REPORT SIXTH MEETING OF THE ICG-HNPCC
MILAN, 11 SEPTEMBER 1994

1. Opening

Henry Lynch opened the meeting and welcomed the participants. Hans Vasen gave a short outline of the program. He memorized that during the previous meetings the group focused on discussions on collaborative studies. For the present meeting there were unfortunately no new proposals. Probably the members were confused by the enormous progress which has been made in molecular genetic studies. Therefore, the council decided to focus this meeting on discussion of the new genetic findings and their implications for general practice. Vasen said that he hoped that the following questions will be answered. Regarding DNA-testing: how many genes are respon- sible for HNPCC?; How reliable are the tests?; How can quality control be assured?; Who should be tested? With respect to counselling of the families: Who should counsel the family: the clinical geneticist, the gastroenterologist or both?; Should a special protocol be used?; Is there a need for special facilities e.g. family cancer clinics?; Regarding the surveillance programs: Are the available programs safe or is prophylactic surgery a better alternative? To get more information on he effect of surveillance Vasen stressed the need for cooperation in the ICG in establising such studies.

2. Molecular genetic studies

Data from Finland
Lauri Aaltonen summarized the results of the molecular genetic studies. He revealed that in Finland 14 families were found to have the same mutation in MLH1. An extensive genealogic study, going back to the 16th century, disclosed that these families were related. Another mutation in MLH1 was found in 5 families which also were found to be related. Aaltonen said that knowledge about the indicence of HNPCC is very important. Unfortunately, the epidemiological studies on the incidence which have been reported so far are not very successful. In his opinion the only way to solve this problem is by molecular genetic screening of a population-based series of colorectal cancer patients. In Finland such a study is initiated in an area which covers about one quarter of Finland with a population of 1.4 million people.

Data from the United Kingdom (Nigel Hall)
The content of the presentation related mainly to work carried out in Leeds at the ICRF Genetic Epidemiology Laboratory and in collaboration with Dr. Richard Kolodner (Dana Farber Cancer Center, Boston), but includes a summary of the linkage results from Dr. Eamonn Maher in Cambridge. Linkage analysis was performed for markers around hMSH2 and hMLH1 in 21 families with HNPCC, 15 of which fulfill the Amsterdam criteria. Results suggest linkage to hMSH2 in 7 families and hMLH1 in 4 families. In seven families the results were compatible with linkage to either locus but in four there was apparent exclusion of both loci. These remaining four families are therefore candidates for linkage to hPMS1 en hPMS2. Mutations were sought (this work performed by Richard Kolodner) in two large Muir-Torre pedigrees which were found to be linked to chomosome 2p. In family 1 a T to C transition was found in the splice acceptor region six bases upstream from the beginning of exon 13 in hMSH2. There was theoretical and circumstantial evidence that this was a good candidate disease-producing mutation. However, a study of the frequency of this sequence change in 111 colorectal cancer cases and 114 controls failed to demonstrate a difference between the two groups. The sequence variation is thus considered to be a polymorphism. Further analysis revealed the true mutation in this family to be a single base change in exon 12, producing a stop codon and a protein truncation. In family 2 the mutation was found to be a two base pair deletion (also in exon 12), resulting in a frame shift and stop codon. We have found strong evidence for linkage to chromosome 3p markers in a very large family from the north of England in which thirty-six members have had colorectal cancer out of a total of 54 persons with cancer in the family. Mutation analysis demonstrated a single base insertion in exon 13, predicted to result in an immediate stop codon and a truncated protein. In Cambridge, sixteen HNPCC families were studied with 2p and 3p markers. In six there was good evidence for linkage to hMSH2 and in 4 the data suggested hMLH1 linkage. In the remaining six families it was difficult to exclude linkage to either locus. Work from both Cambridge and Leeds found that there was no apparent correlation be- tween the spectrum of tumours in a family and the likelihood of finding linkage to any particular locus. There thus appears to be considerable genetic and phenotypic heterogeneity.

Data from The Netherlands
Meera Khan revealed that in Holland mutations in hMSH2 have been found in 5 families. The total number of first degree relatives in these families was 93. Due to the DNA testing 36 relativeswere found to be carrier of the gene. This means that the others can be excluded from follow-up. Meera stressed that future efforts should be addressed to optimalize the laboratory tests, to develop effective screening protocols, to identify (population-based) high risk groups, and to assess the psychological implications. Also close coperation between physicians and laboratories is of importance.

Future perspective
Albert de la Chapelle shortly summarized the molecular genetic findings on HNPCC repor- ted in the literature. He said that the mutations in hMSH2 were spread from exon 5 to 15. This gene and the other repair genes should probably be considered as tumour suppressor genes. Future goals of investigations should include: (1) development of easy methods of mutation detection; (2) determination of penetrance and expression; (3) estimation of incidence of HNPCC; (4) development of strategies regarding clinical handling of mutation information, and (5) elucidating the pathogenetic mechanisms. De la Chapelle emphasized the importance to be informed about the incidence of HNPCC. Unfortunately, the epidemiological studies carried out so far didn't solve this problem considering the wide variety in the calculated frequencies (1/200 - 1/2000). The available estimates of the incidence of HNPCC are probably underestimates because (1) of the occurrence of germline mutations; (2) a small size of families preclude fulfillment of the Amsterdam criteria; (3) in the reported studies HNPCC was only identified through CRC patients and not through patients with endometrial cancers. The only way to get reliable data on this issue will be by molecular genetic screening of a population-based series of CRC patients. Such a study is initiated in Eastern Finland (population: 1.3 million). Out of all newly diag- nosed colorectal cancers, the patients with RER+ tumours will be selected for further molecular genetic study.

3. Genetic counselling and presymptomatic DNA-testing in HNPCC

Fred Menko:
The identification of germ line mutations that underlie HNPCC now allows presymptomatic DNA-testing in selected families. This possibility evokes a series of questions on the impact of testing. Questions pertain to the issue of informed consent, psychosocial consequences of testing, the impact on surveillance behaviour, testing of children, use of research samples and application for prenatal testing.

Helena Kaariainen:
Kaariainen said that in Finland the results of DNA-testing in 20 families made counselling possible. These families include a total of 250 relatives. In Finland extensive experience exists on counselling of Huntington families. A comparable protocol will be used in the HNPCC families. The frequency of the sessions in HNPCC, however, will be less considering the possibility of early diagnosis and treatment of HNPCC in contrast with Huntington disease. The protocol is as follows: during the first session extensive information is provided about the disease, the possibilities of surveillance, the test itself etc.; after an interval of several weeks the patients are asked (by telephone) about their decision. During the second session the results will be delivered. At a third session the follow up will be discussed. At all times psychological support is available. In the discussion the question was posed whether the test in children also should be performed. Most members were of the opinion that this should be avoided because the results do not have any consequence regarding the screening protocol, which is initiated at an age of 20-25 years. However, if grown-ups absolutely want to know about their gene status, probably this cannot be refused. Problems which are encountered in Finland at present include: the large numbers of counsellings, the large distances, the previous well organized follow up which made it sometimes difficult to convince the relatives without the HNPCC-gene that the examinations may be stopped; the occurrence of coincidental cancer; the fact that the mutation is not known in all families.

4. Surveillance

Hans Vasen presented the results of a study on the benefit of surveillance in HNPCC from The Netherlands. The aims were to assess the compliance, the yield of screening, and the occurrence of interval cancers. The study included 43 families which met the Amsterdam criteria. The screening protocol comprises colonoscopy or barium enema in combination with sigmoidoscopy. The interval between examinations is 2-3 years. Approximately 79% of the relatives who could be traced agreed with participation. 8% of them were found to have an adenoma and 3.5% colorectal cancer. About half of the patients with symptomatic colorectal cancer had Dukes C or distant metastases, while most of the screen-detected cases had an early stage of colorectal cancer. However, there were 3 patients who were found to have colorectal cancer within 3 years after the last screening examination. Based on these findings the appropriate screening interval may be 2 years.

Jukka-Pekka Mecklin reported the data from Finland obtained from a 10-year screening programme implemented among 22 HNPCC families (colonoscopy every 2-3 years). The incidence of colorectal cancer and mortality was compared with that in a control group including relatives who could not be traced or relatives who refused screening. It was found that screening resulted in a reduction of colorectal cancer by 62%. Although there was no CRC-related mortality in the screened group the overall mortality was not significantly different. In the Finnish series 2 "interval cancers" were detected within 3 years. Jukka-Pekka Mecklin showed the program which he would recommend in proven gene carriers: for CRC: colonoscopy every 2 years, from the age of 20-25 years; for endometrial cancer: transvaginal sonography and microcurettage from 30-35 years; for urinary tract cancer: urine cytology and abdominal sonography.

5. Collaborative studies

Rectal cancer risk in HNPCC after abdominal colectomy (Miquel Rodriquez-Bigas, Hans Vasen)
The incidence of rectal cancer and adenomatous polyps in HNPCC after abdominal colec- tomy is not known. A questionnaire was sent to 30 ICG-HNPCC members in 18 countries in order to evaluate the incidence of rectal cancer and adenomas in HNPCC patients meeting the Amsterdam criteria after a total abdominal colectomy and ileo-rectal anastomosis (TAC-IR) or a subtotal abdominal colectomy and ileosigmoid anastomosis (STC-IS). Six investigators responded. There were 20 males and 11 females. Nineteen patients underwent TAC-IR and 12 patients underwent STC-IS at a median age of 45 years (range 22-71). These patients have been followed a median of 51 months after TAC-IR or STC-IS (range 1 - 284 months). Sixteen patients had 21 surgical procedures prior to TAC-IR or STC-IS. At the time of TAC-IR or STC-IS 27 out of 31 patients had 34 colorectal carcinomas (CRC) of which 70% were located proximal to the splenic flexure. Sixteen patients (52%) had synchronous adenomas. Of 22 patients who underwent surveillance after abdominal colectomy, 2 patients (9%) developed adenomas in the rectum at 6 and 48 months after TAC-IR. Three patients developed cancer in the rectum during follow up. However, in one the primary tumor site was not clear, thus was exluded. One patient developed a T2N0M0 moderately differentiated mucin producing rectal cancer and a synchronous 2 mm villous adenoma with severe dysplasia 180 months after TAC-IR. He had been lost to follow up for 90 months. The other patient developed a Dukes B moderately differentiated rectal cancer 12 months after a TAC-IR. Thirteen percent (4/41) of the patients developed adenomas and/or carcinomas in the rectum after abdominal colectomy. Even though the riskappears to be small, HNPCC patients should be closely followed for metachronous lesions after abdominal colectomy. Patients accretion should continue.  

Proposal 1: Database of mutation data (Albert de la Chapelle) Albert de la Chapelle suggested that an interesting activity for the ICG would be to collect all mutation data and publish them at regular times in a newsletter. The advantage of such an activity is that at all times an update of such a data base of the ICG is available. He referred to such an activity in cystic fibrosis. The group was very enthousiastic about this proposal. The structure of this collaboration will be further elaborated. The question is how often an update should be published and whether this approach is more effective than comparable systems (McKusik).

Proposal 2: Estimation of the penetrance and survival after CRC in HNPCC (Torben Myrhoj) Aim: To estimate the penetrance of the HNPCC genes in different age groups and possibly adjust the surveillance programme. To study survival after CRC in HNPCC. Background: A lifetable analysis of date of diagnosis (CRC) in children of affected persons in HNPCC-families will provide us with the risk of developing CRC at all ages for first degree relatives, if all children are included. Since the risk of having inherited the gene is 50%, the penetrance can then be calculated. If for example the risk of having CRC at age 45 is found to be 15%, the penetrance at age 45 is 30%. If 50% of the persons have had CRC at age 70, the penetrance is 100% at age 70 and consequently surveillance should stop at age 70. If only a few percent develop CRC before age 30 start of surveillance could be postponed to age 30. If the lifetable analysis is based on a large number of persons from several centers the data will be very valid. With a few additional data a valid survival study can be performed. Material: Children of affected persons from families fulfilling the Amsterdam criteria. Suffi- cient data must be available on all children of included affecteds. Necessary data: Date of birth and status (affected/at risk) of all persons, diagnosis and date of first diagnosis in affecteds, date of death or date of last observation of unaffecteds. Method: A lifetable analysis is performed based on the above mentioned data. The risk of having CRC can be read from the Kaplan-Meyer plot and the penetrance at relevant ages is then calculated. The Danish HNPCC-group will distribute the protocol and registration forms to all interested centers, collect and analyse all data and write a manuscript.

Proposal 3: Surveillance for endometrial cancer (Hans Vasen) Hans Vasen suggested a collaborative study on surveillance for endometrial cancer. He summarized the results of two studies conducted by the ICG on (1) the epidemiology of endometrial cancer, and (2) risk estimation in HNPCC. The following step could be to initiate acollaborative study to assess the benefit of screening and to determine the appropriate screening interval. The protocol which is recommended in The Netherlands includes gynecological examination, transvaginal sonograpy and in the case of abnormalities vacuum microcurettage from the age of 30-35 years at 1-2 years intervals. A randomized controlled trial was advised during the discussion. The first step should be evaluation of results of screening on a national basis; the second step installation of a subcommittie of gynecologists, which should coordinate such a trial.

Proposal 4: Surveillance in gene carriers (Hans Vasen) Colonoscopy in gene carriers, 1 versus 2 years. Problems: Low number of gene carriers at present. Annual examination in gene carriers probably too often and too burdensome. Alternative: Intervals 2 versus 3 years: 3 years probably not appropriate in gene carriers based on Finnish and Dutch data on interval cancers. Possible solution: Just collect data from centers and don't influence the screening intervals. The discussion will be continued.

6. Business meeting

1. Anticancer research issue
At the fourth meeting of the International Collaborative Group on Hereditary Nonpolyposis Colorectal Cancer (ICG-HNPCC) in Greece in October 1992, it was decided to publish a special issue of Anticancer Research devoted entirely to Hereditary Nonpolyposis Colorectal Cancer (HNPCC). The main aim of this effort was to increase awareness of HNPCC among specialists throughout the world. Meera Khan said that the issue will be published very soon and he said that it is dedicated to Henry Lynch because of his immense contributions to the study of HNPCC.

2. New members
Hans Vasen said that last year 11 specialists (J. Bufill, S. Gallinger, Z. Cohen, T. Berk, B. Danes, V. Tassi, J. Giria, R. Madoff, R. Meier, W. Doe and G. Stevenson) showed interest in joining the group. The group unanimously voted for their membership.

3. Next meeting
The next meeting will be organized in Helsinki (Finland) 14-16 September 1994.

4. Closing remarks
Henry Lynch thanked the organizers and adjourned the meeting.