GYNECOLOGIC
ONCOLOGY
47, 167-171
(1992)
Endometrial Adenocarcinoma: Genetic Analysis Suggesting Heritable Site-Specific Uterine Cancer LISA G. SANDLES,* LEE P. SHULMAN,*
GUY J. PHOTOPULOS,*
SHERMAN ELIAS,*
LINDA M. SMMILEY ,* WILLIAM M.
POSTEN,?AND JOE LEIGH SIMPSON* ‘Department
of Obstetrics
and Gynecology, Memorial
University Hospital,
of Tennessee, Memphis, Tennessee 38103; 810 Madison Avenue, Memphis, Tennessee Received
January
Genetic factors are clearly integral to the etiology of neoplasia. A cancer family syndrome (Lynch syndrome II) consistingof uterine, colon, and ovarian cancer is recognized, but the heritability of isolatedendometrialadenocarcinomahasnot otherwise beenthoroughly investigated.We have performed pedigreestudies in index caseswith endometrial adenocarcinoma,using spouses ascontrols. Preliminary resultsfrom 64 probandsshowed four families in which endometrial adenocarcinomawas diagnosedin at leastone first-degreerelative of the proband(mother, daughter, sister); none showedrelatives with colon or ovarian cancer. In none of the 34 control pedigreesdid either a mother or sisterhave endometrialadenocarcinoma.In four other families, multiple first- and second-degree relatives of probandshad adenocarcinomaof the uterus, colon, or ovary, presumablyrepresentinga cancerfamily syndrome(Lynch syndromeII). Conclusion: Our preliminary data not only show familial and probably heritabletendenciesfor endometrialadenocarcinoma,but further suggestthat thereare at leasttwo distinct forms: (1) the previously describedLynch syndromeII (cancer family syndrome),and (2) a heretoforeunemphasizedentity characterizedby a tendency to endometrialadenocarcinomaalone. Qi~?z.&ad&c PWS, IIIC.
Meeting of the Society March 15-18, 1992.
Baptist
patterns of inheritance: a site-specific ovarian cancer syndrome [2], a breast-ovarian cancer syndrome [3], and a cancer family syndrome [4,5]. In the cancer family syndrome, males and females are at increased risk for colon cancer and other adenocarcinomas; females are also at risk for ovarian and endometrial adenocarcinoma [4]. In a pedigree analysis of 51 consecutive cases of endometrial adenocarcinoma, Boltenberg et al. [S] noted a similarly affected mother or sister in 4 cases (8%) and observed three families that likely had the cancer family syndrome. These authors suggested that two genetic models may play a role in the development of endometrial adenocarcinoma: the cancer family syndrome and a predisposition for endometrial adenocarcinoma alone. The present study was undertaken to investigate the heritability of endometrial adenocarcinoma. We sought to determine the proportion of cases manifesting heritable tendencies and to ascertain the likely mode of inheritance(s). AND METHODS
From August 1989 through August 1991,70 consecutive patients with endometrial adenocarcinoma diagnosed at the University of Tennessee, Memphis, were identified and invited to participate in our study. Our referral region includes primarily western Tennessee, eastern Arkansas, and northern Alabama and Mississippi. Although a few local obstetrician/gynecologists do not refer all well-differentiated endometrial adenocarcinomas, most early grade and higher grade lesions are referred to our regional center. Major participating hospitals include (1) the Regional Medical Center, staffed by University of Tennessee, Memphis faculty, and the major health care facility for indigent patients in this area; and (2) the Baptist
Endometrial adenocarcinoma is the most common female genital malignancy. The American Cancer Society estimates that 33,000 new cases of endometrial adenocarcinoma will be diagnosed in 1991, with approximately 5500 deaths from this disease [l]. Many risk factors are known, including obesity, late menopause, and unopposed exogenous estrogen. However, relatively little attention has been given to genetic (familial) factors in endometrial adenocarcinoma. In contrast, ovarian cancer is known to involve at least three autosomal dominant at the 23rd Annual San Antonio, TX,
of Pathology,
22, 1992
PATIENTS
INTRODUCTION
Presented Oncologists,
and t Department 38103
of Gynecologic
167 0090-8258192 $4.00 Copyright 0 1992 by Academic Press, Inc. All rights of reproduction in any form reserved.
168
SANDLES
Memorial Hospital system, whose patient base is primarily middle class. Overall, 300 to 350 new gynecologic cancers are seen per year. This investigation was conducted under the auspices of the University of Tennessee, Memphis, Institutional Review Board, and all information remained strictly confidential. At the time of their initial ascertainment, patients were asked to complete a detailed seven-page questionnaire eliciting the following information: demograhic data (age, race, occupation, education), presence or absence of chronic medical problems, gynecologic history, and family history regarding the occurrence of cancer among firstand second-degree relatives. Current spouses served as controls. Questionnaires were reviewed by either of two investigators (L.G.S. or L.P.S.) and discussed in person with the patient and her family or via telephone. Written permission was requested to interview affected relatives to verify medical and genealogic data and to obtain such medical records as pathology reports, microscopic slides, and death certificates. Pathologic diagnoses were confirmed (W.M.P.) from all 70 probands (index cases) after review of microscopic slides. Confirmation of cancer in affected relatives is still underway and will be included in a subsequent report. Although cancers from all sities were recorded, this report is restricted to cases of adenocarcinoma of the endometrium, ovary, and colon. RESULTS Of 70 consecutive patients with endometrial adenocarcinema, complete pedigrees were obtained from 64 probands and 34 spouses. In 6 cases, complete family histories could not be obtained because the proband was either lost to follow-up (n = 2) or died (n = 4). In 30 cases, information about the spouse’s family was not available because of death, divorce, separation, or the patient never having been married. In 8 of the 64 proband pedigrees (Figs. 1 and 2), there was a first-degree relative affected with either endometrial, colon, or ovarian adenocarcinoma. These 8 pedigrees can heuristically be divided into two groups: (1) families showing relatives affected by endometrial adenocarcinoma alone (n = 4, Fig. l), and (2) families showing relatives affected by endometrial, colon, or ovarian adenocarcinoma (n = 4, Fig. 2), suggesting a diagnosis of the cancer family syndrome. The remaining 56 probands showed no relatives with endometrial, colon, or ovarian adenocarcinoma. In those four families (Fig. 1) in which endometrial adenocarcinoma occurred in at least one first-degree relative (mother, daughter, or sister) with no family member showing colon or ovarian cancer (Table l), the probands were diagnosed at a mean age of 61.8 years (ages 51, 60,
ET AL.
FIG. 1. Pedigrees (A-D) of four families in which at least one firstdegree relative (mother, daughter, sister) had endometrial adenocarcinema but no other cancers. Open squares and circles denote men and women (respectively) with no cancer; closed symbols denote men and women with cancer confirmed by pathology; stippled symbols denote men and women with cancer established by family history alone. The type of cancer is found next to the square or circle: U, endometrial adenocarcinoma; LA, laryngeal cancer; E, esophageal cancer, P, pancreatic cancer, and LU, lung cancer. Probands are denoted by an arrow and age of individual at time of detection is found next to their circle. Numbers within symbols denote numbers of individuals.
65, and 71 years, respectively). This compared to a mean age of 65.3 + SD 12.2 years in 56 probands who had no affected relative. One proband was Black and three were Caucasian. One patient was hypertensive. None had a history of diabetes mellitus, obesity, or exogenous estrogen use. Stage at diagnoses was IB grade I in one case, IB grade II in two patients, and IIIA grade II in the fourth case. In one family (DO, Fig. l), there were two affected first-degree relatives (a mother and sister) and also affected second- and third-degree relatives (maternal aunt and maternal niece). Of note is kindred GM (Fig. l), in which the index case had endometrial adenocarcinema diagnosed at age 65. Her only child, a daughter, had endometrial adenocarcinoma. The daughter verbally confirmed her own diagnosis of endometrial adenocarcinema, but declined signing a release for medical records, In the four pedigrees suggestive of the cancer family syndrome (Fig. 2), the probands were diagnosed at a mean age of 64.5 years (ages 51, 65, 67, and 75 years, respectively). All four probands were Caucasian, two were hypertensive, and none had a history of diabetes mellitus, obesity, or exogenous estrogen use. Stage at diagnosis was IB grade I in two probands, stage IIIC grade II in one patient, and stage IV in the fourth patient. In kindred ET (Fig. 2), the proband was diagnosed at age 65. Her father had colon cancer, and a paternal grand-
HERITABLE
Ptdrgrcc
2A ET
Pcdigrcc
Pedigree
SITE-SPECIFIC
2B : SL
2C TC
ENDOMETRIAL
169
CANCER
Stage of disease was as follows: stage I, n = 41 (73%, with 7 patients having stage IA, 25 patients stge IB, and 9 patients stage IC); stage IIB, n = 2; stage IIIA, n = 3; stage IIIC, n = 1; and stage IV, n = 5. Four additional patients, all with clinical stage I disease, did not undergo surgical staging: 3 had serious underlying medical problems and 1 patient declined treatment. None of the four probands having first-degree relatives with site-specific endometrial adenocarcinoma had multiple primary malignancies. Among the 56 cases of sporadic endometrial adenocarcinoma, 4 had such a history (all pathologically confirmed). One patient had renal cell carcinoma antedating her diagnosis of endometrial adenocarcinoma, a second had breast cancer prior to having endometrial adenocarcinoma, a third had breast cancer 1 year after her diagnosis of endometrial adenocarcinoma, and a fourth had ovarian cancer diagnosed at the time of hysterectomy for endometrial adenocarcinoma. None of these patients had relatives affected in a fashion that would suggest a familial cancer syndrome. DISCUSSION
Pcdrgrtc
ZD SC
FIG. 2. Pedigrees (A-D) of four families in which at least one firstdegree relative (mother, daughter, sister) had endometrial, colon, or ovarian adenocarcinoma. Open squares and circles denote men and women (respectively) with no cancer; closed symbols denote men and women with cancer confirmed by pathology; stippled symbols denote men and women with cancer established by family history alone; stippled symbols with a vertical line denote cancer confirmed by medical record or death certificate. The type of cancer is found next to the square or circle: U, endometrial adenocarcinoma; C, Colon cancer; 0, ovarian cancer; B, breast cancer; S, stomach cancer; LI, liver cancer; LY, lymphoma; SK, skin cancer; and UNK, specific site unknown. Probands are denoted by an arrow and age of individual at time of detection is found next to their circle. Numbers within symbols denote numbers of individuals.
mother had uterine cancer. However, the proband also showed maternal relatives with endometrial cancer. Thus, heritability in this family could reflect either the cancer family syndrome gene(s) transmitted through paternal lineage or a gene(s) predisposing solely to endometrial adenocarcinoma through maternal lineage. Kindred SL (Fig. 2) more plausably showed the cancer family syndrome transmitted through paternal lineage, whereas TC and SC (Fig. 2) showed transmission through maternal lineage. Among the 56 cases of sporadic endometrial adenocarcinoma, the racial composition was as follows: Caucasian, n = 45; Black, n = 8; Hispanic, n = 2; and American Indian, n = 1. Twenty-three patients (41%) were obese, 10 (187 o ) were hypertensive, 8 (14%) had diabetes mellitus, and 4 patients had all three risk factors. Two patients had taken unopposed exogenous estrogens.
Few studies investigating the heritability of endometrial adenocarcinoma have been reported [6-81. In 1966, Lynch et al. [6] reported 154 probands with endometrial carcinoma of whom 20 (13.0%) had a similarly affected first-degree relative. In a pedigree study of 51 patients with endometrial adenocarcinoma, Boltenberg et al. [8] noted a similarly affected mother or sister in 4 cases (8%). Applying similar criteria to our study, 6 of 64 (9.4%) probands with endometrial adenocarcinoma showed a similarly affected first-degree relative. The mode of inheritance cannot be determined definitively, but both our data and those of Lynch et al. [6] and Boltenberg et al. [8] are consistent with autosomal dominant inheritance. Other studies have shown that endometrial adenocarcinoma is one component of the cancer family syndrome. In Boltenberg’s series of 51 probands, there were three families that had pedigrees consistent with the cancer family syndrome, representing 5.8% of all patients with endometrial adenocarcinoma [8]. Our data are only preliminary. Like all studies dealing TABLE 1 Frequenciesof Adenocarcinomaof the Uterus, Colon, and Ovary ObservedamongProbandsand Controls Relatives with endometrial cancer
Probands Controls
Relatives with colon or ovarian cancer
Mothers
Sisters
Parents
Siblings
4/64 (6.3%) o/34
2/106 (1.9%) O/36
2/124 (1.6%) 2/68 (2.9%)
6/128 (4.7%) l/74 (1.4%)
170
SANDLES
with genetic heterogeneity, a large sample size will be required to identify the relatively few kindreds having potential Mendelian tendencies. Nonetheless, we believe two different genetic mechanisms are responsible for the heritable tendencies we and others have observed. In 4 of our 64 (6.25%) families, multiple first-degree relatives had endometrial, colon, or ovarian adenocarcinoma, suggesting the cancer family syndrome (Lynch syndrome II). Another patient had simultaneous endometrial and ovarian adenocarcinoma. Although none of her relatives had a history of colon, ovarian, or endometrial adenocarcinoma, this patient may represent a new germline mutation for the cancer family syndrome. In 4 other families, at least one first-degree relative had endometrial adenocarcinema, but no family member showed colon or ovarian cancer. In 1 of these latter 4 families, two first-degree, one third- and one fourth-degree relative were similarly affected. In aggregate, our findings suggest not only existence of a cancer family syndrome, but existence of a site-specific endometrial adenocarcinoma syndrome as noted by Boltenberg et al. [8]. This would be analogous to the site-specific patterns of inheritance previously described for ovarian cancer [2] and colon cancer [9]. A characteristic of all hereditary cancers is early age of onset [5,7]. Hakala et al. [7] observed the age of onset of endometrial carcinoma in 26 women from 19 families with the cancer family syndrome to be 48.3 years (range, 34 to 63 years). Our data are limited, but to date show only modest differences in age of onset in the four probands having presumptive site-specific endometrial adenocarcinoma compared to sporadic cases (mean ages, 61.8 and 65.3 years, respectively). Another characteristic of hereditary cancer is an increased frequency of multiple primary malignancies. In our series, multiple primary malignancies were observed in none of the four probands in which cancer was restricted to endometrial adenocarcinoma in first-degree relatives. Among the four cancer family syndrome families, a single family member had both colon and ovarian cancer. Again, the sample size is too small to draw definitive conclusions. Among the eight probands in our series with apparent hereditary endometrial adenocarcinoma, five presented with stage IB disease, two with stage III disease, and 1 with stage IV disease. Until a larger series is ascertained, we cannot comment on whether the familial cases of endometrial adenocarcinoma are associated with a more advanced stage at presentation and/or a poorer prognosis compared to sporadic cases. Boltenberg er al. [8] observed that familial cases were associated with more anaplastic tumors and decreased survival. Although mutant genes represent the most likely explanation for our observations of familial tendencies in uterine cancer, other explanations are possible. Such findings could be attributable to chance alone. Unidentified
ET AL.
environmental factors could also exist, although this seems unlikely given that affected women have been raised in multiple households as well as living in different regions. A third possibility relates to reporting errors in pathologically unconfirmed cases. However, one would not expect preferential overreporting of a particular cancer diagnosis (i.e, endometrial adenocarcinoma) in probands as opposed to controls. Nonetheless, we are aggressively pursuing histologic confirmation of diagnoses made in relatives. A fourth possible explanation is misclassification with respect to the diagnosis of site-specific endometrial adenocarcinoma versus the cancer family syndrome. Classification of a kindred as having a sitespecific syndrome could reflect merely the small number of affected family members or the limited time of observation. Over time, other cancers may be expressed; thus, a kindred formerly classified as a site-specific syndrome might actually be shown to represent the cancer family syndrome. These caveats notwithstanding, genetic factors are increasingly recognized as playing integral roles in the etiology of cancer. If a site-specific endometrial adenocarcinoma syndrome exists, it could be characterized by the simple Mendelian (autosomal dominant) inheritance shown by the cancer family syndrome gene. If so, firstdegree relatives would have a 50% risk for inheriting the mutant gene and, hence, possibly developing endometrial adenocarcinoma. Of course, not all inheriting a dominant gene will be affected (incomplete penetrance or variable expressivity, namely atypical endometrial hyperplasia). Nevertheless, increased surveillance among potentially affected female family members may be warranted. Such surveillance might include patient education and counseling, avoidance of periods of anovulation, verification of achieving a secretory endometrium with exogenous hormone replacement, periodic endometrial biopsies, and possibly even prophylactic hysterectomy in selected cases. In the future, the identification of genetic markers, such as specific oncogene mutations or amplification, may make possible identification of affected individuals before phenotypic expression of clinically recognizable cancer. REFERENCES 1. Boring, C. C., Squires, T. S., and Tong, T. Cancer statistics, 1991, Cancer J. Clin. 41, 19-36 (1991). 2. Lynch, H. T., Albano, W., Black, L., Lynch, J. F., Recabaren, J., and Pierson, R. Familial excess of cancer of the ovary and other anatomic sites, J. Am. Med. Assoc. 245, 261-264 (1981). 3. Lynch, H. T., Harris, R. E., Guirgis, H. A., Maloney, K., Carmody, L. L., and Lynch, J. F. Familial association of breast/ovarian carcinoma, Cancer 41, 1543-1549 (1978). 4. Lynch, H. T., Albano, W. A., Lynch, J. F., Lynch, P. M., and Campbell, A. Surveillance and management of patients at high risk for ovarian cancer, Obstet. Gynecol. 59, 589-596 (1982).
HERITABLE
SITE-SPECIFIC
5. Lynch, H. T., Kimberling, W., Albano, W. A., Lynch, J. F., Biscone, K., Schuelke, G. S., Sandberg, A. A., Lipkin, M., Deschner, E. E., Mikol, Y. B., Elston, R. C., Bailey-Wilson, J. E., and Danes, B. S. Hereditary nonpolyposis colorectal cancer (Lynch syndromes I and II). I. Clinical description of resource, Cancer 56, 934-938 (1985). 6. Lynch, H. T., Krush, A. L., and Magnuson, C. W. Endometrial carcinoma: Multiple primary malignancies, constitutional factors, and heredity, Am. J. Med. Sci. 36, 381-390 (1966). 7. Hakala, T., Mecklin, J. P., Forss, M., Jarvinen, H., and Lehtovirta,
ENDOMETRIAL
CANCER
171
P. Endometrial carcinoma in the cancer family syndrome, Cancer 68, 1656-1659 (1991). 8. Boltenberg, A., Furgyik, S., and Kullander, S. Familial cancer aggregation in cases of adenocarcinoma corporis uteri, Actu O&et. Gynecol. Scund. 69, 249-258 (1990). 9. Lynch, H. T., Schuelke, G. S., Kimberling, W. J., Albano, W. A., Lynch, J. F., Biscone, K. A., Lipken, M. L., Deschner, E. E., Mikol, Y. B., Sandberg, A. A., Elston, R. C., Bailey-Wilson, J. E., and Danes, B. S. Hereditary nonpolyposis colorectal cancer (Lynch syndromes I and II). II. Biomarker studies, Cancer 56,939951 (1985).
ONCOLOGY
47, 167-171
(1992)
Endometrial Adenocarcinoma: Genetic Analysis Suggesting Heritable Site-Specific Uterine Cancer LISA G. SANDLES,* LEE P. SHULMAN,*
GUY J. PHOTOPULOS,*
SHERMAN ELIAS,*
LINDA M. SMMILEY ,* WILLIAM M.
POSTEN,?AND JOE LEIGH SIMPSON* ‘Department
of Obstetrics
and Gynecology, Memorial
University Hospital,
of Tennessee, Memphis, Tennessee 38103; 810 Madison Avenue, Memphis, Tennessee Received
January
Genetic factors are clearly integral to the etiology of neoplasia. A cancer family syndrome (Lynch syndrome II) consistingof uterine, colon, and ovarian cancer is recognized, but the heritability of isolatedendometrialadenocarcinomahasnot otherwise beenthoroughly investigated.We have performed pedigreestudies in index caseswith endometrial adenocarcinoma,using spouses ascontrols. Preliminary resultsfrom 64 probandsshowed four families in which endometrial adenocarcinomawas diagnosedin at leastone first-degreerelative of the proband(mother, daughter, sister); none showedrelatives with colon or ovarian cancer. In none of the 34 control pedigreesdid either a mother or sisterhave endometrialadenocarcinoma.In four other families, multiple first- and second-degree relatives of probandshad adenocarcinomaof the uterus, colon, or ovary, presumablyrepresentinga cancerfamily syndrome(Lynch syndromeII). Conclusion: Our preliminary data not only show familial and probably heritabletendenciesfor endometrialadenocarcinoma,but further suggestthat thereare at leasttwo distinct forms: (1) the previously describedLynch syndromeII (cancer family syndrome),and (2) a heretoforeunemphasizedentity characterizedby a tendency to endometrialadenocarcinomaalone. Qi~?z.&ad&c PWS, IIIC.
Meeting of the Society March 15-18, 1992.
Baptist
patterns of inheritance: a site-specific ovarian cancer syndrome [2], a breast-ovarian cancer syndrome [3], and a cancer family syndrome [4,5]. In the cancer family syndrome, males and females are at increased risk for colon cancer and other adenocarcinomas; females are also at risk for ovarian and endometrial adenocarcinoma [4]. In a pedigree analysis of 51 consecutive cases of endometrial adenocarcinoma, Boltenberg et al. [S] noted a similarly affected mother or sister in 4 cases (8%) and observed three families that likely had the cancer family syndrome. These authors suggested that two genetic models may play a role in the development of endometrial adenocarcinoma: the cancer family syndrome and a predisposition for endometrial adenocarcinoma alone. The present study was undertaken to investigate the heritability of endometrial adenocarcinoma. We sought to determine the proportion of cases manifesting heritable tendencies and to ascertain the likely mode of inheritance(s). AND METHODS
From August 1989 through August 1991,70 consecutive patients with endometrial adenocarcinoma diagnosed at the University of Tennessee, Memphis, were identified and invited to participate in our study. Our referral region includes primarily western Tennessee, eastern Arkansas, and northern Alabama and Mississippi. Although a few local obstetrician/gynecologists do not refer all well-differentiated endometrial adenocarcinomas, most early grade and higher grade lesions are referred to our regional center. Major participating hospitals include (1) the Regional Medical Center, staffed by University of Tennessee, Memphis faculty, and the major health care facility for indigent patients in this area; and (2) the Baptist
Endometrial adenocarcinoma is the most common female genital malignancy. The American Cancer Society estimates that 33,000 new cases of endometrial adenocarcinoma will be diagnosed in 1991, with approximately 5500 deaths from this disease [l]. Many risk factors are known, including obesity, late menopause, and unopposed exogenous estrogen. However, relatively little attention has been given to genetic (familial) factors in endometrial adenocarcinoma. In contrast, ovarian cancer is known to involve at least three autosomal dominant at the 23rd Annual San Antonio, TX,
of Pathology,
22, 1992
PATIENTS
INTRODUCTION
Presented Oncologists,
and t Department 38103
of Gynecologic
167 0090-8258192 $4.00 Copyright 0 1992 by Academic Press, Inc. All rights of reproduction in any form reserved.
168
SANDLES
Memorial Hospital system, whose patient base is primarily middle class. Overall, 300 to 350 new gynecologic cancers are seen per year. This investigation was conducted under the auspices of the University of Tennessee, Memphis, Institutional Review Board, and all information remained strictly confidential. At the time of their initial ascertainment, patients were asked to complete a detailed seven-page questionnaire eliciting the following information: demograhic data (age, race, occupation, education), presence or absence of chronic medical problems, gynecologic history, and family history regarding the occurrence of cancer among firstand second-degree relatives. Current spouses served as controls. Questionnaires were reviewed by either of two investigators (L.G.S. or L.P.S.) and discussed in person with the patient and her family or via telephone. Written permission was requested to interview affected relatives to verify medical and genealogic data and to obtain such medical records as pathology reports, microscopic slides, and death certificates. Pathologic diagnoses were confirmed (W.M.P.) from all 70 probands (index cases) after review of microscopic slides. Confirmation of cancer in affected relatives is still underway and will be included in a subsequent report. Although cancers from all sities were recorded, this report is restricted to cases of adenocarcinoma of the endometrium, ovary, and colon. RESULTS Of 70 consecutive patients with endometrial adenocarcinema, complete pedigrees were obtained from 64 probands and 34 spouses. In 6 cases, complete family histories could not be obtained because the proband was either lost to follow-up (n = 2) or died (n = 4). In 30 cases, information about the spouse’s family was not available because of death, divorce, separation, or the patient never having been married. In 8 of the 64 proband pedigrees (Figs. 1 and 2), there was a first-degree relative affected with either endometrial, colon, or ovarian adenocarcinoma. These 8 pedigrees can heuristically be divided into two groups: (1) families showing relatives affected by endometrial adenocarcinoma alone (n = 4, Fig. l), and (2) families showing relatives affected by endometrial, colon, or ovarian adenocarcinoma (n = 4, Fig. 2), suggesting a diagnosis of the cancer family syndrome. The remaining 56 probands showed no relatives with endometrial, colon, or ovarian adenocarcinoma. In those four families (Fig. 1) in which endometrial adenocarcinoma occurred in at least one first-degree relative (mother, daughter, or sister) with no family member showing colon or ovarian cancer (Table l), the probands were diagnosed at a mean age of 61.8 years (ages 51, 60,
ET AL.
FIG. 1. Pedigrees (A-D) of four families in which at least one firstdegree relative (mother, daughter, sister) had endometrial adenocarcinema but no other cancers. Open squares and circles denote men and women (respectively) with no cancer; closed symbols denote men and women with cancer confirmed by pathology; stippled symbols denote men and women with cancer established by family history alone. The type of cancer is found next to the square or circle: U, endometrial adenocarcinoma; LA, laryngeal cancer; E, esophageal cancer, P, pancreatic cancer, and LU, lung cancer. Probands are denoted by an arrow and age of individual at time of detection is found next to their circle. Numbers within symbols denote numbers of individuals.
65, and 71 years, respectively). This compared to a mean age of 65.3 + SD 12.2 years in 56 probands who had no affected relative. One proband was Black and three were Caucasian. One patient was hypertensive. None had a history of diabetes mellitus, obesity, or exogenous estrogen use. Stage at diagnoses was IB grade I in one case, IB grade II in two patients, and IIIA grade II in the fourth case. In one family (DO, Fig. l), there were two affected first-degree relatives (a mother and sister) and also affected second- and third-degree relatives (maternal aunt and maternal niece). Of note is kindred GM (Fig. l), in which the index case had endometrial adenocarcinema diagnosed at age 65. Her only child, a daughter, had endometrial adenocarcinoma. The daughter verbally confirmed her own diagnosis of endometrial adenocarcinema, but declined signing a release for medical records, In the four pedigrees suggestive of the cancer family syndrome (Fig. 2), the probands were diagnosed at a mean age of 64.5 years (ages 51, 65, 67, and 75 years, respectively). All four probands were Caucasian, two were hypertensive, and none had a history of diabetes mellitus, obesity, or exogenous estrogen use. Stage at diagnosis was IB grade I in two probands, stage IIIC grade II in one patient, and stage IV in the fourth patient. In kindred ET (Fig. 2), the proband was diagnosed at age 65. Her father had colon cancer, and a paternal grand-
HERITABLE
Ptdrgrcc
2A ET
Pcdigrcc
Pedigree
SITE-SPECIFIC
2B : SL
2C TC
ENDOMETRIAL
169
CANCER
Stage of disease was as follows: stage I, n = 41 (73%, with 7 patients having stage IA, 25 patients stge IB, and 9 patients stage IC); stage IIB, n = 2; stage IIIA, n = 3; stage IIIC, n = 1; and stage IV, n = 5. Four additional patients, all with clinical stage I disease, did not undergo surgical staging: 3 had serious underlying medical problems and 1 patient declined treatment. None of the four probands having first-degree relatives with site-specific endometrial adenocarcinoma had multiple primary malignancies. Among the 56 cases of sporadic endometrial adenocarcinoma, 4 had such a history (all pathologically confirmed). One patient had renal cell carcinoma antedating her diagnosis of endometrial adenocarcinoma, a second had breast cancer prior to having endometrial adenocarcinoma, a third had breast cancer 1 year after her diagnosis of endometrial adenocarcinoma, and a fourth had ovarian cancer diagnosed at the time of hysterectomy for endometrial adenocarcinoma. None of these patients had relatives affected in a fashion that would suggest a familial cancer syndrome. DISCUSSION
Pcdrgrtc
ZD SC
FIG. 2. Pedigrees (A-D) of four families in which at least one firstdegree relative (mother, daughter, sister) had endometrial, colon, or ovarian adenocarcinoma. Open squares and circles denote men and women (respectively) with no cancer; closed symbols denote men and women with cancer confirmed by pathology; stippled symbols denote men and women with cancer established by family history alone; stippled symbols with a vertical line denote cancer confirmed by medical record or death certificate. The type of cancer is found next to the square or circle: U, endometrial adenocarcinoma; C, Colon cancer; 0, ovarian cancer; B, breast cancer; S, stomach cancer; LI, liver cancer; LY, lymphoma; SK, skin cancer; and UNK, specific site unknown. Probands are denoted by an arrow and age of individual at time of detection is found next to their circle. Numbers within symbols denote numbers of individuals.
mother had uterine cancer. However, the proband also showed maternal relatives with endometrial cancer. Thus, heritability in this family could reflect either the cancer family syndrome gene(s) transmitted through paternal lineage or a gene(s) predisposing solely to endometrial adenocarcinoma through maternal lineage. Kindred SL (Fig. 2) more plausably showed the cancer family syndrome transmitted through paternal lineage, whereas TC and SC (Fig. 2) showed transmission through maternal lineage. Among the 56 cases of sporadic endometrial adenocarcinoma, the racial composition was as follows: Caucasian, n = 45; Black, n = 8; Hispanic, n = 2; and American Indian, n = 1. Twenty-three patients (41%) were obese, 10 (187 o ) were hypertensive, 8 (14%) had diabetes mellitus, and 4 patients had all three risk factors. Two patients had taken unopposed exogenous estrogens.
Few studies investigating the heritability of endometrial adenocarcinoma have been reported [6-81. In 1966, Lynch et al. [6] reported 154 probands with endometrial carcinoma of whom 20 (13.0%) had a similarly affected first-degree relative. In a pedigree study of 51 patients with endometrial adenocarcinoma, Boltenberg et al. [8] noted a similarly affected mother or sister in 4 cases (8%). Applying similar criteria to our study, 6 of 64 (9.4%) probands with endometrial adenocarcinoma showed a similarly affected first-degree relative. The mode of inheritance cannot be determined definitively, but both our data and those of Lynch et al. [6] and Boltenberg et al. [8] are consistent with autosomal dominant inheritance. Other studies have shown that endometrial adenocarcinoma is one component of the cancer family syndrome. In Boltenberg’s series of 51 probands, there were three families that had pedigrees consistent with the cancer family syndrome, representing 5.8% of all patients with endometrial adenocarcinoma [8]. Our data are only preliminary. Like all studies dealing TABLE 1 Frequenciesof Adenocarcinomaof the Uterus, Colon, and Ovary ObservedamongProbandsand Controls Relatives with endometrial cancer
Probands Controls
Relatives with colon or ovarian cancer
Mothers
Sisters
Parents
Siblings
4/64 (6.3%) o/34
2/106 (1.9%) O/36
2/124 (1.6%) 2/68 (2.9%)
6/128 (4.7%) l/74 (1.4%)
170
SANDLES
with genetic heterogeneity, a large sample size will be required to identify the relatively few kindreds having potential Mendelian tendencies. Nonetheless, we believe two different genetic mechanisms are responsible for the heritable tendencies we and others have observed. In 4 of our 64 (6.25%) families, multiple first-degree relatives had endometrial, colon, or ovarian adenocarcinoma, suggesting the cancer family syndrome (Lynch syndrome II). Another patient had simultaneous endometrial and ovarian adenocarcinoma. Although none of her relatives had a history of colon, ovarian, or endometrial adenocarcinoma, this patient may represent a new germline mutation for the cancer family syndrome. In 4 other families, at least one first-degree relative had endometrial adenocarcinema, but no family member showed colon or ovarian cancer. In 1 of these latter 4 families, two first-degree, one third- and one fourth-degree relative were similarly affected. In aggregate, our findings suggest not only existence of a cancer family syndrome, but existence of a site-specific endometrial adenocarcinoma syndrome as noted by Boltenberg et al. [8]. This would be analogous to the site-specific patterns of inheritance previously described for ovarian cancer [2] and colon cancer [9]. A characteristic of all hereditary cancers is early age of onset [5,7]. Hakala et al. [7] observed the age of onset of endometrial carcinoma in 26 women from 19 families with the cancer family syndrome to be 48.3 years (range, 34 to 63 years). Our data are limited, but to date show only modest differences in age of onset in the four probands having presumptive site-specific endometrial adenocarcinoma compared to sporadic cases (mean ages, 61.8 and 65.3 years, respectively). Another characteristic of hereditary cancer is an increased frequency of multiple primary malignancies. In our series, multiple primary malignancies were observed in none of the four probands in which cancer was restricted to endometrial adenocarcinoma in first-degree relatives. Among the four cancer family syndrome families, a single family member had both colon and ovarian cancer. Again, the sample size is too small to draw definitive conclusions. Among the eight probands in our series with apparent hereditary endometrial adenocarcinoma, five presented with stage IB disease, two with stage III disease, and 1 with stage IV disease. Until a larger series is ascertained, we cannot comment on whether the familial cases of endometrial adenocarcinoma are associated with a more advanced stage at presentation and/or a poorer prognosis compared to sporadic cases. Boltenberg er al. [8] observed that familial cases were associated with more anaplastic tumors and decreased survival. Although mutant genes represent the most likely explanation for our observations of familial tendencies in uterine cancer, other explanations are possible. Such findings could be attributable to chance alone. Unidentified
ET AL.
environmental factors could also exist, although this seems unlikely given that affected women have been raised in multiple households as well as living in different regions. A third possibility relates to reporting errors in pathologically unconfirmed cases. However, one would not expect preferential overreporting of a particular cancer diagnosis (i.e, endometrial adenocarcinoma) in probands as opposed to controls. Nonetheless, we are aggressively pursuing histologic confirmation of diagnoses made in relatives. A fourth possible explanation is misclassification with respect to the diagnosis of site-specific endometrial adenocarcinoma versus the cancer family syndrome. Classification of a kindred as having a sitespecific syndrome could reflect merely the small number of affected family members or the limited time of observation. Over time, other cancers may be expressed; thus, a kindred formerly classified as a site-specific syndrome might actually be shown to represent the cancer family syndrome. These caveats notwithstanding, genetic factors are increasingly recognized as playing integral roles in the etiology of cancer. If a site-specific endometrial adenocarcinoma syndrome exists, it could be characterized by the simple Mendelian (autosomal dominant) inheritance shown by the cancer family syndrome gene. If so, firstdegree relatives would have a 50% risk for inheriting the mutant gene and, hence, possibly developing endometrial adenocarcinoma. Of course, not all inheriting a dominant gene will be affected (incomplete penetrance or variable expressivity, namely atypical endometrial hyperplasia). Nevertheless, increased surveillance among potentially affected female family members may be warranted. Such surveillance might include patient education and counseling, avoidance of periods of anovulation, verification of achieving a secretory endometrium with exogenous hormone replacement, periodic endometrial biopsies, and possibly even prophylactic hysterectomy in selected cases. In the future, the identification of genetic markers, such as specific oncogene mutations or amplification, may make possible identification of affected individuals before phenotypic expression of clinically recognizable cancer. REFERENCES 1. Boring, C. C., Squires, T. S., and Tong, T. Cancer statistics, 1991, Cancer J. Clin. 41, 19-36 (1991). 2. Lynch, H. T., Albano, W., Black, L., Lynch, J. F., Recabaren, J., and Pierson, R. Familial excess of cancer of the ovary and other anatomic sites, J. Am. Med. Assoc. 245, 261-264 (1981). 3. Lynch, H. T., Harris, R. E., Guirgis, H. A., Maloney, K., Carmody, L. L., and Lynch, J. F. Familial association of breast/ovarian carcinoma, Cancer 41, 1543-1549 (1978). 4. Lynch, H. T., Albano, W. A., Lynch, J. F., Lynch, P. M., and Campbell, A. Surveillance and management of patients at high risk for ovarian cancer, Obstet. Gynecol. 59, 589-596 (1982).
HERITABLE
SITE-SPECIFIC
5. Lynch, H. T., Kimberling, W., Albano, W. A., Lynch, J. F., Biscone, K., Schuelke, G. S., Sandberg, A. A., Lipkin, M., Deschner, E. E., Mikol, Y. B., Elston, R. C., Bailey-Wilson, J. E., and Danes, B. S. Hereditary nonpolyposis colorectal cancer (Lynch syndromes I and II). I. Clinical description of resource, Cancer 56, 934-938 (1985). 6. Lynch, H. T., Krush, A. L., and Magnuson, C. W. Endometrial carcinoma: Multiple primary malignancies, constitutional factors, and heredity, Am. J. Med. Sci. 36, 381-390 (1966). 7. Hakala, T., Mecklin, J. P., Forss, M., Jarvinen, H., and Lehtovirta,
ENDOMETRIAL
CANCER
171
P. Endometrial carcinoma in the cancer family syndrome, Cancer 68, 1656-1659 (1991). 8. Boltenberg, A., Furgyik, S., and Kullander, S. Familial cancer aggregation in cases of adenocarcinoma corporis uteri, Actu O&et. Gynecol. Scund. 69, 249-258 (1990). 9. Lynch, H. T., Schuelke, G. S., Kimberling, W. J., Albano, W. A., Lynch, J. F., Biscone, K. A., Lipken, M. L., Deschner, E. E., Mikol, Y. B., Sandberg, A. A., Elston, R. C., Bailey-Wilson, J. E., and Danes, B. S. Hereditary nonpolyposis colorectal cancer (Lynch syndromes I and II). II. Biomarker studies, Cancer 56,939951 (1985).
