Genetic Counseling for Hereditary Cancer

January 1, 1996
Henry T. Lynch, MD
Henry T. Lynch, MD

Volume 10, Issue 1

This paper provides an overview of the current approach to genetic counseling for cancer, using hereditary

ABSTRACT: This paper provides an overview of the currentapproach to genetic counseling for cancer, using hereditary nonpolyposiscolorectal cancer (HNPCC) as a prototype. Heretofore, when evaluatingthe possibility of an HNPCC diagnosis, physicians had to relyexclusively on a detailed family history of cancer in the contextof an extended pedigree. Patients in the direct genetic lineagewho had one or more first-degree relatives with an HNPCC syndromecancer were told that they had a 50% likelihood of inheritingthe deleterious gene. However, with the discovery of the HNPCCgenes (hMSH2, hMLH1, hPMS1, hPMS2), genetic counseling can nowprovide a more precise determination of a patient's lifetime cancerdestiny. Since these DNA findings are new, guidelines for sharingthis information with patients remain preliminary. One must becertain that the patient wants to receive DNA information andthat he or she is aware of potential discrimination by insurancecompanies and employers, as well as the possibility of psychologicalsequelae. [ONCOLOGY 10(1):27-34, 1996]

Introduction

Genetic counseling should be an integral component of the managementof hereditary cancer. This paper will provide an overview of geneticcounseling in cancer using hereditary nonpolyposis colorectalcancer (HNPCC) as a prototype. Our purpose will be to describean approach to genetic counseling that is in concert with thegenetics, natural history, surveillance, and management of HNPCC.

Hereditary Nonpolyposis Colorectal Cancer

Hereditary colorectal cancers are exceedingly heterogeneous (Figure1). Hereditary nonpolyposis colorectcal cancer, one subset ofthese disorders, is subdivided into Lynch syndromes I and II.Lynch syndrome I is characterized by early-onset (age, 40 years)colorectal cancer that shows a predilection (70%) for the proximalcolon (ie, to the right of the splenic flexure) and has a highrate of metachronous colorectal carcinoma. (In patients who undergoless than a subtotal colectomy for the initial cancer, the rateof metachronous cancer is approximately 45% by age 40.) The Lynchsyndrome II variant has all of the features of Lynch syndromeI, but in addition, includes a litany of extracolonic cancers.Foremost among these is endometrial carcinoma, followed by carcinomaof the ovary, small bowel, stomach, and pancreas, and transitional-cellcarcinoma of the ureter and renal pelvis. The penetrance of theHNPCC genes is about 90% [1,2].

Although HNPCC is not a polyposis disorder, affected individualsdo form colonic adenomas, and there is strong circumstantial evidencethat these adenomas are precursor lesions to malignancy.

Histologic Features

The colorectal malignancies in HNPCC do not have unique histologicfeatures, but some differences have been observed when large seriesof HNPCC and sporadic colon cancer cases have been compared. Mostcomparison studies indicate that HNPCC is more likely to producepoorly differentiated carcinoma, coupled with an excess of mucinousand signet-cell histologies. A small, but significant percentage(9%) [3] of colorec tal cancers in HNPCC form solid sheets ofcells with an expanding growth pattern and a marked host lymphocyticresponse-a pattern that has been referred to as "medullarycarcinoma" by Jessurun and Manivel [4]. This pattern is interestingbecause it is the usual histology of sporadic colon cancers withdefective mismatch repair [5]. In turn, this immunologic responsemay be responsible for the improved prognosis noted in some HNPCCpatients [6].

Molecular Genetics

Molecular genetic studies have identified germ-line mutationsin an increasingly large number of hereditary cancer syndromes(Table 1). The genetic basis for HNPCC has been proved by geneticlinkage between cancer occurrences and chromosome 2p in some familiesand chromosome 3p in other families [7,8].

Localization of a DNA mismatch repair gene in the critical regionof chromosome 2p was documented by Leach et al [9] and Fishelet al [10] with the discovery of hMSH2 mutations in this genein several HNPCC families. Subsequently, a second mismatch repairgene (hMLH1) was identified in the critical region of chromosome3p, and mutations of that gene were found in the HNPCC familiespreviously linked to chromosome 3p [11,12]. Mutations in thesegenes appear to account for 90% of all known HNPCC families [13].Also, hPMS1 and hPMS2 have been demonstrated to be mismatch repairgenes in HNPCC [14].

Defective DNA mismatch repair results in a steady accumulationof mutations. The mutation load can be detected as errors in longtandem repeat sequences, which produce microsatellite instability.A tumor with microsatellite instability is said to show replicationerror phenotype (RER+).

Interval Cancers

Vasen et al [15] followed a cohort of 51 HNPCC families that included394 first-degree relatives who participated in a nationwide colonoscopysurveillance program. During mean follow-up of 5 years, theseresearchers observed an unexpectedly high rate of advanced colorectalcancers, which were diagnosed within 2 to 5 years after a negativecolon screening examination. The shortest interval between a negativescreening examination and the diagnosis of colorectal carcinomawas 2 to 3 years (six patients).

Our experience with interval cancers has been comparable. Forexample, a woman with Lynch syndrome II who had manifested endometrialcarcinoma at age 36 years underwent colonoscopies every 2 years.Eighteen months after a normal colonoscopic evaluation, she wasdiagnosed with a Duke's B1 cancer of the transverse colon andunderwent a segmental colonic resection. At that point, she insistedon having colonoscopies every 6 months. Five months after hermost recent colonoscopy, she was found to have two primary coloncancers-a Duke's B1 tumor in the cecum and a Duke's A lesion inthe lower rectum.

There are several possible explanations for these interval cancers:First, lesions may have been missed at the time of colonoscopy.Also, it is possible that the gene accelerates the progressionof adenomas, even those that are very small, to cancer. The latterexplanation is consonant with postulates of Jass et al [16] whosuggested that adenomas in HNPCC are more likely to show a villousgrowth pattern with a high degree of dysplasia than are adenomasin a necropsy series, and that these features cause the HNPCCadenomas to undergo more rapid malignant transformation. The hypothesizedrapid progression of the adenoma-carcinoma sequence in HNPCC,which may be as short as 2 to 3 years, is in striking contrastto findings from the National Polyp Study, which found that theadenoma-carcinoma sequence in the general population is about8 to 10 years.

Another potential explanation for interval cancers in HNPCC isthat gene carriers develop "de novo" cancers [17,18].

Adenomas and Accelerated Carcinogenesis

As mentioned above, strong circumstantial evidence suggests thatadenomas are precursor lesions to malignancy in patients withHNPCC [19-21]. Leach et al [9], however, found that the HNPCCgenes do not initiate adenoma development, but rather accelerateadenoma evolution. In turn, Cannon-Albright et al [22] have shownthat genetic factors are responsible for proneness to developadenomas.

It is hypothesized that when adenomas do occur in patients withHNPCC, they are more likely to progress to carcinoma than areadenomas in the general population. In theory, the "mutatorphenotype" of HNPCC could drive the promotion of adenomato carcinoma, once the neoplastic process has been initiated.The existence of "aggressive adenomas" in HNPCC is supportedby prospective surveillance studies showing a much lower adenoma-to-carcinomaratio than in the general population. Jarvinen et al [23], forexample, found 22 adenomas and 5 cancers in HNPCC families overa 3-year period. In the Netherlands, 14 adenomas and 6 cancerswere discovered in a screening program of HNPCC kindreds [24],and Jass et al [25] in New Zealand found 25 adenomas and 2 cancers.In contrast, data from the National Polyp Study suggest that 41to 119 polypectomies are required to prevent 1 colorectal cancerin the general population [17,18].

Surveillance and Management Guidelines

Knowledge of the natural history and molecular genetics of HNPCChas been useful in developing the following guidelines for surveillance(Table 2) and management of high-risk patients:

Genetic Counseling is initiated at age 20 years.

Full Colonoscopy--We recommend that these patients undergofull colonoscopy beginning sometime between age 20 and 25 (flexiblesigmoidoscopy is not adequate). Colonoscopy is repeatedevery 2 years to age 35 and annually thereafter. If colonoscopyexpertise is unavailable, double air-contrast barium enema canbe used instead.

Endometrial Curettage should be done yearly, beginningat age 30 years.

Ovarian Screening--Transvaginal ultrasound, Doppler colorblood flow imaging, and CA 125 determinations are all initiatedat age 30 years and are repeated annually thereafter.

Patients Who Develop Colorectal Cancer--If colorectal canceroccurs, surgical resection must be no less than a subtotalcolectomy because the entire colonic mucosa is at risk formalignant transformation. Women who have completed their familiesat the time they manifest colorectal cancer are given the optionof undergoing a prophylactic total abdominal hysterectomy andbilateral salpingo-oophorectomy at the time of their subtotalcolectomy.

Prophylactic Surgery--Individuals who have germ-line mutations(hMSH2, hMLH1, hPMS1, hPMS2) are given the option of prophylacticsubtotal colectomy or lifetime colonoscopy screening.

Cancer Screening vs Prophylactic Surgery

Patients' lack of faith in their physician's ability to diagnosecancer early may impact heavily on their decisions about cancerscreening practices. Those who are known to be gene carriers maynot have confidence in the efficacy of colonoscopy, endometrialcurettage, or transvaginal ovarian ultrasound, Doppler color bloodflow imagery, and CA 125 for detecting early-onset colon, endometrial,and ovarian cancer, respectively. In these situations, the questionfrequently emerges as to whether prophylactic surgery should beentertained as a viable option to surveillance. This concern isparticularly cogent when the patient's quality of life is worsenedby a severe cancer phobia.

In patients with familial adenomatous polyposis (FAP), these mattersoften have been worked out satisfactorily since prophylactic colectomyhas become an acceptable option. Unfortunately, these patientsremain at risk for intra-abdominal desmoids, a disease that maybe promoted by the prophylactic life-saving surgical intervention(colectomy) per se. Furthermore, other forms of cancer in FAP,particularly periampullary carcinoma, may become a worrisome issue.

Familial adenomatous polyposis is mentioned in this context simplybecause there is a larger body of information about its geneticcounseling and management [26]. Nevertheless, it remains a geneticcounseling challenge because of limited information about themanagement of extracolonic cancers in this disease. The same problemexists with regard to extracolonic cancers in HNPCC [2,27].

Genetic Counseling

Importance of Family History

Historically, the first example of genetic counseling in HNPCCtook place in 1895, when Aldred Warthin queried his seamstressabout the reasons for her depression. She told him that she wasfearful that she would die at an early age from cancer, sinceit was so common in her family. Warthin explored the family historyin depth, hoping that he could reassure her. Unfortunately, asthe patient had predicted, she did die of cancer (endometrial)at an early age [28].

As in Warthin's experience of a century ago, the compilation ofa thorough family history remains the first step for genetic counselingin any situation where hereditary disorder is suspected. It isvirtually impossible to establish an accurate hereditary cancersyndrome diagnosis and, in turn, to determine a given patient'srisk for cancer, without knowing how this disease is manifestedin his or her family. Regrettably, compilation of the family historyremains one of the most neglected parts of the medical work-upof patients [29].

Warthin clearly had insight into this matter, given the prodigiouslydetailed family history he provided, including pathologic documentation,in his 1913 publication about the seamstresses' family (FamilyG) [28]. Subsequently, the history of this family was updatedand found to be consonaHNPCith HNPCC of the Lynch syndrome IIvariant [30].

Basic Principles

A quarter of a century ago, several basic principles of geneticcounseling were proposed [31]:

  • Genetic counseling implies that we are dealing with an intenselyconcerned patient.
  • Genetic counseling is medically oriented, in that the counselorhas a deep interest in and obligation to explore the disease statusof the patient. Therefore, it follows that the counselor willmake use of all available diagnostic, preventive, and therapeuticregimens available.
  • Genetic counseling must encompass "case finding"in the family of HNPCC patients. In other words, high-risk relativesof the proband need to be educated about the disease's naturalhistory and genetic risk so that they can benefit from this information.
  • Genetic counseling must be psychodynamically oriented. Thisimplies that the counselor is sufficiently versed in psychologyand psychiatry to recognize and deal effectively with the rangeof emotions that may arise in the family in general and the patientin particular. In short, the counselor must appraise the fullextent of emotional factors, and must be sympathetic to them,providing an "empathetic listening ear."
  • The counselor must provide accurate genetic information onlywhen the patient is emotionally capable of grasping its significance.

These guidelines appear to be as important today as they werewhen they were first formulated, particularly given the prodigiousadvances in knowledge about cancer genetics and molecular genetics.

Genetic Counseling Difficulties in HNPCC

Over the years, genetic counseling in HNPCC had been imprecisebecause premonitory physical stigmata of HNPCC's genetic susceptibilityto cancer are lacking. Investigators had to establish a diagnosisbased exclusively on the natural history and the pattern of cancerdistribution within a given extended HNPCC kindred. An exceptionwas the subset of Lynch syndrome II patients who harbor the phenotypicexpression of cutaneous signs of cancer risk relevant to the Muir-Torresyndrome [32,33], namely, sebaceous adenomas, sebaceous carcinomas,and multiple keratoacanthomas, in concert with multiple visceralcancers.

When counseling first-degree relatives of individuals affectedwith an HNPCC syndrome cancer, one could only estimate the riskto be approximately 50% for developing cancer at specific anatomicsites. It was impossible to state whether family members wereat approximately 90% risk for these cancers (based on the gene'spenetrance), or whether they had escaped inheriting the deleteriousgene and had the same risk for colorectal cancer as the generalpopulation (5%).

Fortunately, the genetic predictability of HNPCC has changed remarkablyas a result of the identification of the HNPCC genes [7-10]. Inthe near future, when a simple blood test for HNPCC germ-linemutations becomes available, it will be possible to utilize thisinformation for highly definitive genetic risk assessment. Specifically,it will be possible to tell high-risk patients whether or notthey have inherited the cancer-prone susceptibility gene, basedon an approach similar to that employed for FAP and its adenomatouspolyposis of the colon and rectum (APC) mutation(s) [26].

Patients who are found to be HNPCC germ-line carriers could thenplan to follow the vigorous surveillance and management recommendationsoutlined above. The gene-negative patients could be told thatthey do not have to undergo these rigorous, expensive, and time-consumingsurveillance/management measures and should follow the AmericanCancer Society's recommendations for the general population.

Unresolved Issues

Once it becomes possible to identify HNPCC germ-line carriers,various important issues will need to be addressed: For example,when is it appropriate to initiate genetic counseling? Is it prudentto test a child or young adult for gene carriage status?

Some people argue that children and adolescents may not be capableof making informed decisions about the implications of this knowledge.Gene mutation information may even influence child-parent bonding.Since HNPCC cancers rarely occur in the teenage years, it wouldseem prudent to delay genetic testing until the patient reacheshis or her 20s. Not only would counseling probably be more meaningfulif begun during adulthood but also would coincide with the timeframe for initiating cancer screening. Unfortunately, there areno clear-cut answers to these vexing questions. Clearly, moreresearch will be required.

A priori, given the wide range of differences in maturity betweenyouthful individuals, it is likely that one will find marked variabilityin how this information may be received and interpreted by youngadults. Some may find it extremely beneficial psychologicallywhile others may find it emotionally devastating. For example,a patient in the early teens who receives information that heor she is a gene carrier with the strong likelihood of eventuallymanifesting cancer may have an otherwise normal adolescence significantlydisrupted. This knowledge could compromise interpersonal relationships,as well as education, career, and marriage plans. However, ona more positive side, we are dealing with a disorder that can,in many circumstances, be highly amenable to appropriate managementand treatment. This optimism contrasts strikingly with other hereditarydisorders, such as Huntington's disease, for which medical treatmentis unavailable but genetic counseling experience is accruing [34-36].

Another question is whether or not members of HNPCC families wantto know their genetic cancer risk status. The anxiety of learningone is at high risk for cancer must be balanced against the anxietyof coping with uncertainty. Thus, is it psychologically more advantageousto know whether or not one is a gene carrier or to remain in thedark? Again, these issues clearly require more psychological research.

Barriers to Effective Counseling of First- and Second-DegreeRelatives

Once the HNPCC syndrome diagnosis has been established, cancersurveillance and management recommend- ations, within the contextof genetic counseling, should be extended to all availableprimary and secondary relatives of the proband. Unfortunately,this is not a common practice in the clinical setting. For example,Arvanitis et al [37] showed that even in such a strikingly recognizabledisorder as FAP, where proctosigmoidoscopy could lead to theprevention of colorectal cancer through prophylactic colectomy,59% of patients are not identified early and, consequently,die of metastatic disease. Clearly, this means that the messageabout the hereditary susceptibility to colorectal cancer amongmembers of FAP families simply is not being adequately addressed.

In HNPCC, where reliable premonitory physical signs of cancerrisk are lacking, the problem will admittedly be much more difficult.Nevertheless, the solution to this problem should be facilitatedonce a simple blood test for hMSH2, hMLH1, hPMS1, and hPMS2 becomesclinically available.

The lack of genetic counseling of high-risk family members stems,in part, from shortcomings in the medical education of physicians.Typically, physicians focus on the medical problems of individualpatients who are under their care. Consequently, appropriate cancercontrol measures for their patients' high-risk relatives oftenare not initiated.

Physicians may also be concerned about contacting high-risk collateralrelatives because of how this might be misconstrued by the medicalcommunity. For example, in some medical circles, this type ofsolicitation is disconcerting, and may unjustifiably be construedas tantamount to attempts at economic gain on the part of thephysician.

Other barriers to effective genetic counseling and cancer controlinclude patients' fear and denial, which may impact heavily oncompliance with recommended screening procedures.

There also may be socioeconomic and educational barriers to compliance.Health insurance carriers may not pay for screening proceduresand, therefore, high-risk patients may simply forego surveillance.Based on past experiences of relatives, patients may fear thatif their insurance company is aware of the increased cancer risk,the policy could be canceled and/or future policies may be denied.Others may fear that their own cancer risk will translate intosimilar discrimination problems for their siblings and children.Hence, they may prefer to keep this information from the insurancecompany and, in turn, not request coverage for surveillance. Wehave witnessed many of these problems over the years.

How can these problems be resolved? It will be important to educatephysicians, insurance actuaries, and medical directors of insurancecompanies about the needs of high-genetic-risk patients, includingHNPCC germ-line carriers. Research demonstrating cost-benefitof decreased morbidity and mortality through early cancerdiagnosis and cancer prevention (should it be adequatelydemonstrated) would be useful to third-party carriers.

Special funding to address all facets of these issues, includinggenetic counseling, has been initiated through the National Centerfor Human Genome Research and the NIH [38].

International Collaborative GroupOn HNPCC

The International Collaborative Group on HNPCC (ICG-HNPCC), anorganization committed to the advancement of knowledge about HNPCC,is attempting to resolve all these problems. The following arethe overall objectives of the ICG-HNPCC:

Conclusion

Genetic counseling in HNPCC (and all forms of hereditary cancer)must encompass the myriad social-psychological-economic needsof the patient, in concert with education about the natural historyof the cancer and the most up-to-date available surveillance andmanagement strategies. Furthermore, the genetic counselor mustbe fully abreast of molecular genetic knowledge about HNPCC sothat the patient and family can derive the maximum benefit fromDNA testing. Clearly, given that knowledge about many of theseissues is in its infancy, priority must be given to research toestablish appropriate genetic counseling guidelines for HNPCC.

References:

1. Watson P, Lynch HT: Extracolonic cancer in hereditary nonpolyposiscolorectal cancer. Cancer 71(3):677-685, 1993.

2. Lynch HT, Smyrk TC, Watson P, et al: Genetics, natural history,tumor spectrum, and pathology of hereditary nonpolyposis colorectalcancer: An updated review. Gastroenterology 104:1535, 1993.

3. Jass JR, Smyrk TC, Stewart SM, et al: Pathology of hereditarynonpolyposis colorectal cancer. Anticancer Res 14:1631-1634, 1994.

4. Jessurun, MR, Manivel JC: Cecal, poorly-differentiated adenocarcinoma,medullary type. Mod Pathol 5:43A, 1992.

5. Kim, H, Jen J, Vogelstein B, et al: Clinical and pathologicalcharacteristics of sporadic colorectal carcinomas with DNA replicationerrors in microsatellite sequences. Am J Pathol 145:148, 1994.

6. Lynch HT, Smyrk TC, Watson P, et al: Genetics, natural history,tumor spectrum, and pathology of hereditary nonpolyposis colorectalcancer. Gastroenterology 104:1535, 1993.

7. Peltomäki P, Aaltonen L, Sistonen P, et al: Genetic mappingof a locus predisposing to human colorectal cancer. Science 260:810,1993.

8. Lindblom A, Tannergard P, Werelius B, et al: Genetic mappingof a second locus predisposing to hereditary nonpolyposis colorectalcancer. Nat Genet 5:279, 1993.

9. Leach FS, Nicolaides NC, Papadopoulos N, et al: Mutations ofa MutS homolog in hereditary non-polyposis colorectal cancer Cell75:1215, 1993.

10. Fishel R, Lescoe MK, Rao MRS, et al: The human mutator genehomolog MSH2 and its association with hereditary nonpolyposiscolon cancer. Cell 75:1027-1038, 1993.

11. Papadopoulos N, Nicolaides NC, Wei Y-F, et al: Mutation ofa mulL homolog in hereditary colon cancer. Science 262:1625-29,1994.

12. Bonner CE, Baker SM, Morrison PT, et al: Mutation in the DNAmismatch repair gene homologue hMLH1 is associated with hereditarynon-polyposis colon cancer. Nature 368:258, 1994.

13. Liu B, Parsons RE, Hamilton SR, et al: hMSH2 mutations inhereditary nonpolyposis colorectal cancer kindreds. Cancer Res54:4590-4594, 1994.

14. Nicholaides NC, Papadopoulos N, Liu B, et al: Mutations oftwo PMS homologues in hereditary nonpolyposis colon cancer. Nature371:75, 1994.

15. Vasen GFA, Nagenhast FM, Meera Khan P: Interval cancers inhereditary nonpolyposis colorectal cancer (Lynch syndrome). Lancet345:1183-1184, 1995.

16. Jass JR, Stewart SM, Stewart J, et al: Hereditary non-polyposiscolorectal cancer: morphologies, genes and mutations. MutationRes 290:125-133, 1994.

17. Winawer S, Zauber AG, O'Brien MJ, et al: National Polyp StudyWorkgroup: Randomized comparison of surveillance intervals aftercolonoscopic removal of newly diagnosed adenomatous polyps. NEngl J Med 328:901-906, 1993.

18. Winawer SJ, Zauber AG, Ho MN, et al: National Polyp StudyWorkgroup. Prevention of colorectal cancer by colonoscopic polypectomy.N Engl J Med 329:1977-1981, 1993.

19. Love RR, Morrisey JF: Colonoscopy in asymptomatic individualswith a family history of colorectal cancer. Arch Intern Med 144:2209-2211,1984.

20. Lanspa SJ, Lynch HT, Smyrk TC, et al: Colorectal adenomasin the Lynch syndromes-results of a colonoscopy screening program.Gastroenterology 98:1117-1122, 1990.

21. Mecklin JP, Jarvinen HJ: Clinical feature of colorectal carcinomain the cancer family syndrome. Dis Colon Rectum 29:160-164, 1986.

22. Cannon-Albright AL, Skolnick MH, Bishop DT, et al: Commoninheritance of susceptibility to colonic adenomatous polyps andassociated colorectal cancers. N Engl J Med 319:533-537, 1988.

23. Jarvinen HJ, Mecklin JP, Sistonen P: Screening reduces colorectalcancer rate in families with hereditary nonpolyposis colorectalcancer. Gastroenterology 108:1405-1411, 1995.

24. Vasen HF, Hartog Jager FCA, Menko FH, et al: Screening forhereditary non-polyposis colorectal cancer: A study of 22 kindredsin the Netherlands. Am J Med 86:278-281, 1989.

25. Jass JR, Young PJ, Robinson EM: Predictors of presence, multiplicity,size and dysplasia of colorectal adenomas: A necropsy study inNew Zealand. Gut 33:1508-1514, 1992.

26. Petersen GM, Francomano C, Kinzler K: Presymptomatic directdetection of adenomatous polyposis coli (APC) gene mutations infamilial adenomatous polyposis. Hum Genet 91:307-311, 1993.

27. Watson P, Lynch HT: Extracolonic cancer in hereditary nonpolyposiscolorectal cancer. Cancer 71,3:677-685, 1993.

28. Warthin AS: Heredity with reference to carcinoma. Arch InternMed 12:546-555, 1913.

29. Lynch HT: Cancer and the family history trail. NY State JMed 91:145-47, 1991.

30. Lynch HT, Krush, AJ: Cancer family "G" revisited:1895-1970. Cancer 27:1505-1511, 1971.

31. Lynch HT: Dynamic Genetic Counseling for Clinicians. Springfield,MA, CC Thomas, 1969.

32. Lynch HT, Lynch PM, Pester J, et al: The cancer family syndrome:Rare cutaneous phenotypic linkage of Torre's syndrome. Arch InternMed 141:607, 1980.

33. Lynch HT, Fusaro RM, Roberts L, et al: Muir-Torre syndromein several members of a family with a variant of the cancer familysyndrome. Br J Dermatol 113:295-301, 1985.

34. The World Federation of Neurology Research Group on Huntington'sDisease: Pre-symptomatic testing for Huntington's disease: A worldwide survey. J Med Genet 30:1020-1022, 1993.

35. Thies U, Bockel B, Bochdalofsky V: Attitudes of neurologists,psychiatrists, and psychotherapists towards predictive testingfor Huntington's disease in Germany. J Med Genet 30:1023-1027,1993.

36. Simpsom SA, Harding AE: Predictive testing for Huntington'sdisease: after the gene. J Med Genet 30:1036-1038, 1993.

37. Arvanitis ML, Jagelman DG, Fazio VN, et al: Mortality in patientswith familial adenomatous polyposis. Dis Colon Rectum 33:639-642,1990.

38. National Institutes of Health: NIH Guide to Grants and Contracts,vol 19, pp 12-15. Bethesda, MD, National Institutes of Health,1990.