- Email: [email protected]
Cancer family syndrome
GASTROENTEROLOGY 1986:90:3X%-33
Cancer Family Syndrome Genetic Analysis of 22 Finnish Kindreds JUKKA-PEKKA MECKLIN, HEIKKI J. JARVINEN, and PEKKA PELTOKALLIO Second Department
of Surgery, University
Central Hospital, Helsinki, Finland
The genetic properties cind the frequency of cancer family syndrome (CFS) were evaluated on the basis of 22 kindreds identified in Finland by investigating the family histories of young patients with colorectal carcinoma. These families had 196 members with malignancies: 120 colorer&d (Sl%), 30 undefined intraabdominal (15%), 20 endometrial (zo%), 7 gastric (4%), 5 biliary tract (3%) adenocarcinomas, and 24 (7%) other cancers. The present CFS patients represented 0.4% of all colorectal cancers diagnosed in Finland during 1961-1980, and the estimated minimum frequency of CFS was 4-4.7 x lop5 (1 in 25,100-21,400). An autosomal dominant pattern of genetic transmission for colorectal cancer was demonstrated through two to four generations, but the trait was also transferred by mothers with uterine carcinoma. The highest incidence of colorectal carcifioma occurred between 40 and 49 yr of age (15% per decade), and the cumulative risk in the descendants increased to 50% at 69 yr. The penetrance of CFS was 0.5-0.9. The analysis suggests that heredity is more significant in the etiology of colorectal cancer than has been previously believed. Identification of CFS seems to provide a good opportunity for screening the descendants of affected patients for colorectal cancer. Cancer family syndrome (CFS) was first reported by Warthin in 1913 (l), and was characterized in more detail by Lynch et al. (2,s). Its main features include
(a) an autosomal dominant mode of genetic transmission of colorectal and some other adenocarcinomas, (b) mainly proximal location of colorectal tumors, (c) early age of onset, and (d) multiple primary malignancies. Several reports of CFS from different parts of the world support its existence as a real disease entity (4-B), but comprehensive studies (for example, on the genetic features of CFS) are few. Estimat& of the frequency of all hereditary colon cancer types are speculative and range from ~1% of all colorectal carcinomas to 5% (9,lO). Recognition of CFS can be of great importance in providing a basis for screening and follow-up‘&symptomless relatives at risk, because this allows early detection of colorectal cancer, thus improving the prognosis. We have identified 22 Finnish CFS kindreds, one of the largest series reported to date. Therefore, we attempted to evaluate the basic genetic characteristics of CFS: its mode of inheritance, penetrance, and risk of colorectal carcinoma in the descendants of an affected parent. This was considered important in planning screening programs for nonaffected, highrisk family members. We are fully aware that the present CFS kindreds may represent only a fraction of all such kindreds. However, we compared the number of these known CFS cases to all colorectal carcinoma cases in Finland, and estimated the frequency of the CFS phenotype in the Finnish population.
Patients and Methods Received January 14, 1985. Accepted July 3, 1985. Address requests for reprints to: Heikki J. JBrvinen, M.D., Second Department of Surgery, University Central Hospital, 00290 Helsinki, Finland. This work was supported by the Finnish Cancer Foundation, the Yrja Jahnsson Foundation, and the Ida Montin Foundation. The authors thank Dr. R. Norio and Dr. E. Kivilaakso for their critical review of the manuscript, K. Pitkgnen and M. Kirsi for excellent technical assistance, and L. Tarkkonen for statistical advice. 0 1986 by the American Gastroenterological Association 0016-5085/86/$3.50
The majority of the present kindreds fied in a study of all 251 colorectal carcinoma
were identipatients ~40
yr of age, diagnosed in Finland between 1970 and 1979, and entered into the Finnish Cancer Registry (11). When hospital records of these 251 patients noted colorectal cancer irt a first-degree relative, all family members were examined uging parish registers. This included the gathering of biographic and cause of death data. This full Abbreviation
used in this paper: CFS, cancer family syndrome.
February
Table
CANCER
1986
1. Distribution of Malignancies Family Syndrome Kindreds Malignancy
Colon Rectum Intraabdominal uterus Stomach Biliary tract Lung Breast Other Total
(undefined)
Men 64 11 21 -
in 22 Cancer
Women
Total patients (%) 102 18 30 20
4
38 7 9 20 3 3 1 2 4
109
87
196 (100)
4 2 3
7 5 4 2 8
(52)" (9)" (15) (10)" (41 (31 (2) (11 (41
" One hundred
five of the colorectal carcinomas verified histologically. b Fifteen histologically verified endometrial carcinomas, and five undefined carcinomas of the uterus.
examination was performed irrespective of any family history on a smaller regional sample of 20 patients from the province of Central Finland. These examinations resulted in the identification of 12 CFS kindreds. Two others of the present kindreds had been identified earlier, and one of these had been reported in part during the 1960s (5). Since 1979 more emphasis has been placed on family histories of all colorectal carcinoma patients in our department, and eight additional CFS kindreds have been identified during routine clinical procedures. The only criterion of CFS was a demonstration of at least three first-degree relatives with colorectal carcinoma, without familial adenomatosis coli. Other characteristics of CFS were not used as criteria because they are not definitive, even though such features were demonstrable in all of our 22 kindreds. No attempt was made to distinguish between CFS and hereditary site-specific colorectal (12) because no clear criteria for such a division carcinoma are available. In two kindreds with four and five affected members, however, colorectal carcinoma was the only cancer type observed. In the 22 kindreds fulfilling our
Figure I. Pedigrees of cancer family syndrome I, undefined intraabdominal cancer: number of healthy children.
kindreds 0, other
FAMILY
SYNDROME
329
definition, data (including dates of birth and death, cause of death, marital status, and number of children) were examined for all members for four to seven generations using parish registers, until a generation with no incidence of cancer was reached. For all malignant cases thus revealed, accurate verification was achieved in 82% (n = 161) using hospital records: 120 of these cases were also verified histologically. In 18% (n = 35) of the malignant cases, only the cause of death was obtained from the Register of the Central Statistical Office in Finland. The frequency of CFS in Finland during 1961-1980 was calculated according to Read and Neel by the forumla f = P/D, where P is the number of patients with colorectal carcinoma in CFS families who died between 1961 and 1980 and D is the number of all people who died in Finland during the same period (13-15). The number of colorectal cancer patients in CFS families was also compared with the number of all cases of colorectal carcinoma in Finland during 1961-1980 based on the data of the Finnish Cancer Registry (founded in 1952) (16). It has been obligatory since 1961 to report cases of cancer to the Finnish Cancer Registry, and this reporting has been found to be complete (I 7). Cumulative risk and penetrance of cancer were calculated for 14 kindreds. Eight families were excluded because the fate of some of these family members was not known. The cumulative risk of colorectal cancer among children of affected patients was calculated according to Lynch et al. (12,18) by the formula Risk = R, = c
(1 ~ q,]
,=o
where x, is the number of persons affected by colorectal cancer in the ith decade of life, ni is the number of persons under observation and unaffected at the beginning of the ith decade, wi is the number of withdrawals during the ith
1 and 2. 0, female; Cl, male; 0, colorectal cancer; d, endometrial malignancy: 0. , colorectal cancer with other primary malignant
cancer; @and neoplasms;a,
330
MECKLIN ET AL.
5.
GASTROENTEROLOGYVol. 90, No. 2
PO
6.
6.
9.
7.
PO
10.
w
7
13.
Figure 2. Pedigrees of cancer family syndrome kindreds 3-14. See legend in Figure 1
decade, and yij is the number of years remaining in the ith decade for the jth withdrawal. According to Alm and Licznerski (14) the gene penetrance was determined by counting the number of affected and unaffected parents in whose progeny colorectal cancer was diagnosed.
Results Occurrence
of Cancer Family
Syndrome
The present 22 CFS kindreds comprised 196 patients with malignancy. Of these patients, 120 had colorectal carcinoma, 30 had an undefined intraab-
dominal cancer, 20 had carcinoma of the uterus, 7 had gastric carcinoma, 5 had biliary tract cancer, and 14 had other malignancies (Table 1). During 1961-1980, a total of 18,748 colorectal carcinomas were diagnosed in Finland (9929 colon, 8819 rectum). Of these, 75 belonged to the present 22 CFS families, representing a frequency of 0.4% for CFS as compared with all colorectal carcinomas. When colorectal carcinoma alone was included, the evaluated frequency of CFS during the period 1961-1980 in the whole Finnish population (about 5 million inhabitants) on the basis of the present 22 kindreds was 3.99 X 10F5, or 1 in 25,100 If = 35/877,369). If those CFS members with cancer of the
CANCER FAMILY SYNDROME
February 1986
Table 2. Cumulative Risk of Colorectal Carcinoma in the Descendants With 50% Risk in 14 Cancer Family
Syndrome
CumuAge (yrl
Kindreds
No. of affected members
lative No.
10-19 Cl0
2 14 48 107 206 330 390 428 456
0 (0) 0 (0) 3 (111 13 (25) 30 (43) 35 (40) 11 (12) 1 (11 0 (0)
Total
456
93 (132)
>a0 70-79 60-69 50-59 40-49 30-39 20-29
The figures in parentheses intraabdominal cancer.
include
Incidence per decade (%I
Cumulative risk of cancer (%)
0 (01 0 (01 6.2(22.9) 12.1(23.4) 14.6(20.8) 10.6(12.1) 2.8(3.1) 0.2(0.2) 0 (01
49.9(76.9) 49.9(76.9) 49.9(76.9) 44.0(60.9) 32.3(41.2) 11.6(17.7) 3.2(3.5) 0.2(0.2) 0 (0)
endometrial
dominal carcinoma gives a penetrance figure of 89% (40 of 45 cases). In the five instances of failing penetrance, the causes of death of the parents were tonsillitis, pulmonary tuberculosis, gastric cancer, lung cancer, and undefined liver disease, at the ages of 35, 38, 44, 57, and 64 yr, respectively. The total number of all members of the second, third, and fourth generations of the 14 CFS kindreds is given in Table 3. The figures for patients affected by colorectal carcinoma, endometrial carcinoma, or undefined intraabdominal carcinoma were 58%, 26%, and ll%, closely corresponding to a Mendelian dominant mode of inheritance, even though many of the healthy members in the third and fourth generations are still at the risk age.
and undefined
Descendants
uterus were also taken into account, the frequency increased to 4.67 x 10p5, or 1 in 21,400 If = 411877,369). Cumulative
331
Risk of Cancer
The 14 kindreds, of which all descending family branches could be followed-up, comprised 456 family members at risk. The pedigrees of these kindreds are shown in Figures 1 and 2, revealing a total of 151 patients with at least one malignancy. The male to female ratio of all family members was 1:0.85, and this ratio was the same for those with cancer. The mean age of the 196 cancer patients was 44.8 yr; for the 120 patients with colorectal carcinoma it was 41.3 yr. Multiple primary malignancies, ranging from two to eight, occurred in 34 patients (22.5%). The cumulative risk of colorectal cancer and the incidence per decade are given in Table 2. The cumulative risk shows a gradual increase in occurrence from 19 yr up to 50% at the age of 69 yr. Peak incidence (15%) occurred between 40 and 49 yr, and the oldest patient with cancer of the colon was an 83-yr-old man, an ancestor of one of the kindreds. If those cases with undefined intraabdominal cancer and those with endometrial carcinoma were included, the corresponding cumulative risk figures would be 42%, 62%, and 77% at the ages of 49, 59, and 69 yr, respectively.
The CFS patients of the present 22 kindreds have 287 children apparently at 50% risk of contracting colorectal cancer. Of these 287 children, 34 are under 20 yr of age, 242 are between 20 and 60 yr of age, and 11 are over 60 yr of age. Thus, at least 242 adults should be screened at present, and an additional 34 in the future.
Discussion The present report is the first that evaluates the frequency of CFS on a national level. The estimated minimum rate was 0.4% of all colorectal cancers, based on data of 22 kindreds identified in Finland. The frequency of 1 in 21,000 for CFS nearly equals the frequency of familial adenomatous polyposis in Finland (1 in 18,000) (19). Thus, the overall frequency of hereditary colorectal cancer approaches that of Mulvihill (9), who estimated the frequency to be about 1% of all colorectal carcinomas. There are, however, several reasons to believe that our preliminary estimate for the frequency of CFS Table
3. Occurrence of Cancer in 14 Cancer Family Syndrome Kindreds by Generation (All Descending Branches Included) No. of members in each generation Affected
Generation
Penetrance
Total No.
All malignancies
In 21 of 45 cases, a patient with colorectal cancer had a parent with the same tumor, giving rise to a penetrance of 47%. Inclusion of those parents
3 4”
with carcinoma
’Four kindreds only.
of the uterus and undefined
intraab-
at Risk
2
84 222 195
53 (63%) 68 (31%) 25 (13%)
Colorectal, uterine, or intraabdominal undefined carcinoma 49 (58%) 58 (26%) 21 (11%)
Unaffected >60 vr
<60 vr
15 28 7
16 126 163
33.2
MECKLIN
ET AL.
represents a minimum rate. First, rather strict criteria for at least three first-degree relatives with colorectal cancer were used before inclusion in the series because the identification of CFS is not possible on the basis of a single patient. However, solitary cases of CFS must occur, as they occur in familial adenomatosis in about 30%45% of families. In these solitary cases there is still a propensity to transfer the gene to the children (1920). Second, the basic heredity data, which were gathered from hospital records, were often inadequate, and accurate family histories were taken only in those cases with some suggestive evidence of heredity. In a smaller sample comprising the province of Central Finland (0.25 million inhabitants), the frequency of CFS seemed to be much higher because the family history was examined fully in every case. Furthermore, CFS also occurs in patients older than 40 yr of age, who were not screened in this context. Several earlier studies have demonstrated that 16%-26% of the patients with colorectal carcinoma have at least one first-degree relative with the same disease (21-24). These studies were not performed from the standpoint of CFS, and certainly also include coincidental cases or cases in which polygenic inheritance might be in question. Therefore, it is reasonable to assume that the true frequency of CFS is less than the approximate 20% anticipated on the basis of the epidemiologic studies cited above. Further studies to evaluate the incidence of CFS are clearly needed. Nevertheless, we think that our 0.4% minimum rate of all colorectal cancers is a significant preliminary finding, and more accurate than previous estimates. In fact, the study now going on in the province of Central Finland aims at even more accuracy. The transmission of CFS is in good agreement with the inheritance pattern of an autosomai dominant trait suggested by Lynch et al. (10). The sex distribution of the patients was even, and 50% of the descendants of the progenitors were affected by colorectal cancer through two to four generations. In fact, the frequency of all cancers exceeded 50% at the age of 60 yr. This may be explained in part by coincidental cases of cancer in the CFS kindreds, and in part by the selection of the family material. It is clear that families with more affected members are detected easier, whereas those with few affected members remain unidentified. The syndrome seems to have a high penetrance of up to 90%. Failing penetrance was due more to incomplete knowledge of the condition of parents, or to their early death by other causes, than to verified failing penetrance. Although cancer of the colon is by far the most common malignancy in CFS, mothers with endometrial carcinoma also transferred this trait in at least
GASTROENTEROLOGY
Vol. 90, No. 2
seven families described here, and in several cases gastric, uterine, or other cancer occurred simultaneously or metachronously in patients with colorectal carcinoma. Patients with these other malignancies probably belong to the CFS entity. We found it difficult to differentiate subgroups of CFS in the present kindreds such as “hereditary site-specific colon cancer,” as did Lynch et al. (12). The rather common occurrence’ of malignant tumors in CFS, other than colorectal, speaks in favor of a generalized disease or of a metabolic disorder, rather than of a disturbance located in the colon itself. The relatively high prevalence of CFS provides many opportunities for the etiologic consideration of colorectal carcinoma, and its metabolic etiology in particular. Furthermore, the main significance of CFS lies in the fact that it makes possible the screening for colorectal cancer among the descendants of patients identified. The youngest patient with colorectal cancer was a 19yr-old man, and the greatest incidence was observed between 30 and 50 yr of age, which was also the observation in earlier analyses of single families (3,12). Thus, screening examinations should be instituted at 20 yr of age, or soon after, and continued to 60 yr of age. There are already a few reports of successful results (25-27). Such screening programs still pose many difficult questions: what are the examination methods to be used, how often should they be repeated, and what would be the cost-benefit ratio? Due to the predominantly proximal location of colorectal tumors in CFS, colonoscopy or high-quality, double-contrast colography should be used (10). The impact of other cancer types on the screening programs is even more problematic, because they are less common (12,28). On the other hand, the expected benefit of screening for cancer in CFS families gives a much higher yield than screening of the general population-the value of which is still under debate (29), especially as the risk in CFS concerns young adults (27). The present 22 CFS kindreds provided 276 descendants at risk, and they will be used to evaluate these questions.
References AS. Heredity with reference to carcinoma. .Arch 1. Warthin Intern Med 1913;12:546-55. CW, Larsen AL, 2. Lynch HT, Shaw MW, Arbor A, Magnuson Krush AJ. Hereditary factors in cancer. Arch Intern Med 1966;117:206-12. and adenocarcinoma of the 3. Lynch HT, Krush AJ. Heredity colon. Gastroenterology 1967;53:517-27. cancer of the colon. Acta Chir Stand 4. Kluge T. Familial 1964;127:392-8. P, Peltokallio V. Relationship of familial factors to 5. Peltokallio carcinoma of the colon. Dis Colon Rectum 1966;9:367-70. GH, Knaggs TW. Familial cancer of the colon and 6. Dunstone rectum. J Med Genet 1972;9:451-6.
CANCER FAMILY SYNDROME
February 1986
tract. Br J Surg 7. Lovett E. Familial cancer of gastro-intestinal 1976;63:19-22. 8. Sheehan MP, Metzmaker CO. Cancer family syndrome manifested in an extended kindred. Surg Gynecol Obstet 1984; 158:450-6. 9. Mulvihill JJ. The frequency of hereditary large bowel cancer. In: Ingall JR, Mastromarino AJ, eds. Prevention of hereditary large bowel cancer. New York: Alan R. Liss, 198361-75. 10. Lynch HT, Lynch PM, Albano WA, Lynch JF. The cancer family syndrome: a status report. Dis Colon Rectum 1981; 24:311-21. 11. Jarvinen HJ, Turunen MJ. Colorectal cancer before 40 years of age: prognosis and predisposing conditions. Stand J Gastroenter01 1984;19:634-8. 12. Lynch HT, Lynch PM, Lynch JF. What is hereditary colon cancer. In: Ingall JR, Mastromarino AJ, eds. Prevention of hereditary large bowel cancer. New York: Alan R. Liss, 1983: 3-38. 13. Yearbook of the Central Statistical Office of Finland. New Series. Helsinki: Central Statistical Office, 1982:78. 14. Alm T, Licznerski G. The intestinal polyposes. Clin Gastroenter01 1973;2:577-602. 15. Read TE, Neel JV. A genetic study of multiple polyposes of the colon. Am Hum Genet 1955;7:236-63. 16. Finnish Cancer Registry. Cancer incidence in Finland, Helsinki: 1961-80;6-25. 17. Saxen E, Teppo L. Finnish Cancer Registry 1952-1977: twenty-five years of a nationwide cancer registry. Helsinki: Finnish Cancer Registry, 1978. 18. Lynch HT. Harris RE, Organ CH, et al. The surgeon, genetics and cancer control: the cancer family syndrome. Ann Surg 1977;185:435-40.
19.
333
Jarvinen HJ, Husa A, Aukee S, Laitinen S, Matikainen M, Havia T. Finnish registry for familial adenomatosis coli. Stand J Gastroenterol 1984;19:941-6. 20. Bussey HJR. Familial polyp&is coli. Family studies, histopathology, differential diagnosis and results of treatment. London: Johns Hopkins University Press, 1975. 21. Woolf CM. A genetic study of carcinoma of the large intestine. Am Hum Genet 1958;10:42-52. 22. Macklin MT. Inheritance of cancer of the stomach and large intestine in man. J Nat1 Cancer Inst 1960;24:551-71. 23. Lovett E. Family studies in cancer of colon and rectum. Br J Surg 1976;63:13-8. 24. Duncan JL, Kyle J. Family incidence of carcinoma of the colon and rectum in North-East Scotland. Gut 1982;23:169-71. 25. Woolf CM, Richards RC, Gardner EJ. Occasional discrete polyps of the colon and rectum showing an inherited tendency in a kindred. Cancer 1955;8:403-8. 26. Anderson DE, Romsdahl MM. Family history: a criterion for selective screening. In: Mulvihill JJ, Miller RW, Fraumeni JF, eds. Genetics of human cancer. New York: Raven, 1977: 257-62. 27. Lynch HT, Albano WA, Ruma TA, Schmitz GD, Castello KA, Lynch JF. Surveillance/management of an obligate gene carrier: the cancer family syndrome. Gastroenterology 1983;84: 404-8. 28. Sivak MV. Hereditary large bowel cancer: surveillance criteria for at risk patients. In: Ingall JR, Mastromarino AJ, eds. Prevention of hereditary large bowel cancer. New York: Alan R. Liss, 1983:149-56. 29. Simon JB. Occult blood screening for colorectal carcinoma: a critical review. Gastroenterology 1985;88:820-37.
Cancer Family Syndrome Genetic Analysis of 22 Finnish Kindreds JUKKA-PEKKA MECKLIN, HEIKKI J. JARVINEN, and PEKKA PELTOKALLIO Second Department
of Surgery, University
Central Hospital, Helsinki, Finland
The genetic properties cind the frequency of cancer family syndrome (CFS) were evaluated on the basis of 22 kindreds identified in Finland by investigating the family histories of young patients with colorectal carcinoma. These families had 196 members with malignancies: 120 colorer&d (Sl%), 30 undefined intraabdominal (15%), 20 endometrial (zo%), 7 gastric (4%), 5 biliary tract (3%) adenocarcinomas, and 24 (7%) other cancers. The present CFS patients represented 0.4% of all colorectal cancers diagnosed in Finland during 1961-1980, and the estimated minimum frequency of CFS was 4-4.7 x lop5 (1 in 25,100-21,400). An autosomal dominant pattern of genetic transmission for colorectal cancer was demonstrated through two to four generations, but the trait was also transferred by mothers with uterine carcinoma. The highest incidence of colorectal carcifioma occurred between 40 and 49 yr of age (15% per decade), and the cumulative risk in the descendants increased to 50% at 69 yr. The penetrance of CFS was 0.5-0.9. The analysis suggests that heredity is more significant in the etiology of colorectal cancer than has been previously believed. Identification of CFS seems to provide a good opportunity for screening the descendants of affected patients for colorectal cancer. Cancer family syndrome (CFS) was first reported by Warthin in 1913 (l), and was characterized in more detail by Lynch et al. (2,s). Its main features include
(a) an autosomal dominant mode of genetic transmission of colorectal and some other adenocarcinomas, (b) mainly proximal location of colorectal tumors, (c) early age of onset, and (d) multiple primary malignancies. Several reports of CFS from different parts of the world support its existence as a real disease entity (4-B), but comprehensive studies (for example, on the genetic features of CFS) are few. Estimat& of the frequency of all hereditary colon cancer types are speculative and range from ~1% of all colorectal carcinomas to 5% (9,lO). Recognition of CFS can be of great importance in providing a basis for screening and follow-up‘&symptomless relatives at risk, because this allows early detection of colorectal cancer, thus improving the prognosis. We have identified 22 Finnish CFS kindreds, one of the largest series reported to date. Therefore, we attempted to evaluate the basic genetic characteristics of CFS: its mode of inheritance, penetrance, and risk of colorectal carcinoma in the descendants of an affected parent. This was considered important in planning screening programs for nonaffected, highrisk family members. We are fully aware that the present CFS kindreds may represent only a fraction of all such kindreds. However, we compared the number of these known CFS cases to all colorectal carcinoma cases in Finland, and estimated the frequency of the CFS phenotype in the Finnish population.
Patients and Methods Received January 14, 1985. Accepted July 3, 1985. Address requests for reprints to: Heikki J. JBrvinen, M.D., Second Department of Surgery, University Central Hospital, 00290 Helsinki, Finland. This work was supported by the Finnish Cancer Foundation, the Yrja Jahnsson Foundation, and the Ida Montin Foundation. The authors thank Dr. R. Norio and Dr. E. Kivilaakso for their critical review of the manuscript, K. Pitkgnen and M. Kirsi for excellent technical assistance, and L. Tarkkonen for statistical advice. 0 1986 by the American Gastroenterological Association 0016-5085/86/$3.50
The majority of the present kindreds fied in a study of all 251 colorectal carcinoma
were identipatients ~40
yr of age, diagnosed in Finland between 1970 and 1979, and entered into the Finnish Cancer Registry (11). When hospital records of these 251 patients noted colorectal cancer irt a first-degree relative, all family members were examined uging parish registers. This included the gathering of biographic and cause of death data. This full Abbreviation
used in this paper: CFS, cancer family syndrome.
February
Table
CANCER
1986
1. Distribution of Malignancies Family Syndrome Kindreds Malignancy
Colon Rectum Intraabdominal uterus Stomach Biliary tract Lung Breast Other Total
(undefined)
Men 64 11 21 -
in 22 Cancer
Women
Total patients (%) 102 18 30 20
4
38 7 9 20 3 3 1 2 4
109
87
196 (100)
4 2 3
7 5 4 2 8
(52)" (9)" (15) (10)" (41 (31 (2) (11 (41
" One hundred
five of the colorectal carcinomas verified histologically. b Fifteen histologically verified endometrial carcinomas, and five undefined carcinomas of the uterus.
examination was performed irrespective of any family history on a smaller regional sample of 20 patients from the province of Central Finland. These examinations resulted in the identification of 12 CFS kindreds. Two others of the present kindreds had been identified earlier, and one of these had been reported in part during the 1960s (5). Since 1979 more emphasis has been placed on family histories of all colorectal carcinoma patients in our department, and eight additional CFS kindreds have been identified during routine clinical procedures. The only criterion of CFS was a demonstration of at least three first-degree relatives with colorectal carcinoma, without familial adenomatosis coli. Other characteristics of CFS were not used as criteria because they are not definitive, even though such features were demonstrable in all of our 22 kindreds. No attempt was made to distinguish between CFS and hereditary site-specific colorectal (12) because no clear criteria for such a division carcinoma are available. In two kindreds with four and five affected members, however, colorectal carcinoma was the only cancer type observed. In the 22 kindreds fulfilling our
Figure I. Pedigrees of cancer family syndrome I, undefined intraabdominal cancer: number of healthy children.
kindreds 0, other
FAMILY
SYNDROME
329
definition, data (including dates of birth and death, cause of death, marital status, and number of children) were examined for all members for four to seven generations using parish registers, until a generation with no incidence of cancer was reached. For all malignant cases thus revealed, accurate verification was achieved in 82% (n = 161) using hospital records: 120 of these cases were also verified histologically. In 18% (n = 35) of the malignant cases, only the cause of death was obtained from the Register of the Central Statistical Office in Finland. The frequency of CFS in Finland during 1961-1980 was calculated according to Read and Neel by the forumla f = P/D, where P is the number of patients with colorectal carcinoma in CFS families who died between 1961 and 1980 and D is the number of all people who died in Finland during the same period (13-15). The number of colorectal cancer patients in CFS families was also compared with the number of all cases of colorectal carcinoma in Finland during 1961-1980 based on the data of the Finnish Cancer Registry (founded in 1952) (16). It has been obligatory since 1961 to report cases of cancer to the Finnish Cancer Registry, and this reporting has been found to be complete (I 7). Cumulative risk and penetrance of cancer were calculated for 14 kindreds. Eight families were excluded because the fate of some of these family members was not known. The cumulative risk of colorectal cancer among children of affected patients was calculated according to Lynch et al. (12,18) by the formula Risk = R, = c
(1 ~ q,]
,=o
where x, is the number of persons affected by colorectal cancer in the ith decade of life, ni is the number of persons under observation and unaffected at the beginning of the ith decade, wi is the number of withdrawals during the ith
1 and 2. 0, female; Cl, male; 0, colorectal cancer; d, endometrial malignancy: 0. , colorectal cancer with other primary malignant
cancer; @and neoplasms;a,
330
MECKLIN ET AL.
5.
GASTROENTEROLOGYVol. 90, No. 2
PO
6.
6.
9.
7.
PO
10.
w
7
13.
Figure 2. Pedigrees of cancer family syndrome kindreds 3-14. See legend in Figure 1
decade, and yij is the number of years remaining in the ith decade for the jth withdrawal. According to Alm and Licznerski (14) the gene penetrance was determined by counting the number of affected and unaffected parents in whose progeny colorectal cancer was diagnosed.
Results Occurrence
of Cancer Family
Syndrome
The present 22 CFS kindreds comprised 196 patients with malignancy. Of these patients, 120 had colorectal carcinoma, 30 had an undefined intraab-
dominal cancer, 20 had carcinoma of the uterus, 7 had gastric carcinoma, 5 had biliary tract cancer, and 14 had other malignancies (Table 1). During 1961-1980, a total of 18,748 colorectal carcinomas were diagnosed in Finland (9929 colon, 8819 rectum). Of these, 75 belonged to the present 22 CFS families, representing a frequency of 0.4% for CFS as compared with all colorectal carcinomas. When colorectal carcinoma alone was included, the evaluated frequency of CFS during the period 1961-1980 in the whole Finnish population (about 5 million inhabitants) on the basis of the present 22 kindreds was 3.99 X 10F5, or 1 in 25,100 If = 35/877,369). If those CFS members with cancer of the
CANCER FAMILY SYNDROME
February 1986
Table 2. Cumulative Risk of Colorectal Carcinoma in the Descendants With 50% Risk in 14 Cancer Family
Syndrome
CumuAge (yrl
Kindreds
No. of affected members
lative No.
10-19 Cl0
2 14 48 107 206 330 390 428 456
0 (0) 0 (0) 3 (111 13 (25) 30 (43) 35 (40) 11 (12) 1 (11 0 (0)
Total
456
93 (132)
>a0 70-79 60-69 50-59 40-49 30-39 20-29
The figures in parentheses intraabdominal cancer.
include
Incidence per decade (%I
Cumulative risk of cancer (%)
0 (01 0 (01 6.2(22.9) 12.1(23.4) 14.6(20.8) 10.6(12.1) 2.8(3.1) 0.2(0.2) 0 (01
49.9(76.9) 49.9(76.9) 49.9(76.9) 44.0(60.9) 32.3(41.2) 11.6(17.7) 3.2(3.5) 0.2(0.2) 0 (0)
endometrial
dominal carcinoma gives a penetrance figure of 89% (40 of 45 cases). In the five instances of failing penetrance, the causes of death of the parents were tonsillitis, pulmonary tuberculosis, gastric cancer, lung cancer, and undefined liver disease, at the ages of 35, 38, 44, 57, and 64 yr, respectively. The total number of all members of the second, third, and fourth generations of the 14 CFS kindreds is given in Table 3. The figures for patients affected by colorectal carcinoma, endometrial carcinoma, or undefined intraabdominal carcinoma were 58%, 26%, and ll%, closely corresponding to a Mendelian dominant mode of inheritance, even though many of the healthy members in the third and fourth generations are still at the risk age.
and undefined
Descendants
uterus were also taken into account, the frequency increased to 4.67 x 10p5, or 1 in 21,400 If = 411877,369). Cumulative
331
Risk of Cancer
The 14 kindreds, of which all descending family branches could be followed-up, comprised 456 family members at risk. The pedigrees of these kindreds are shown in Figures 1 and 2, revealing a total of 151 patients with at least one malignancy. The male to female ratio of all family members was 1:0.85, and this ratio was the same for those with cancer. The mean age of the 196 cancer patients was 44.8 yr; for the 120 patients with colorectal carcinoma it was 41.3 yr. Multiple primary malignancies, ranging from two to eight, occurred in 34 patients (22.5%). The cumulative risk of colorectal cancer and the incidence per decade are given in Table 2. The cumulative risk shows a gradual increase in occurrence from 19 yr up to 50% at the age of 69 yr. Peak incidence (15%) occurred between 40 and 49 yr, and the oldest patient with cancer of the colon was an 83-yr-old man, an ancestor of one of the kindreds. If those cases with undefined intraabdominal cancer and those with endometrial carcinoma were included, the corresponding cumulative risk figures would be 42%, 62%, and 77% at the ages of 49, 59, and 69 yr, respectively.
The CFS patients of the present 22 kindreds have 287 children apparently at 50% risk of contracting colorectal cancer. Of these 287 children, 34 are under 20 yr of age, 242 are between 20 and 60 yr of age, and 11 are over 60 yr of age. Thus, at least 242 adults should be screened at present, and an additional 34 in the future.
Discussion The present report is the first that evaluates the frequency of CFS on a national level. The estimated minimum rate was 0.4% of all colorectal cancers, based on data of 22 kindreds identified in Finland. The frequency of 1 in 21,000 for CFS nearly equals the frequency of familial adenomatous polyposis in Finland (1 in 18,000) (19). Thus, the overall frequency of hereditary colorectal cancer approaches that of Mulvihill (9), who estimated the frequency to be about 1% of all colorectal carcinomas. There are, however, several reasons to believe that our preliminary estimate for the frequency of CFS Table
3. Occurrence of Cancer in 14 Cancer Family Syndrome Kindreds by Generation (All Descending Branches Included) No. of members in each generation Affected
Generation
Penetrance
Total No.
All malignancies
In 21 of 45 cases, a patient with colorectal cancer had a parent with the same tumor, giving rise to a penetrance of 47%. Inclusion of those parents
3 4”
with carcinoma
’Four kindreds only.
of the uterus and undefined
intraab-
at Risk
2
84 222 195
53 (63%) 68 (31%) 25 (13%)
Colorectal, uterine, or intraabdominal undefined carcinoma 49 (58%) 58 (26%) 21 (11%)
Unaffected >60 vr
<60 vr
15 28 7
16 126 163
33.2
MECKLIN
ET AL.
represents a minimum rate. First, rather strict criteria for at least three first-degree relatives with colorectal cancer were used before inclusion in the series because the identification of CFS is not possible on the basis of a single patient. However, solitary cases of CFS must occur, as they occur in familial adenomatosis in about 30%45% of families. In these solitary cases there is still a propensity to transfer the gene to the children (1920). Second, the basic heredity data, which were gathered from hospital records, were often inadequate, and accurate family histories were taken only in those cases with some suggestive evidence of heredity. In a smaller sample comprising the province of Central Finland (0.25 million inhabitants), the frequency of CFS seemed to be much higher because the family history was examined fully in every case. Furthermore, CFS also occurs in patients older than 40 yr of age, who were not screened in this context. Several earlier studies have demonstrated that 16%-26% of the patients with colorectal carcinoma have at least one first-degree relative with the same disease (21-24). These studies were not performed from the standpoint of CFS, and certainly also include coincidental cases or cases in which polygenic inheritance might be in question. Therefore, it is reasonable to assume that the true frequency of CFS is less than the approximate 20% anticipated on the basis of the epidemiologic studies cited above. Further studies to evaluate the incidence of CFS are clearly needed. Nevertheless, we think that our 0.4% minimum rate of all colorectal cancers is a significant preliminary finding, and more accurate than previous estimates. In fact, the study now going on in the province of Central Finland aims at even more accuracy. The transmission of CFS is in good agreement with the inheritance pattern of an autosomai dominant trait suggested by Lynch et al. (10). The sex distribution of the patients was even, and 50% of the descendants of the progenitors were affected by colorectal cancer through two to four generations. In fact, the frequency of all cancers exceeded 50% at the age of 60 yr. This may be explained in part by coincidental cases of cancer in the CFS kindreds, and in part by the selection of the family material. It is clear that families with more affected members are detected easier, whereas those with few affected members remain unidentified. The syndrome seems to have a high penetrance of up to 90%. Failing penetrance was due more to incomplete knowledge of the condition of parents, or to their early death by other causes, than to verified failing penetrance. Although cancer of the colon is by far the most common malignancy in CFS, mothers with endometrial carcinoma also transferred this trait in at least
GASTROENTEROLOGY
Vol. 90, No. 2
seven families described here, and in several cases gastric, uterine, or other cancer occurred simultaneously or metachronously in patients with colorectal carcinoma. Patients with these other malignancies probably belong to the CFS entity. We found it difficult to differentiate subgroups of CFS in the present kindreds such as “hereditary site-specific colon cancer,” as did Lynch et al. (12). The rather common occurrence’ of malignant tumors in CFS, other than colorectal, speaks in favor of a generalized disease or of a metabolic disorder, rather than of a disturbance located in the colon itself. The relatively high prevalence of CFS provides many opportunities for the etiologic consideration of colorectal carcinoma, and its metabolic etiology in particular. Furthermore, the main significance of CFS lies in the fact that it makes possible the screening for colorectal cancer among the descendants of patients identified. The youngest patient with colorectal cancer was a 19yr-old man, and the greatest incidence was observed between 30 and 50 yr of age, which was also the observation in earlier analyses of single families (3,12). Thus, screening examinations should be instituted at 20 yr of age, or soon after, and continued to 60 yr of age. There are already a few reports of successful results (25-27). Such screening programs still pose many difficult questions: what are the examination methods to be used, how often should they be repeated, and what would be the cost-benefit ratio? Due to the predominantly proximal location of colorectal tumors in CFS, colonoscopy or high-quality, double-contrast colography should be used (10). The impact of other cancer types on the screening programs is even more problematic, because they are less common (12,28). On the other hand, the expected benefit of screening for cancer in CFS families gives a much higher yield than screening of the general population-the value of which is still under debate (29), especially as the risk in CFS concerns young adults (27). The present 22 CFS kindreds provided 276 descendants at risk, and they will be used to evaluate these questions.
References AS. Heredity with reference to carcinoma. .Arch 1. Warthin Intern Med 1913;12:546-55. CW, Larsen AL, 2. Lynch HT, Shaw MW, Arbor A, Magnuson Krush AJ. Hereditary factors in cancer. Arch Intern Med 1966;117:206-12. and adenocarcinoma of the 3. Lynch HT, Krush AJ. Heredity colon. Gastroenterology 1967;53:517-27. cancer of the colon. Acta Chir Stand 4. Kluge T. Familial 1964;127:392-8. P, Peltokallio V. Relationship of familial factors to 5. Peltokallio carcinoma of the colon. Dis Colon Rectum 1966;9:367-70. GH, Knaggs TW. Familial cancer of the colon and 6. Dunstone rectum. J Med Genet 1972;9:451-6.
CANCER FAMILY SYNDROME
February 1986
tract. Br J Surg 7. Lovett E. Familial cancer of gastro-intestinal 1976;63:19-22. 8. Sheehan MP, Metzmaker CO. Cancer family syndrome manifested in an extended kindred. Surg Gynecol Obstet 1984; 158:450-6. 9. Mulvihill JJ. The frequency of hereditary large bowel cancer. In: Ingall JR, Mastromarino AJ, eds. Prevention of hereditary large bowel cancer. New York: Alan R. Liss, 198361-75. 10. Lynch HT, Lynch PM, Albano WA, Lynch JF. The cancer family syndrome: a status report. Dis Colon Rectum 1981; 24:311-21. 11. Jarvinen HJ, Turunen MJ. Colorectal cancer before 40 years of age: prognosis and predisposing conditions. Stand J Gastroenter01 1984;19:634-8. 12. Lynch HT, Lynch PM, Lynch JF. What is hereditary colon cancer. In: Ingall JR, Mastromarino AJ, eds. Prevention of hereditary large bowel cancer. New York: Alan R. Liss, 1983: 3-38. 13. Yearbook of the Central Statistical Office of Finland. New Series. Helsinki: Central Statistical Office, 1982:78. 14. Alm T, Licznerski G. The intestinal polyposes. Clin Gastroenter01 1973;2:577-602. 15. Read TE, Neel JV. A genetic study of multiple polyposes of the colon. Am Hum Genet 1955;7:236-63. 16. Finnish Cancer Registry. Cancer incidence in Finland, Helsinki: 1961-80;6-25. 17. Saxen E, Teppo L. Finnish Cancer Registry 1952-1977: twenty-five years of a nationwide cancer registry. Helsinki: Finnish Cancer Registry, 1978. 18. Lynch HT. Harris RE, Organ CH, et al. The surgeon, genetics and cancer control: the cancer family syndrome. Ann Surg 1977;185:435-40.
19.
333
Jarvinen HJ, Husa A, Aukee S, Laitinen S, Matikainen M, Havia T. Finnish registry for familial adenomatosis coli. Stand J Gastroenterol 1984;19:941-6. 20. Bussey HJR. Familial polyp&is coli. Family studies, histopathology, differential diagnosis and results of treatment. London: Johns Hopkins University Press, 1975. 21. Woolf CM. A genetic study of carcinoma of the large intestine. Am Hum Genet 1958;10:42-52. 22. Macklin MT. Inheritance of cancer of the stomach and large intestine in man. J Nat1 Cancer Inst 1960;24:551-71. 23. Lovett E. Family studies in cancer of colon and rectum. Br J Surg 1976;63:13-8. 24. Duncan JL, Kyle J. Family incidence of carcinoma of the colon and rectum in North-East Scotland. Gut 1982;23:169-71. 25. Woolf CM, Richards RC, Gardner EJ. Occasional discrete polyps of the colon and rectum showing an inherited tendency in a kindred. Cancer 1955;8:403-8. 26. Anderson DE, Romsdahl MM. Family history: a criterion for selective screening. In: Mulvihill JJ, Miller RW, Fraumeni JF, eds. Genetics of human cancer. New York: Raven, 1977: 257-62. 27. Lynch HT, Albano WA, Ruma TA, Schmitz GD, Castello KA, Lynch JF. Surveillance/management of an obligate gene carrier: the cancer family syndrome. Gastroenterology 1983;84: 404-8. 28. Sivak MV. Hereditary large bowel cancer: surveillance criteria for at risk patients. In: Ingall JR, Mastromarino AJ, eds. Prevention of hereditary large bowel cancer. New York: Alan R. Liss, 1983:149-56. 29. Simon JB. Occult blood screening for colorectal carcinoma: a critical review. Gastroenterology 1985;88:820-37.