criteria for hereditary non polyposis colon cancer (HNPCC, Lynch syndrome) proposed by the International collaborative group on HNPCC. Gastroenterology 116:1453-1456, 1999 10. Fuchs CS, Giovannuci EL, Colditz GA, et al: A prospective study of family history and the risk of colorectal cancer. N Engl J Med 331:16691674, 1994 11. Evans DG, Walsh S, Jeacock J, et al: The incidence of hereditary nonpolyposis colorectal cancer in a population based study of 1137 consecutive cases of colorectal cancer. Br J Surg 84:1281-1285, 1997 12. Aarnio M, Sankila R, Pukkala E, et al: Cancer risk in mutation carriers of DNA-mismatch-repair genes. Int J Cancer 81:214-218, 1999 13. Dunlop MG, Farrington SM, Carothers AD, et al: Cancer risk associated with germline DNA mismatch repair gene mutations. Hum Mol Genet 6:105-110, 1997 14. Jarvinen, HJ, Aarnio M, Mustonen H, et al: Controlled 15-year trial on screening for colorectal cancer in families with hereditary nonpolyposis colon cancer. Gastroenterology 118:829-834, 2000 15. Hill J, Walsh S, Evans DG: Screening of patients at high risk of colorectal cancer. Colorectal Dis 3:308-311, 2001 16. Pinol V, Castells A, Andreu M, et al: Accuracy of revised Bethesda guidelines, microsatellite instability, and immunohistochemistry for the identification of patients with hereditary nonpolyposis colorectal cancer. JAMA 293:1986-1994, 2005 17. Evans DG, Lalloo F, Mak A, et al: Is it time to abandon microsatellite instability as a pre-screen for selecting families for mutation testing for mismatch repair genes? J Clin Oncol 24:1960-1962, 2006 18. Park JG, Vasen HF, Park YJ, et al: Suspected HNPCC and Amsterdam criteria II: Evaluation of mutation detection rate, an international collaborative study. Int J Colorectal Dis 17:109-114, 2002 19. Barnetson RA, Tenesa A, Farrington SM, et al: Identification and survival of carriers of mutations in DNA mismatch-repair genes in colon cancer. N Engl J Med 354:2751-2763, 2006 20. Boland CR, Thibodeau SN, Hamilton SR, et al: A National Cancer Institute Workshop on Microsatellite Instability for cancer detection and familial predisposition: Development of international criteria for the determination of microsatellite instability in colorectal cancer. Cancer Res 58:5248-5257, 1998 21. Calistri D, Presciuttini S, Buonsanti G, et al: Microsatellite instability in colorectal cancer patients with suspected genetic predisposition. Int J Cancer 849:87-91, 2000 22. Southey MC, Jenkins MA, Mead L, et al: Use of molecular tumor characteristics to prioritize mismatch repair gene testing in early onset colorectal cancer. J Clin Oncol 23:1-6, 2005 23. Wahlberg SS, Schmeits J, Thomas G, et al: Evaluation of microsatellite instability and immunohistochemistry for the prediction of germ-line MSH2 and MLH1 mutations in hereditary nonpolyposis colon cancer families. Cancer Res 62:3485-3492, 2002 24. Pinol V, Castells A, Andreu M, et al: Accuracy of revised Bethesda guidelines, microsatellite instability, and immunohistochemistry for
the identification of patients with hereditary nonpolyposis colorectal cancer. JAMA 293:1986-1994, 2005 Evans DG, Wu CL, Walsh S, et al: Characterisation of hereditary nonpolyposis colon cancer families: population based series of cases. J Natl Cancer Inst 93:716-717, 2001 Salovaara R, Loukola A, Kristo P, et al: Population-based molecular detection of hereditary nonpolyposis colorectal cancer. J Clin Oncol 18:2193-2200, 2000 de Leon MP, Pedroni M, Benatti P, et al: Hereditary colorectal cancer in the general population: From cancer registration to molecular diagnosis. Gut 45:32-38, 1999 Katballe N, Christensen M, Wikman FP, et al: Frequency of hereditary non-polyposis colorectal cancer in Danish colorectal cancer patients. Gut 50:43-51, 2002 Percesepe A, Borghi F, Menigatti M, et al: Molecular screening for hereditary nonpolyposis colorectal cancer: A prospective, populationbased study. J Clin Oncol 19:3944-3950, 2001 Hampel H, Frankel WL, Martin E, et al: Screening for the Lynch syndrome (hereditary nonpolyposis colorectal cancer). N Engl J Med 352: 1851-1860, 2005 Kievit W, de Bruin JH, Adang EM, et al: Cost effectiveness of a new strategy to identify HNPCC patients. Gut 54:97-102, 2005 Ramsey SD, Clarke L, Etzioni R, et al: Cost-effectiveness of microsatellite instability screening as a method for detecting hereditary nonpolyposis colorectal cancer. Ann Intern Med 135:577-588, 2001 Reyes CM, Allen BA, Terdiman JP, et al: Comparison of selection strategies for genetic testing of patients with hereditary nonpolyposis colorectal carcinoma: Effectiveness and cost-effectiveness. Cancer 95:18481856, 2002 Chen S, Wang W, Lee S, et al: Prediction of germline mutations and cancer risk in the Lynch syndrome. JAMA 296:1479-1487, 2006 Marroni F, Pastrello C, Benatti P, et al: A genetic model for determining MSH2 and MLH1 carrier probabilities based on family history and tumor microsatellite instability. Clin Genet 69:254-262, 2006 Balmana J, Stockwell DH, Steyerberg EW, et al: Prediction of MLH1 and MSH2 mutations in Lynch syndrome. JAMA 296:1469-1478, 2006 Hampel H, Frankel W, Panescu J, et al: Screening for Lynch syndrome (hereditary nonpolyposis colorectal cancer) among endometrial cancer patients. Cancer Res 66:7810-7817, 2006 Ericson KM, Isinger AP, Isfoss BL, et al: Low frequency of defective mismatch repair in a population-based series of upper urothelial carcinoma. BMC Cancer 5:23, 2005 Taylor CF, Charlton RS, Burn J, et al: Genomic deletions in MSH2 or MLH1 are a frequent cause of hereditary non-polyposis colorectal cancer: Identification of novel and recurrent deletions by MLPA. Hum Mutat 22:428-433, 2003 Grabowski M, Mueller-Koch Y, Grasbon-Frodl E, et al: Deletions account for 17% of pathogenic germline alterations in MLH1 and MSH2 in hereditary nonpolyposis colorectal cancer (HNPCC) families. Genet Test 9:138-146, 2005 Bunyan DJ, Eccles DM, Sillibourne J, et al: Dosage analysis of cancer predisposition genes by multiplex ligation-dependent probe amplification. Br J Cancer 91:1155-1159, 2004 Domingo E, Laiho P, Ollikainen M, et al: BRAF screening as a low-cost effective strategy for simplifying HNPCC genetic testing. J Med Genet 41:664-668, 2004 Ramsey SD, Burke W, Clarke L: An economic viewpoint on alternative strategies for identifying persons with hereditary nonpolyposis colorectal cancer. Genet Med 5:353-363, 2003