Portal de Periódicos da CAPES

Prospective Results of Surveillance Colonoscopy in Dominant Familial Colorectal Cancer With and Without Lynch Syndrome

2006; Elsevier BV; Volume: 130; Issue: 7 Linguagem: Inglês

10.1053/j.gastro.2006.03.018

1528-0012

Isis Dove‐Edwin, Andrea E. van der Meulen–de Jong, J T Adams, David Mesher, Lara Lipton, Peter Sasieni, Hans F. A. Vasen, Huw Thomas,

Colorectal Cancer Screening and Detection

Background & Aims: Lynch syndrome is an autosomal dominant predisposition to colorectal cancer caused by mutations in DNA mismatch repair genes; colorectal cancer risk is high. Few studies have addressed colorectal cancer risk in individuals from dominant families without mismatch repair deficiency. We sought to establish whether these individuals are also at increased risk by examining the incidence of advanced neoplasia during surveillance. Methods: In this prospective cohort study, BAT26 testing of tumors was carried out at 2 tertiary centers on 125 individuals from 97 families (with a dominant colorectal cancer history) to classify families as Lynch syndrome (microsatellite unstable) or non-Lynch syndrome (microsatellite stable). Colonoscopy results in 288 at-risk family members were compared. Results: Twenty-nine families were classified as Lynch syndrome and 68 as non-Lynch syndrome. Seven hundred seventy-six colonoscopies were undertaken. High-risk adenomas occurred in 7 of 91 (7.7%) Lynch syndrome individuals and 15 of 197 (7.6%) non-Lynch syndrome individuals, adjusted relative risk 1.15 (95% CI: 0.6–2.3). Cancer was observed only in Lynch syndrome individuals (4/91; 4.4%), Fisher exact test, P = .010. Multiple adenomas were only seen in non-Lynch syndrome individuals (13/197; 6.6%), Fisher exact text, P = .06. Conclusions: Individuals with an autosomal dominant family history of colorectal cancer with and without evidence of Lynch syndrome are at equal risk of high-risk adenomas during surveillance, but colorectal cancer was only seen in Lynch syndrome. Therefore non-Lynch syndrome individuals do require colonoscopic surveillance, but the interval could be lengthened because risk of (interval) cancer is low. Lynch syndrome individuals require short surveillance intervals as is the recommended practice. Background & Aims: Lynch syndrome is an autosomal dominant predisposition to colorectal cancer caused by mutations in DNA mismatch repair genes; colorectal cancer risk is high. Few studies have addressed colorectal cancer risk in individuals from dominant families without mismatch repair deficiency. We sought to establish whether these individuals are also at increased risk by examining the incidence of advanced neoplasia during surveillance. Methods: In this prospective cohort study, BAT26 testing of tumors was carried out at 2 tertiary centers on 125 individuals from 97 families (with a dominant colorectal cancer history) to classify families as Lynch syndrome (microsatellite unstable) or non-Lynch syndrome (microsatellite stable). Colonoscopy results in 288 at-risk family members were compared. Results: Twenty-nine families were classified as Lynch syndrome and 68 as non-Lynch syndrome. Seven hundred seventy-six colonoscopies were undertaken. High-risk adenomas occurred in 7 of 91 (7.7%) Lynch syndrome individuals and 15 of 197 (7.6%) non-Lynch syndrome individuals, adjusted relative risk 1.15 (95% CI: 0.6–2.3). Cancer was observed only in Lynch syndrome individuals (4/91; 4.4%), Fisher exact test, P = .010. Multiple adenomas were only seen in non-Lynch syndrome individuals (13/197; 6.6%), Fisher exact text, P = .06. Conclusions: Individuals with an autosomal dominant family history of colorectal cancer with and without evidence of Lynch syndrome are at equal risk of high-risk adenomas during surveillance, but colorectal cancer was only seen in Lynch syndrome. Therefore non-Lynch syndrome individuals do require colonoscopic surveillance, but the interval could be lengthened because risk of (interval) cancer is low. Lynch syndrome individuals require short surveillance intervals as is the recommended practice. In up to 30% of colorectal cancer cases, individuals report a family history of colorectal cancer. In developed countries, approximately 20% of individuals with colorectal cancer have a family history of 1 or more affected first- or second-degree relatives,1Lynch H.T. de la Chapelle A. Hereditary colorectal cancer.N Engl J Med. 2003; 348: 919-932Crossref PubMed Scopus (1712) Google Scholar and epidemiologic evidence suggests that up to 15% of colorectal cancers may be due to dominant inherited predispositions; most of which are uncharacterized.2Houlston R.S. Collins A. Slack J. Morton N.E. Dominant genes for colorectal cancer are not rare.Ann Hum Genet. 1992; 56: 99-103Crossref PubMed Scopus (132) Google Scholar Lynch syndrome is caused by mutations in DNA mismatch repair genes, mainly MLH1, MSH2, and MSH6. It is the most common autosomal dominant syndrome predisposing to colorectal cancer and is associated with a high frequency of microsatellite instability [MSI-H] phenotype in tumor tissue. Lynch syndrome is also associated with an increased risk of cancer of the endometrium and, to a lesser extent, ovary, stomach, brain, small bowel, and urinary tract (renal pelvis and ureter). Colorectal cancer risk in Lynch syndrome is known to be high, and the International Collaborative Group on Hereditary Nonpolyposis Colorectal Cancer (HNPCC) has recommended regular surveillance for gene carriers and at-risk family members. Several studies have been published as to how best to identify Lynch syndrome in the general population, including the feasibility of direct mutation testing.3Lipton L.R. Johnson V. Cummings C. Fisher S. Risby P. Eftekhar Sadat A.T. et al.Refining the Amsterdam Criteria and Bethesda Guidelines testing algorithms for the prediction of mismatch repair mutation status in the familial cancer clinic.J Clin Oncol. 2004; 22: 4934-4943Crossref PubMed Scopus (80) Google Scholar, 4Aaltonen L.A. Salovaara R. Kristo P. Canzian F. Hemminki A. Peltomaki P. et al.Incidence of hereditary nonpolyposis colorectal cancer and the feasibilty of molecular screening for the disease.N Engl J Med. 1998; 338: 1481-1487Crossref PubMed Scopus (996) Google Scholar, 5Loukola A. de la Chapelle A. Aaltonen L. Strategies for screening for hereditary non-polyposis colorectal cancer.J Med Genet. 1999; 36: 819-822PubMed Google Scholar, 6Terdiman J.P. Gum J.R. Conrad P.G. Miller G.A. Weinberg V. Crawley S.C. et al.Efficient detection of hereditary nonpolyposis colorectal cancer gene carriers by screening for tumor microsatellite instability before germline genetic testing.Gastroenterology. 2001; 120: 21-30Abstract Full Text Full Text PDF PubMed Scopus (148) Google Scholar, 7Hampel H. Frankel W.L. Martin E. Arnold M. Khanduja K. Kuebler P. et al.Screening for the Lynch syndrome (hereditary nonpolyposis colorectal cancer).N Engl J Med. 2005; 352: 1851-1860Crossref PubMed Scopus (1176) Google Scholar Interestingly, the most recent study also showed that the majority of Lynch syndrome cases, identified by routine molecular screening of patients with colorectal cancer, did not fulfill the Amsterdam Criteria.7Hampel H. Frankel W.L. Martin E. Arnold M. Khanduja K. Kuebler P. et al.Screening for the Lynch syndrome (hereditary nonpolyposis colorectal cancer).N Engl J Med. 2005; 352: 1851-1860Crossref PubMed Scopus (1176) Google Scholar However, a substantial proportion of families that fulfill the Amsterdam Criteria do not have molecular evidence of Lynch syndrome (referred to as non-Lynch families in this article). Few, if any, studies have addressed the colorectal cancer risk in individuals from such families. We sought to establish whether individuals from non-Lynch families were also at increased risk of developing colorectal cancer. We have carried out a prospective study of the outcome of colonoscopic surveillance in at-risk individuals with a strong family history of colorectal cancer and compared the results in families with and without Lynch syndrome. All families included in this study were registered with either the Cancer Research UK Family Cancer Clinic at St. Mark's Hospital, London, United Kingdom, or The Netherlands Foundation for the Detection of Hereditary Tumours and were enrolled in colonoscopic surveillance. All families had a history suggestive of autosomal dominant pedigree inheritance of colorectal cancer. Families either fulfilled the Amsterdam Criteria I or II8Vasen H.F.A. Mecklin J.-P. Khan P.M. Lynch H.T. The International Collaborative Group on Hereditary Non-Polyposis Colorectal Cancer.Dis Colon Rectum. 1991; 34: 424-425Crossref PubMed Scopus (1778) Google Scholar, 9Vasen H.F.A. Watson P. Mecklin J.-P. Lynch H.T. ICG-HNPCC at. New clinical criteria for hereditary nonpolyposis colorectal cancer (HNPCC, Lynch syndrome) proposed by the International Collaborative Group on HNPCC.Gastroenterology. 1999; 116: 1453-1456Abstract Full Text Full Text PDF PubMed Scopus (2146) Google Scholar or were classified as a dominant pattern pedigree family (colorectal cancer in 3 or more relatives, 1 a first-degree relative of the other 2, involving at least 2 generations) but no cases were diagnosed below the age of 50 years. Families were invited to join the study, and informed consent was obtained. All families had at least 1 colorectal tumor from an affected individual tested for MSI, and families were classified as Lynch syndrome if ≥50% of tumors tested showed evidence of MSI-H or if there was a known mismatch repair gene mutation. MSI was defined as instability at the BAT26 locus. Amsterdam Criteria families from The Netherlands Foundation had previously undergone testing for mutations in MLH1, MSH2, and MSH6, but no mutation had been found. Families with no known mutation and with <50% MSI were classified as non-Lynch syndrome. Data for the study were collected between March 1, 1987, and December 31, 2003. Ethical approval was obtained from the Harrow Research Ethical Committee for St. Mark's patients, and informed consent was obtained for DNA studies on blood and tumor tissue from affected family members. At the time of registration with The Netherlands Registry, all subjects were informed about the aims of the Registry, which include evaluation of follow-up and molecular markers, and gave written informed consent for their data to be used by the Registry. Colonoscopic surveillance was carried out on (unaffected) at-risk first-degree relatives of any affected member of the family. Results of colonoscopies, including histologic analysis of any tissue removed, were obtained from a secure database. If colonoscopic surveillance was being carried out at another hospital, results of the examination were requested by letter and then entered onto the database. Colonoscopies carried out after an individual had been diagnosed with colorectal cancer were excluded. At St. Mark's, initial surveillance protocols offered a colonoscopy every 5 years from age 25 years, increasing to 3 years if adenomas were seen. From 1997, Amsterdam Criteria families were offered 1- to 3-year colonoscopic surveillance in keeping with the International Collaborative Group on HNPCC guidelines.10Weber T. Clinical surveillance recommendations adopted for HNPCC.Lancet. 1996; 348: 465Abstract Full Text Full Text PDF Google Scholar In dominant pattern pedigree families, 5-year surveillance (with an interval of 3 years if adenomas are seen) was continued, commencing at age 40–45 years. In The Netherlands, individuals from Amsterdam Criteria families have been offered colonoscopic surveillance at least every 2 years since 1995 and, prior to this, every 3 years. At-risk members of dominant pattern pedigree families are offered 5-year surveillance, starting from age 45–50 years. The end points of the study were the presence of high-risk adenomas or cancer. High-risk adenomas are defined as adenomas with villous histology or a diameter of at least 10 mm or high-grade dysplasia. Advanced neoplasia is defined as either a high-risk adenoma or cancer. Genomic DNA was extracted from tumor blocks and blood using standard methods. MSI analysis was carried out using the BAT26 microsatellite marker, which has been shown to have a high sensitivity for the MSI-H phenotype in colorectal tumors.11de la Chapelle A. Testing tumours for microsatellite instability.Eur J Hum Genet. 1999; 7: 407-408Crossref PubMed Scopus (48) Google Scholar, 12Hoang J.-M. Cottu P.H. Thuille B. Salmon R.J. Thomas G. Hamelin R. BAT-26, an indicator of the replication error phenotype in colorectal cancers and cell lines.Cancer Res. 1997; 57: 300-303PubMed Google Scholar Fluorescent primers were used, and polymerase chain reaction was carried out using previously described methods.12Hoang J.-M. Cottu P.H. Thuille B. Salmon R.J. Thomas G. Hamelin R. BAT-26, an indicator of the replication error phenotype in colorectal cancers and cell lines.Cancer Res. 1997; 57: 300-303PubMed Google Scholar, 13de Jong A.E. van Puijenbroek M. Hendriks Y. Tops C. Wijnen J. Ausems M.G. et al.Microsatellite instability, immunohistochemistry, and additional PMS2 staining in suspected hereditary nonpolyposis colorectal cancer.Clin Cancer Res. 2004; 10: 972-980Crossref PubMed Scopus (192) Google Scholar Familial age of onset is defined as the median age of onset within a family. Results at colonoscopy are presented as the most advanced lesion found at the examination. Tests of significance included Pearson χ2 test, Fisher exact test, and 2 sample t-tests. To calculate the odds ratios for high-risk adenomas and advanced neoplasia associated with belonging to a Lynch syndrome (compared with a non-Lynch syndrome) family, we used logistic regression adjusting throughout for the effect of age, gender, family extent, center, and nonindependence of individuals within a family. To calculate the odds ratio of the location (left or right sided) of adenomas in the bowel associated with belonging to a Lynch syndrome family, we used logistic regression controlling for center and nonindependence of individuals within a family. Logistic regression analysis was performed in Stata (StataCorp 2003; Statistical Software: Release 8.0; Stata Corporation, College Station, TX). Twenty-nine out of the 97 families tested exhibited MSI and were therefore classified as Lynch syndrome families (Table 1). The median (range) age at diagnosis of cancers tested was 44 (25–75) years in the Lynch syndrome group and 54 (18–82) years in the non-Lynch syndrome group. Only 3 of the 34 MSI tumors were from individuals aged 60 years or older.Table 1Family Classification and Median Age at First Cancer Diagnosis Within PedigreeFulfilling ACDominant pattern pedigreeTotalLynch syndrome26 [30]3 [3]29 [33]48 (25–65)60 (48–75)48 (25–75)Non-Lynch syndrome45 [62]23 [30]68 [92]50 (18–82)58 (28–79)54 (18–82)Total N = 97 [125]71 [92]26 [33]97 [125]48 (18–82)58 (28–79)50 (18–82)NOTE. Figures are number of families [number of individuals]: median (range) age (y) of diagnosis. Open table in a new tab NOTE. Figures are number of families [number of individuals]: median (range) age (y) of diagnosis. A total of 776 surveillance colonoscopies were carried out in first-degree relatives from 97 families. There were 288 individuals: 91 Lynch syndrome and 197 non-Lynch syndrome. A cecal intubation of 97% was achieved. If the colonoscopy was not complete, it was repeated or a barium enema carried out to visualize the right colon. The mean number of colonoscopies per individual was 2.7: this was 3.2 in the Lynch group and 2.5 in the non-Lynch group. The surveillance intervals were similar for Lynch and non-Lynch. The median age at first colonoscopy was 36 (range, 20–73) years in Lynch syndrome and 44 (range, 20–77) years in non-Lynch syndrome, P = .0005. The findings at surveillance colonoscopy overall are presented in Table 2. A normal colonoscopy was seen in 162 of 288 (56%) individuals. The most advanced neoplastic finding was at least 1 adenoma or cancer in 91 of 288 (32%) individuals. Hyperplastic polyps were the most advanced finding in 35 of 288 (12%) individuals. Thirteen of 197 (6.6%) non-Lynch individuals had multiple adenomas (either ≥3 adenomas in 1 single colonoscopy or ≥5 adenomas during surveillance history). These were only seen in individuals from non-Lynch families. The highest number of adenomas found in any colonoscopy was 6, and this occurred in 3 individuals.Table 2Most Advanced Neoplastic Lesion Found at Each Surveillance Colonoscopy in At-Risk Family Members From Lynch Syndrome vs Non-Lynch Syndrome FamiliesColonoscopies, n (%)Lynch syndromeNon-Lynch syndromeNormal50 (54.9)112 (56.8)Hyperplastic polyp13 (14.3)22 (11.2)Adenoma17 (18.7)39 (19.8)Three or more adenomas0 (0.0)9 (4.6)High-risk adenoma7 (7.7)15 (7.6)Cancer4 (4.4)0 (0.0)Total91 (100.0)197 (100.0) Open table in a new tab High-risk adenomas were seen in 22 individuals: 7 of 91 (7.7%) from Lynch syndrome families and 15 of 197 (7.6%) from non-Lynch syndrome families. Eight individuals had an adenoma with villous histology only, 4 had adenomas ≥10 mm only, 3 had adenomas with severe dysplasia only, and 7 had 2 high-risk features. The median age for individuals with high-risk adenomas was 49 (range, 32–65) years in Lynch syndrome and 53 (range, 29–72) years in non-Lynch syndrome individuals. After adjusting for age, gender, family history, center, and possible nonindependence of individuals within a family, the odds ratio of high-risk adenomas associated with belonging to a Lynch syndrome (vs non-Lynch syndrome) family is 1.15 (95% CI: 0.6–2.3), P = .69. Cancer was seen in 4 of 91 (4.4%) individuals from Lynch syndrome families, but no cases occurred in non-Lynch syndrome individuals, Fisher exact test, P = .010. One cancer was seen on the first colonoscopy, whereas the other 3 were identified after a normal colonoscopy 35 months, 37 months, and 77 months, respectively, previously. The median (range, 31–52) age at diagnosis with cancer was 48.5 years. Three of the 4 individuals have been found to be HNPCC mutation carriers; the results of the fourth individual are awaited. Three of the cancers were Dukes' stage A, and the fourth, detected 37 months after a clear colonoscopy, was widely metastatic at presentation (Table 3).Table 3Characteristics of Individuals With Advanced NeoplasiaIndividualAge (y)FindingLocationaRight-sided lesions are situated proximal to the splenic flexure.Interval between last colonoscopy (y)bInterval between last colonoscopy considered as NA if first recorded colonoscopy.Lynch syndrome132HGDRight2.13242VALeftNA347VALeft5.33449VA, HGDLeft1.1954TA ≥ 10 mmRight2.01553TA ≥ 10 mmRightNA665VARight2.21765VALeftNA831CA (metastatic)Left3.05948CA (A)Right2.881049CA (A)Right6.381152CA (A)RightNANon-Lynch syndrome1329VALeftNA1432VA, TA ≥ 10 mmLeftNA1544VALeft2.651645TA ≥ 10 mmLeftNA60VA, HGDLeft15.141747VA, HGDLeftNA1849VA, TA ≥ 10 mmRight1.491952HGDRight2.502053HGDLeft5.172153TA ≥ 10 mm, HGDLeftNA2256TA ≥ 10 mmLeftNA2356VALeftNA2460VALeftNA2562TA ≥ 10 mm, HGDLeftNA2669VA, HGDRightNA2772TA ≥ 10 mmRight3.29VA, villous histology; TA, tubular adenoma; HGD, high-grade dysplasia; CA, cancer (Dukes' stage).a Right-sided lesions are situated proximal to the splenic flexure.b Interval between last colonoscopy considered as NA if first recorded colonoscopy. Open table in a new tab VA, villous histology; TA, tubular adenoma; HGD, high-grade dysplasia; CA, cancer (Dukes' stage). Fifty-three percent of cancers or adenomas were located in the left side of the bowel, distal to the splenic flexure. There was very little evidence of any difference between Lynch syndrome and non-Lynch syndrome individuals. The odds ratio of having an adenoma on the right associated with belonging to a Lynch syndrome (compared with a non-Lynch syndrome) family is 0.9 (95% CI: 0.4–2.3; P = .90) when adjusting for center and possible nonindependence of individuals from the same family (Table 4).Table 4Location of Adenomas or Cancers Found at Surveillance Colonoscopy in At-Risk Family Members From Lynch Syndrome and Non-Lynch Syndrome FamiliesColonoscopies, n (%)TotalLeft colonRight colonLynch syndrome22 (53.7)19 (46.3)41 (100.0)Non-Lynch syndrome62 (52.1)57 (47.9)119 (100.0)Total84 (52.5)76 (47.5)160 (100.0)NOTE. Right-sided lesions are located proximal to the splenic flexure. Open table in a new tab NOTE. Right-sided lesions are located proximal to the splenic flexure. In this study, we prospectively evaluated the incidence of neoplasia during endoscopic surveillance in dominant families at risk of colorectal cancer with and without Lynch syndrome. We found that non-Lynch syndrome families (without MSI/mismatch repair [MMR] deficiency) when compared with Lynch syndrome families (with MSI/MMR deficiency) are actually at equal risk of developing high-risk adenomas but at significantly lower risk of developing (interval) cancers. Individuals from non-Lynch families are at an increased risk of developing multiple adenomas. Diagnosing Lynch syndrome has been the focus of several studies of familial colorectal cancer. Until recently, not much attention has been given to the families who do not have Lynch syndrome. Prospective assessments of colorectal cancer risk/incidence in these families are still lacking. Lindor et al retrospectively evaluated the cancer incidence in Amsterdam Criteria families with and without evidence of MSI/DNA MMR deficiency and reported a decreased incidence of colorectal cancer and HNPCC-related cancers in families who did not have evidence of MMR deficiency.14Lindor N.M. Rabe K. Petersen G.M. Casey G. Baron J. Gallinger S. et al.Lower cancer incidence in Amsterdam-I criteria families without mismatch repair deficiency. Familial colorectal cancer type X.JAMA. 2005; 293: 1979-1985Crossref PubMed Scopus (493) Google Scholar Our results are in line with these findings. The genetic basis of non-Lynch syndrome colorectal cancer predispositions remains unclear. Familial clustering of colorectal cancer is common, and this group is likely to be genetically diverse and include families in which clustering occurs by chance. The actual risk of developing colorectal cancer will vary widely; individuals who do not carry a cancer-predisposing gene may have a risk that is no higher than the general population, whereas the risk in mutation carriers may be substantial. It is not known whether the mutations involved are in novel, as yet undiscovered, genes or less penetrant mutations in genes that are known to cause colorectal cancer predispositions. For example, it is possible that the multiple adenoma phenotype seen in the non-Lynch group is as a result of MYH or I1307K mutations.15Laken S.J. Petersen G.M. Gruber S.B. Oddoux C. Ostrer H. Giardello F.M. et al.Familial colorectal cancer in Ashkenazim due to a hypermutable tract in APC.Nat Genet. 1997; 17: 79-83Crossref PubMed Scopus (540) Google Scholar, 16Sieber O. Lipton L. Crabtree M. Heinimann K. Fidalgo P. Phillips R. et al.Multiple colorectal adenomas, classic adenomatous polyposis and germ-line mutations in MYH.N Engl J Med. 2003; 348: 791-799Crossref PubMed Scopus (756) Google Scholar At present, the genetic basis of Non-Lynch families is unknown and it is therefore difficult to make definitive recommendations about surveillance. As 40% of neoplastic lesions were located proximal to the splenic flexure, it would be appropriate to recommend full colonoscopic surveillance. Because no cancers developed during surveillance in the non-Lynch group (and in the absence of mismatch repair gene deficiency, the adenoma-carcinoma sequence is not thought to be accelerated), intervals of at least 5 years if colonoscopy is clear as suggested by Lindor et al14Lindor N.M. Rabe K. Petersen G.M. Casey G. Baron J. Gallinger S. et al.Lower cancer incidence in Amsterdam-I criteria families without mismatch repair deficiency. Familial colorectal cancer type X.JAMA. 2005; 293: 1979-1985Crossref PubMed Scopus (493) Google Scholar should be appropriate. In a separate study, the prospective results of surveillance in at-risk members of moderate risk colorectal cancer families—half of whom were classified as a dominant pattern pedigree family (colorectal cancer in 3 or more relatives, 1 a first-degree relative of the other 2, involving at least 2 generations but no cases diagnosed below the age of 50 years)—showed that cancer was extremely rare when surveillance was carried out at 5-year intervals, which was reduced to 3 years if an adenoma was detected.17Dove-Edwin I. Adams J. Sasieni P. Thomas H.J.W. Prevention of colorectal cancer by colonoscopic surveillance in individuals with a family history of colorectal cancer 16 year, prospective, follow-up study.Br Med J. 2005; 331: 1047-1049Crossref PubMed Scopus (175) Google Scholar Advanced neoplasia was also rare in this moderate risk group in individuals under the age of 45 years. However, we have found 3 individuals from non-Lynch families with advanced neoplasia under the age of 45 years. It is interesting to note that all 3 came from families meeting the Amsterdam Criteria. Current suggestions that surveillance should begin within 5–10 years younger than the youngest pedigree cancer are at best empirical but would appear to be reasonable for non-Lynch syndrome families. It may become feasible to screen for Lynch syndrome at a population level, using immunohistochemistry or MSI as a first step. A recent study has shown that, in this situation, most cases of Lynch syndrome were not associated with a strong family history (Amsterdam Criteria).7Hampel H. Frankel W.L. Martin E. Arnold M. Khanduja K. Kuebler P. et al.Screening for the Lynch syndrome (hereditary nonpolyposis colorectal cancer).N Engl J Med. 2005; 352: 1851-1860Crossref PubMed Scopus (1176) Google Scholar At present, however, the majority of cases of familial colorectal cancer is identified on the basis of a strong family history, the importance of which has been emphasised previously by Lynch and Smyrk.18Lynch H.T. Smyrk T.C. Identifying hereditary nonpolyposis colorectal cancer.N Engl J Med. 1998; 338: 1537-1538Crossref PubMed Scopus (57) Google Scholar Family history should remain an integral part of the risk assessment to individualize surveillance strategies in families with strong (non-DNA mismatch repair) genetic predispositions, as well as to improve the sensitivity of molecular testing in identifying Lynch syndrome. A possible limitation of this study is that a small proportion of Lynch syndrome tumors do not exhibit MSI, and this may have led to some families being misclassified as non-Lynch. In contrast, a proportion of tumors exhibit MSI because of somatic methylation of the MMR gene MLH1, and these tumors may have been misclassified as Lynch syndrome. Kakar et al report a significant increase in loss of MLH1 expression with advancing age in sporadic colorectal cancer.19Kakar S. Burgart L.J. Thibodeau S.N. Rabe K.G. Petersen G.M. Goldberg R.M. et al.Frequency of loss of hMLH1 expression in colorectal carcinoma increases with advancing age.Cancer. 2003; 97: 1421-1427Crossref PubMed Scopus (90) Google Scholar In the setting of a strong family history, the proportion is likely to be less; in this study, only 3 out of 34 tumors showing MSI were from individuals above the age of 60 years. The families from The Netherlands Registry had been previously tested for HNPCC mutations, and none was found. The classification of the dominant families in this study was based on the MSI status of the proband's tumor, and it is not uncommon to be unable to identify MMR gene mutations in dominant families with tumors exhibiting MSI using current mutation detection techniques. Colorectal cancer surveillance for individuals with a strong family history will become increasingly targeted and specific. It will be informed by evidence of DNA MMR deficiency and further knowledge of the phenotype, genotype, and biologic behavior of the tumors from non-Lynch syndrome families. In this prospective study, we have shown that at-risk individuals from non-Lynch syndrome families are at significant risk of developing advanced neoplasia and, therefore, require targeted colonoscopic surveillance. Non-Lynch syndrome families did not develop interval cancers when surveillance was carried out every 5 years (3 years if adenomas are seen), suggesting that these intervals are appropriate. The authors thank Helen Crowne, Steve Edmeades, and Linda Godfrey from Cancer Research UK Clinical Information Systems for setting up the Bobby Moore Database and Professor John Northover and Dr Joan Slack, who established the Family Cancer Clinic; Dr Christopher Williams and other members of the Wolfson Endoscopy Unit at St. Mark's Hospital for their contributions to the colonoscopic surveillance; Maggie Stevens and Julie Stokes for their administrative input; nurses Carole Cummings and Sheila Goff for their help with the patients; and all the individuals and families who have taken part in this study.