Hereditary Nonpolyposis Colorectal Cancer (Lynch Syndromes I and 11)

I. Clinical Description of Resource

HENRY T. LYNCH, MD,' WILLIAM KIMBERLING, PHD,t WILLIAM A. ALBANO, MD$ JANE F. LYNCH, BSN,' KAREN BISCONE, BSN,' GUY S. SCHUELKE, PHD,' AVERY A. SANDBERG. MD,I) MARTIN LIPKIN, MD,ll ELEANOR E. DESCHNER. PHD,# YVES B. MIKOL, P H D , ~ ROBERT C. ELSTON, PHD," JOAN E. BAILEY-WILSON, PHD" AND B. SHANNON DANES, MD, PHDtt

Hereditary nonpolyposis colorectal cancer (HNPCC) is comprised of the following: (1) the cancer family syndrome (CFS), or Lynch syndrome 11, which shows early-onset proximal colonic cancer predominance and other associated extracolonic adenocarcinomas, particularly endometrial carcinoma; and (2) hereditary site-specific colon cancer (HSSCC), or Lynch syndrome I, which shows all of the same characteristics, except for extracolonic cancer. Nine families with CFS and two with HSSCC provided the resource that was tested for biomarkers (see companion article). All families were meticulously evaluated for genealogy and cancer verification. Biologic specimens were obtained during field visits to areas of closest geographic proximity to the families. Cancer education and recommendations for surveillance/management were provided to patients and their physicians. Additionally, 40 families (about 3000 individuals) with either CFS or HSSCC have been ascertained. Syndrome cancers were restricted to direct-line relatives as opposed to nonbloodline relatives, arguing against involvement of environmental factors. One documented clinical feature was a predilection for proximal versus distal colonic cancer in both CFS and HSSCC kindreds. This has important clinical significance in that it clarifies the need for instituting effective surveillance earlier to detect the predominantly proximal colonic cancers.
Cancer 56934-938. 1985.

UNITED STATESand many other Western countries, colorectal cancer is second in incidence only to carcinoma of the lung. In spite of advances in the treatment of colorectal cancer, little change in survival from this disease has been accomplished during the last 3 decades. Needed, therefore, are improved methods for its earlier detection. This could be greatly facilitated through the identification of patients who are at inordinately high risk for colorectal cancer. Such patients could then be systematically entered into specific surN THE

From the Departments of *Preventive Medicine/Public Health, tOtolaryngology, and +Surgery, Creighton University School of Medicine and the Hereditary Cancer Institute, Omaha, Nebraska, the IIDepartments of Medicine and Genetics, Roswell Park Memorial Institute, Buffalo, the 7Laboratory of Gastrointestinal Cancer Research, and #Laboratory of Digestive Tract Carcinogenesis, Memorial Sloan-Kettering Cancer Center, New York, New York, the **Department of Biometry, Louisiana State University Medical Center, New Orleans, Louisiana, and the TtLaboratory for Cell Biology, Department of Medicine, Cornell University Medical College, New York, New York. Supported by NCI Grant #5 ROl-CA 27831-03 and 5 ROI-CA 28198-05. Q The authors dedicate these articles to the memory of their friend and colleague, William A. Albano, MD, who died July 7. 1983. in the prime of his life at the age of 37. Dr. Albano was active in every phase of this research project, and was loved and respected by all the authors. Accepted for publication September 4, 1984.

veillance/management programs so that acceptable costbenefit yields would be obtainable. Bloodline relatives of patients with any of the hereditary predispositions to colorectal carcinoma may be at as much as a 50% risk for this disease, and are therefore ideal candidates for such surveillance/management programs.' This article describes the clinical-genetic characteristics of 11 extended kindreds manifesting hereditary nonpolyposis colorectal cancer (HNPCC), which provided the subject resource for biomarker determinations aimed at identification of gene carriers as discussed in our companion article (Part 11). Material and Methods
ClassiJicationof Hereditary Nonpolyposis Colon Cancer

Families in the HNPCC category are recognized by the following clinical features: (1) early age at onset of colorectal cancer; (2) proximal colonic cancer excess; (3) increased frequency of multiple primary cancers, including carcinoma of the endometrium; and (4) improved 5-year survival of patients with colon cancer who were compared by appropriate staging with the

934

No. 4

GENETICS, CANCER, AND HNPCC

Lynch ef al.

935

American College of Surgeons long-term audit series.- Families showing these criteria, inclusive of multiple cases of endometrial cancer, are classified as cancer family syndrome (CFS). A subcategory of this second group is hereditary site-specific nonpolyposis colonic cancer (HSSCC). These families show the same criteria as CFS, except that the primary cancer target organ is exclusively the colon.6 Historically, the first description of a family that fulfilled the criteria of CFS was Family G, first encountered by Warthin in 1895 and described by him in 1913. Family G has been updated intermittently over the years, most recently by Lynch and Krush. The first family with hereditary nonpolyposis colorectal cancer that we described was Family N.v3 Ascertainment of the proband was by local physician referral in 1963. This patient was convinced that he was going to die of cancer because all of my relatives eventually die of this disease. He eventually did die of adrenal cortical carcinoma at age 44. For the last 20 years, Family N has served as a prototype model of the CFS. Additionally, we have ascertained more than 40 families composed of about 3000 individuals with either CFS (Lynch syndrome 11) or HSSCC (Lynch syndrome I). These clinical entities have been documented by others within all racial and many national and ethnic f groups represented (for extensive review, see Lynch e al. Eleven families in this resource were selected for intensive biomarker and genetic studies because of their availability, cooperativeness, and size. The current report will focus on the clinical-genetic characteristics of these I I families, 9 of which form the basis for the majority of our biomarker determinations (Part 11). Of these latter 9, seven families displayed criteria of CFS, and the remaining two showed criteria for HSSCC.

TABLE 1 . Sex and Cancer Status of Living and Deceased Members of the I I Nonpolyposis Families Studied.
Unaffected M
CFS Living and deceased Unrelated Group I t Group 2 Group 3
All relatives

Affected M
F

Total

Total

253 120 105

101

217
110

103 95 308

470 230 208 196 634

4 25 0 0

9 34$ 0 0 34

13 (2.7%) SS(25.696) 0 0 59

326

25

HSSCC Living and deceased Unrelated Group I Group 2 Group 3

All relatives

32 77
105

37 70 80
114

94 276

69 147 185 208

540

0 21 0 0

0 13 0 0 13

0 34(18.8%) 0 0 34

264

21

The sex and cancer status (affected/unaffected) of living and deceased first- and second-degree relatives, as well as unrelated relatives from nine families with the cancer family syndrome (CFS) and two families with hereditary site-specific colorectal cancer (HSSCC). A n individual who married into the family or a relative of someone who married into the family. t Group I consists of cancer-affected individuals and their siblings; a Group 2 relative is defined as a child or sibling of an affected gene canier; and a Group 3 relative would be a child of any person in Group 2 or a sibling of an affected proven gene carrier. Probands and their ancestors are not included. $ Nine with endometrium, 19 with colon cancer, and 6 with both types of cancer.

Family Ascertainmenl
Sources of ascertainment have been variable: ( I ) physician referral, primarily because of recognition of a marked excess of cancer within a kindred: (2) initiation of referral by a patient because of concern about an excess of cancer in the family; (3) discovery of the family through review of a consecutive series of medical records of patients with carcinoma of the colon or endometrium; and (4) recognition of a family in an Oncology Clinic where detailed family histories were generated on a consecutive series of patients being treated for cancer of all anatomic sites.

followed by medical-genetic questionnaires that contain signed permission forms. This then provides for the release of primary medical and pathology documents. Formal educational sessions take place during field visits, when biologic samples are collected. These sessions provide an opportunity to discuss all facets of the cancer problem as it relates to risk factors in the kindred and, at the same time, to discuss the importance of highly targeted surveillance programs. These sessions stress the need for close liaison with the family physician. We provide the patients physician with information about these counseling sessions and, at the same time, describe the most current cancer screening measures that should be employed for high risk patients.

Logislies of Data Collection

We have developed a highly detailed protocol for study of genealogy and cancer verification (all sites) among cancer-prone kindreds. This involves detailed interviews of probands and all key available relatives,

Biologic Sample Collectionsfor Biomarker Srudies Small (about 5-mm) rectal mucosal biopsy specimens were obtained from selected patients. These were used for tritiated thymidine uptake labeling of the colonic crypt compartments. Blood was collected for: ( 1 ) linkage studies; (2) cytogenetics on peripheral lymphocytes; and (3) immunologic studies, most notably, quantitative serum i m m u n ~ g l o b u l i n s . ~ Aliquots ~ - ~ ~ of contingency samples were also stored for future use. Skin biopsy specimens were obtained for studies oE ( I ) tetraploidy in dermal fibroblast monolayer culture^'^ and (2) methionine dependency.16

936

CANCER August I5 1985

TABLE 2. Cancer Status of Living Relatives From the I I Nonpolyposis Families Studied Unaffected Affected
M

Vol. 56

M
CFS Living only Unrelated* Group It Group 2 Group 3
All relatives HSSCC Living only Unrelated* Group I t Group 2 Group 3 All relatives

Total

Total

or monthly and all possible errors were routed to the person who entered the data. The computer checked the completeness of data and provided lists of missing information that might be obtained by a telephone call or a personal visit.
Results

200 93 101 95 289

188 96 98 91 285

388 189 199 186 574

8 0 0

2 13 0 0 13

3 (0.8%) 21 (11.1%) 0 0 21

25

85
100 89 274

33 64 78 110 252

58 149 178 199 526

0 9 0 0
9

0 5 0 0
5

0 14(9.3%) 0 0
14

Includes persons who married into the family and, in w m e cases. their relatives. t Group I consists of cancer-affected individuals and their siblings; a Group 2 relative is defined as a child or sibling of an affected gene carrier; and a Group 3 relative would be a child of any person in Group 2 or a sibling of an affected proven gene carrier. Probands and their ancestors are not included. CFS: cancer family syndrome: HSSCC: hereditary site-specific colorectal cancer.

Data Management
The collection, storage, updating, and retrieval of all the clinical and laboratory data for the families was done using the MEGADATS data management system..* MEGADATS (Medical Genetics Acquisition and Data Transfer System) was developed at the Indiana University Medical Genetics Department to manage collection, storage, retrieval, and display of pedigree data. This system allows for flexibility and ease of data entry, editing, and analysis. It enables all of the data files to be fully encrypted by the MEGADATS system to provide a maximal degree of confidentiality. A set of housekeeping programs checked for internal consistency of data. The housekeeping programs were run weekly
TABLE 3. Colon Cancer by Anatomic Site Within the Cancer Family Syndrome and Hereditary Site-Specific Colorectal Cancer Kindreds
No.

Percent 28.2
11.7

Cecum Ascending colon Hepatic flexure Transverse colon Splenic flexure Descending colon Sigmoid Rectum Total no. Proximal Distal

27 17 4
17 1

4.2
17.7 1.o

Table 1 provides a breakdown of the 1 1 subject families. Nine of these fulfilled criteria for the CFS and two for the HSSCC syndrome. Affected is defined for CFS as having colon and/or endometrial cancer and for HSSCC as having colon cancer without endometrial cancer. There is a notable excess of cancer in the firstdegree relatives of CFS and HSSCC kindreds, and a paucity in the nonbloodline relatives. Table 2 focuses on living bloodline and nonbloodline relatives and their cancer status. This table demonstrates the magnitude of the resource and the number of individuals available for marker testing. Table 3 details the distribution of colorectal cancer throughout the colon by specific anatomic locations. There was a marked excess of proximal colonic involvement, with the largest concentration occumng in the cecum. No significant difference was observed in colon cancer location between CFS and HSSCC affecteds. When the significance of proximal colonic cancer in HNPCC was identified, it was based on population expectations for anatomic site distribution. Herein, leftsided involvement was predominant in the general population. Table 4 provides a breakdown by age and sex for development of carcinoma of the colon or endometrium in members from CFS kindreds. This table clearly shows the peak age of onset for manifestation of cancer to be between 40 and 50 years in both men and women. Furthermore, once kindred members reach age 60, it becomes less likely that they will present with the syndrome cancers. This observation is distinct from age distributions for the respective sporadic counterparts of colon and endometrial carcinoma.20 Table 5 depicts the age of onset of colonic cancer in relatives at risk for H S S C C . As in Table 4, there is a strikingly similar pattern for both men and women wherein patients between the ages of 40 and 50 are most vulnerable for clinically recognizable colorectal cancer. Once these kindred members reach age 60, their likelihood of having syndrome cancers also diminishes remarkably.
Discussion

5 II
14

5.2
11.5

14.6

96 66 30 68.8 31.2

Patients lacking premonitory clinical signs of hereditary colonic cancer are seen more frequently than their polyposis counterparts (Fig. They pose a profoundly vexing problem to the cancer geneticist, oncologist, and primary care physician, and are the major focus of our concern.

No. 4

GENETICS,

CANCER, AND HNPCC

.
F

Lynch et al.

937

TABLE 4. Family Members at Risk of Cancer Family Syndrome by Age, Sex, and Phenotypic Status M Age group (yr)
0-4 No.

Both sexes Freq No.

I1

Aff

Freq
O.OO0

No. 6 0 3 5
11

Aff

Aff

Freq
0.000 0.000 0.000 0.000 0.000

5-9 10-14 15-19 20-24 25-29 30-34 35-39

40-44

2
5 6 10 15 13 19 12 13
15

0 0 0 0 0 2 2 6 3
5

O.OO0 O.OO0 O.OO0 O.OO0

0. I33 0. I54 0.316 0.250 0.385 0.267

13 24 16 14
17

45-49 50-54 55-59 60-64 65-69 70-74 75-79 80-84 85-89

Total

8 6 6 3 I 4 2 I45

4 0
1

O.OO0
0. I67
O.OO0

0 0 0 1
1

O.OO0 O.OO0
0.250 0.500 0.172

9 8 3 4 5 2 3
1
144

0 0 0 0 0 2 5 5 8 10 2 I 0 0 I 0 0 0 34

O.OO0 O.OO0 O.OO0 O.OO0 O.OO0

0. I54 0.208 0.3 13 0.57 I 0.588 0.222 0.125

O.OO0 O.OO0
0.200

2 8 I1 21 28 37 35 26 30 24 16 9
10

0 0 0 0 0 4 7 I1
11

15 6
1

O.OO0 O.OO0 O.OO0

0.236

8 3 7 3 289

I 0 I 0
1

0.143 0.189 0.314 0.423 0.500 0.250 0.063 0.111

0.000

0.125
0.000

0.143 0.333 0.204

25

59

Aff; affected; Freq: frequency.

The findings in Tables 1 through 5 clearly define several of the clinical facets of the natural history of hereditary nonpolyposis colon cancer that distinguish it from sporadic forms of this disease. There was a striking difference in syndrome cancers; i.e., excess of colon cancer and endometrial cancer in CFS and site-specific colon cancer in HSSCC kindreds, when compared with observed cancer frequencies in nonbloodline relatives (Tables I and 2). In Table 1, syndrome cancers are also restricted to direct-line relatives as opposed to nonbloodline relatives. This argues against primary environmental factors being responsible for the cancers observed.

As in previous studies, Table 3 shows an excess of proximal colonic involvement, with the greatest frequency occumng in the cecum, among the CFS and HSSCC first-degree relatives. This has important clinical significance in that it mandates the need for instituting colonoscopy and/or double-air-contrast barium enema at an early age to detect proximal colonic cancer since this is the location where disease will usually be manifested initially. The findings in Tables 4 and 5 clearly show that the most vulnerable age for expression of hereditary forms of colon and endometrial cancer in the hereditary subset is between the ages of 40 and 50. Subclinical expression

TABLE 5. Individuals at Risk for Hereditary Site-Specific Colorectal Cancer by Age Group, Sex, and Phenotypic Status M Age group
0 4

F
Freq O.OO0

Both sexes
Freq
O.OO0

No.

Aff
0 0 0 0 1 1 6
1

No.
I
0

Aff

No.

Aff

FW 0.000 0.000 0.000 0.000 0.030 0.133 0.348 0.222 0.571 0.333 0.500 0.167 0.000 0.000 0.200 0.143 0.000

5-9 10-14 15-19 20-24 25-29 30-34 35-39 4044 45-49 50-54 55-59 60-64 65-69 70-74 75-79 80-84
Total

2 0 3 9 16 11 14 6 8 9 8 2 4 I I 4 0 98

0
0

O.Oo0 O.Oo0 O.Oo0

0.063 0.09 1 0.429 0.167 0.625 0.222 0.500

5 2 4 0 0 0
0

3 II 17 4 9 3 6 6

0 0 0
1

O.Oo0
O.OO0 O.OO0 O.OO0

2 4
5 0 4 3 5

2 I 3 3 I
1

O.OO0 O.OO0 O.OO0 O.OO0 0.250 0.222 0.333 0.500 0.500 0.500 0.250
O.OO0 O.OO0

3 0 6 20 33 15 23 9 14
15 10

0 0 0 0 I 2 8 2 8
5

0
0

6 9 I
5

I 0
21

0.250

O.OO0
0.214

I 0 0
13

0.250 O.Oo0

O.OO0
0.157

7 5 181

5 1 0 0 1 1 0

83

34

0.188

Aff: affected; Freq: frequency.

938
FAMILIAL POLYPOSIS CoLl

CANCER August I5 1985

Vol. 56

of families showing clusters of colon cancer and/or endometrial carcinoma, and may be used effectively for indicating when it is worthwhile to extend kindreds showing nuclear cancer clusters in the search for hereditary cancer syndrome identification. Since considerable time, money and, on occasion, emotional stress may occur to high-risk patients when the process of extending a family is begun, this decision becomes a matter of extreme importance.
REFERENCES
I . Lynch HT, Albano WA, Danes BS, Lynch J, Lynch PM. Precursor conditions and monitoring of high risk colon cancer patients. In: Stroehlein JR. Romsdahl MM, eds. Gastrointestinal Cancer. New York: Raven Press, 1981; 297-325. 2. Lynch HT. Shaw MW, Magnuson CW, L a w n AL, Krush AJ. Hereditary factors in two large Midwestern kindreds. Arch Inrern Med 1966: 117:206-212. 3. Lynch HT. Familial cancer prevalence spanning eight years. Arch Intern Med 1974; I34:93 1-938. 4. Albano WA, Recabaren JA, Lynch HT e/ al. Natural history of hereditary cancer of the breast and colon. Cancer 1982; 50:360-363. 5 . Lynch HT, Lynch PM, Albano WA, Lynch JF. The cancer family syndrome: A status report. Dis Colon Rectum 1981; 24:31 I 322. 6. Lynch HT, Hams RE, Bardawil WA e/ a/. Management of hereditary site-specific colon cancer. Arch Surg 1977; I 12: 170- 174. 7. Wanhin AS. Heredity with reference to carcinoma as shown by the study of the cases examined in the pathological laboratory of the University of Michigan, 1895- I9 13. Arch In/ern Med I9 13; 12546-

FIG. 1. Diagrammatic presentation of the contribution to the colon cancer burden attributable to nonpolyposis colon cancer.

of these lesions may have been present for at least several years before their clinical diagnosis. Therefore, detection programs must be initiated early enough to be able to detect the lesions before the time when they present at a stage that would compromise the patients survival. When clinical stigmata are lacking, one must rely heavily upon the pedigree for assessment of cancer risk status. However, this has a major limitation for clinical application in that progeny or siblings of cancer-syndrome-affected individuals can only be assigned a cancer predictability risk of 50%. Therefore, the discovery of a biomarker(s) associating with the cancer-prone genotype would prove invaluable for the elucidation of genotypic status, thereby enabling the clinician to predict with greater confidence those patients who will (and contrariwise, those. who will not) manifest the phenotype (clinical cancer). Such potential biomarkers have been investigated in the various cancer-prone conditions and might harbor cancer control implications that could become legion. Consistently associated biomarkers could also be useful for studies of environmental interaction with cancerprone genotype, thereby enabling a better assessment of pathogenesis and carcinogenesis (see companion article). We do not have sufficient information to critically assess in a quantitative manner the number of colonic polyps that may be present in patients currently considered to be prone to HNPCC. However, we have not encountered patients within these families with florid polyp expression, although on occasion, patients with isolated polyps have been identified. The results in this hereditary resource provide a better basis for interpreting the significance of biomarker findings contained in our companion article (Part 11). They also serve as a basis for evaluating nuclear components

555. 8. Lynch HT. Krush AJ. Cancer family G revisited: 1895-1970. Cancer 1971; 27:1505-1511. 9. Lynch HT, Follett KL, Lynch PM, Albano WA, Mailliard JA, Pierson RL. Family history in an oncology clinic: Implications for cancer genetics. JAMA 1979; 242: 1268-1 272. 10. Lipkin M. Blattner WE, Fraumeni JF, Lynch HT. Deschner E. Winawer S. Tritiated thymidine (bp, @h) labeling distribution as a marker for hereditary predisposition to colon cancer. Cancer Res 1982; 43: 1899- 1904. 1 I. Sandberg AA. The Chromosomes in Human Cancer and Leukemia. New York: Elsevier North-Holland, 1980. 12. Berlinger NT, Good RA. Suppressor cells in healthy relatives of patients with hereditary colon cancer. Cancer 1980; 45: I 112-1 116. 13. Guirgis HA, Lynch HT, Hams RE, Vandevoorde JP. Genetic and communicable effects on carcinoembryonic antigen expressivity in the cancer family syndrome. Cancer Res 1978; 38:2523-2528. 14. Katano M, Fujiwara H, Toyoda K, Torisu M. lmmunogenetic studies of familial large bowel cancer. Gann 1980; 71583-588. 15. Danes BS. Occurrence of in virro tetraploidy in the heritable colon cancer syndromes. Cancer 198 I ; 48: 1596-1 601. 16. Mikol YB, Lipkin M. Increased methionine requirement in skin fibroblasts of humans affected with familial colon cancer or Gardner syndrome. J Nut1 Cancer Inst 1984 72:19-22. 17. Memtt AD. MEGADATS A computer system for family data acquisition, storage, and analysis. In: Emery AEH. Miller JR, eds. Registers for the Detection and Prevention of Genetic Diseases. New York: Stratton Intercontinental Medical Book Corp., 1976. 18. Gersting JE. MEGADATS: A medical genetics database management system: 1. Design, data structures. and pedigree input. In: Proceedings of the Workshop on Computer Health Care Resources 1978; 231-242. 19. Lynch PM, Lynch HT, Hams RE. Hereditary proximal colonic i s Colon Rectum 1977; 20661-668. cancer. D 20. Lynch HT, Lynch PM, Hams RE. Minimal genetic findings and their cancer control implications: a family with the cancer family syndrome. J A M 4 1978; 240535-538.