Report
Familial and histological analyses of 138 breast cancer patients
Nicole Bürki1, Andrej Gencik2, Joachim K.H. Torhorst3, Walter Weber4 and Hansjakob Müller5 1, 2, 4, 5 Laboratory of Human Genetics, Department of Research of the University Clinics, Basel; 3 Department of Pathology, University of Basel, Switzerland
Key words: familial breast cancer, histological type, tumor associations
Summary
A histological analysis was conducted in 138 female breast cancer patients, and the results were classified in accordance with “Histological Typing of Breast Tumours” (WHO, Geneva 1981). Since about half of these tumors showed more than one histological type of carcinoma, a simplified classification system with four groups was adopted. When patients were categorized according to the number and degree of kinship of their relatives with breast cancer, no specific association with the histological types was found. Familial tumors also encompassed a wide spectrum of histopathologic diagnoses. This suggests the absence of a histological marker in familial breast cancer.
Pedigrees of all the patients were then analyzed, special emphasis being placed on relatives suffering from the same and other malignancies. It was found that 13.8% of the probands had at least one first-degree relative with breast cancer and that, compared with the tumor spectra in the male and female population, there was a significantly higher number of esophageal carcinomas in the fathers, of stomach cancers in the uncles and grandfathers, of brain tumors in the mothers, and of sarcomas in the brothers. An accumulation of the same tumors, especially stomach cancer and tumors related to the SBLA syndrome, was observed in families of index patients with tubular or medullary breast cancer. The SBLA syndrome is a complex familial cancer syndrome characterized by a proclivity to Sarcomas, Breast cancers, brain tumors, Lung and laryngeal cancers, leukemia, and Adrenocortical carcinomas.
Introduction
Familial breast cancer (i.e., breast cancer occur- ring in first degree relatives) tends to occur at an earlier age and is more commonly bilateral than breast cancer in the general population [1]. An increased breast cancer risk for relatives of breast cancer patients has been amply documented [2], and an autosomal dominant pattern of transmis- sion of susceptibility to breast cancer in specific families has been suggested [3].
Very little attention has so far been given to identifying a histological marker for women at high risk of breast cancer. However, when considering the morphology of familial breast tumors, perhaps
the first question that arises is: Why should any specificity of histopathological findings exist - es- pecially when the histology of breast carcinoma shows structural heterogeneity [4], and bilateral tumors in the same individual are often of different histological types [5]? However, single breast tu- mors often show a uniform histological pattern, as observed in this study, and bilateral breast tumors tend to be of the same histological type more often than would be expected [5]. This investigation has been conducted on the assumption that specific genetic factors suspected to be responsible for fa- milial concentrations of breast cancer might also be directly expressed in the morphology.
A generally accepted system of nomenclature
and uniform criteria for pathologic diagnosis are lacking in the literature. Hence, a new system of classification has been described below and it has been used in this study.
Patients, methods and controls
Every year about 1000 new cancer cases are diag- nosed in females in Basel, which has a population of about 440,000; twenty-five percent of these ma- lignancies are breast cancers. During 1982-1984 detailed pedigree data were obtained by a human geneticist (A.G.) from 138 women with invasive breast cancer (index patients) from hospital ser- vices and private practices in Basel. The anamnes- tic data were verified as far as possible by reviewing medical reports. About 55% of the probands were diagnosed as having premenopausal breast cancer. The age at menarche ranged from 9 to 18 years, without a statistically significant difference in the median age of 13 years between the histological or familial subgroups. About 75% of all women had had at least one child. The age at the birth of the first child varied from 16 to 41 years with a median of 27. Other parameters considered for these pa- tients are mentioned elsewhere [6].
The histological slides from all the index patients (familial and sporadic cases) were reviewed by a pathologist (J.T.) and were classified according to “Histological Typing of Breast Tumours” [7]. Only carcinomas with at least one infiltrating part were considered. The histological preparations of 45.8% of the first-degree relatives suffering from breast cancer and of 26.7% of all relatives were also re- viewed.
In accordance with the WHO guide lines, a tu- mor was considered to contain more than one his- tological type if the relevant components ac- counted for more than 10% of the total area. The tumors in the index patients included 59 with only one type of histology, 63 with two, 15 with three, and one with as many as four different types. For the sake of simplicity, classification was established according to the following criteria:
All tumors with an invasive lobular component were classified as “invasive lobular”. Of the other
tumors, those with invasive ductal components were classified as “invasive ductal” while the “tu- bular” carcinomas are pure forms and the remain- ing neoplasms are grouped under “other histo- logies”.
The probands were then categorized into 4 groups according to the number and degree of kinship of their female relatives having breast can- cer:
Group 1: Those with one first-degree relative with breast cancer
Group 2: Those with two or more first-degree relatives with breast cancer
Group 3: Those with no first-degree, but with second and/or third-degree relative/s with breast cancer and
Group 4: Those having no relatives with breast cancer (up to the third degree of kin- ship).
Mothers, sisters, and daughters are first-degree relatives, while grandmothers, aunts, nieces, and granddaughters are second-degree relatives. Third-degree relatives are great-grandmothers, grandaunts, female cousins, grandnieces, and great-granddaughters. The index patients of groups 1 and 2 are designated as familial and those of groups 3 and 4 as sporadic.
In the first part of this investigation, we sought an association between the familial vs. sporadic status of the index patients and the histological type of their breast cancer. In the second part, the oc- currence of tumors other than breast cancer in the families of the 138 index patients is demonstrated. The female and male tumor spectrum from the population-based Basel Cancer Registry (J.T.) with 6487 female and 7076 male cases in the years 1975-1982 (skin tumors other than melanomas were excluded) was used to establish controls. Fi- nally, the pedigrees of patients with tubular or medullary breast cancer were analyzed in more detail.
The patient information was computerized. Sta- tistical analyses were carried out using the chi- square test and, for small numbers, Fisher’s exact test. The Kolmogorov-Smirnov test was used for the comparison of age-distribution tables.
Results and discussion
The distribution of index patients within the four familial groups based on the number and kinship degree of their relatives with breast cancer is shown in Table 1. Fifty-three probands had a positive family history of breast cancer, whereas only nine- teen had at least one first-degree relative with breast cancer.
The probands showing the different histological types of breast cancer according to our classifica- tion system are listed in Table 2. The distribution of the main histological types in index patients grouped according to Table 1 is summarized in Table 3. There is no significant association between a histological type and a positive family history of breast cancer. In addition, a comparison of the histological pattern was made between the two groups of familial and sporadic cases (Table 4). However, the differences were not statistically sig-
nificant. The tumors of index patients with familial breast cancer also encompassed a wide spectrum of histopathological diagnoses and displayed an unex- pectedly high frequency of structural heterogen- eity.
In order to investigate a possible association be- tween a specific histological type of breast cancer and a particular family history, the classification problem of tumors showing more than one histo- logical pattern had to be resolved. In this study only 42.8% of all cancers were of a single histological type, while in the investigations of Mulcahy and Platt [8], for example, there were 53.5%, and in those of Fisher et al. [9] 69.9%. The histological categories in this study were adopted only arbitrari- ly, the first category being of special interest be- cause of a suspected variance in its frequency (in- vasive lobular carcinoma). Thirty-three index patients had a tumor whose histology was at least partly of the invasive lobular type; in fourteen of
| Group | Number and degree of relatives with breast cancer | Index patients | |
|---|---|---|---|
| n | % | ||
| 1 | One first-degree relative | 15 | 10.9 |
| 2 | Two or more first-degree relatives | 4 | 2.9 |
| 3 | No first-degree, but second and/or third-degree relatives | 34 | 24.6 |
| 4 | No relatives | 85 | 61.6 |
| Total | 138 | 100.0 | |
| Histological type of breast cancer | Index patients | |
|---|---|---|
| n | % | |
| Invasive lobular | 33 | 23.9 |
| Invasive ductal | 93 | 67.4 |
| Tubular | 5 | 3.6 |
| Others | 7 | 5.1 |
| Total | 138 | 100.0 |
Note the high rate of occurrence of invasive lobular carcinoma among the index patients (in other series around 12% only).
them this was not compounded with another invas- ive type of histology, but eighteen of them also showed an invasive ductal carcinoma component in the same tumor and one was combined with tubu- lar carcinoma. Among the remaining probands ninety-three tumors showed at least one compo- nent with an invasive ductal histology, namely ninety-one of them without a combination and two compounded with tubular carcinoma. There re- mained five pure tubular carcinomas forming a class of their own, and four medullary carcinomas, two mucinous carcinomas, and one large cell car- cinoma are presented in “Other histologies”.
The occurrence of the different histological types in this study is compared with the data of Haggie et al. [10], Mulcahy and Platt [8], Mc Divitt et al. [11] and Lagios et al. [12] in Table 5. For the comparison of these studies using different definitions of “fa-
milial”, the proposal of Lynch [13] was used. He designated all probands familial who had at least one relative of the nuclear pedigree with breast cancer. The nuclear pedigree on the female side comprises mothers, sisters, daughters, aunts, and grandmothers. In all these studies the sporadic cases were not true matched controls to the familial cases. In the present study and that of Mulcahy and Platt [8], invasive lobular carcinomas are observed more often, in comparison with the other investiga- tions, than would be expected in both the familial and the sporadic cases, but no association with a positive family history of breast cancer is reported. McDivitt et al. [11] described a much higher inci- dence of invasive ductal carcinomas among breast cancer than was observed in the three family his- tory studies. Mulcahy and Platt [8] cited the small- est proportion of invasive ductal carcinoma, al-
| Group* | Histological type of breast cancer | |||||||
|---|---|---|---|---|---|---|---|---|
| Inv. lobular | Inv. ductal | Tubular | Others | |||||
| n | % | n | % | n | % | n | % | |
| 1 (n = 15) | 1 | 6.7 | 11 | 73.3 | 1 | 6.7 | 2 | 13.3 |
| 2 (n = 4) | 1 | 25.0 | 3 | 75.0 | - | - | ||
| 3 (n = 34) | 10 | 29.4 | 24 | 70.6 | - | - | ||
| 4 (n = 85) | 21 | 24.7 | 55 | 64.7 | 4 | 4.7 | 5 | 5.9 |
| Total (n= 138) | 33 | 23.9 | 93 | 67.4 | 5 | 3.6 | 7 | 5.1 |
* Refer to Table 1.
| Histological type of breast cancer | ||||||||
|---|---|---|---|---|---|---|---|---|
| Inv. lobular | Inv. ductal | Tubular | Others | |||||
| n | % | n | % | n | % | n | % | |
| Familial cases (n = 19) | 2 | 10.5 | 14 | 73.7 | 1 | 5.3 | 2 | 10.5 |
| Sporadic cases (n = 119) | 31 | 26.0 | 79 | 66.4 | 4 | 3.4 | 5 | 4.2 |
| Total (n = 138) | 33 | 23.9 | 93 | 67.4 | 5 | 3.6 | 7 | 5.1 |
The “familial” cases have at least one first-degree relative with breast cancer.
though they did not have the highest frequency of invasive lobular carcinoma. In this histological group there also was no association with a special familial status of mammary carcinoma. The tubular carcinomas of Lagios’ series [12] showed a signifi- cant association with family history (p<0.05), as well as tendencies toward multicentricity and bilat- erality which could not be confirmed in the present study because of the small number of cases. Haggie et al. [10] also reported more tubular carcinomas than anticipated among the familial cases (not sig- nificant according to the Lynch definition). Mulca- hy and Platt [8] observed a very high frequency of medullary carcinoma with a statistically significant proportion of these carcinomas occurring among familial cases (p<0.05), while Haggie et al. [10] reported only a slightly larger number than ex- pected in these cases. These tendencies likewise could not be evaluated in this study, in that there were only four cases of medullary carcinoma.
Occurrence of tumors other than breast cancer in the families of index patients with breast cancer
Finally, we wanted to know how often relatives of the 138 index patients had carcinomas. We were also interested to know the type and the frequency of occurrence, in order to correlate familial cancer with the occurrence of breast cancer. Pedigree ana- lyses were done on all probands with the data collected from the interviews and hospital records. The pertinent results are presented below. Skin tumors were not included in the study because their occurrence could not be ascertained from the avail- able sources of information. However, melanomas were included. The female and male tumor spec- trum from the population-based Basel Cancer Registry (J.T.) with 6487 female and 7076 male cases in the years 1975-1982 (without skin tumors) was used to establish controls. The estimated oc- currence of the various cancers found in the rela- tives is the following:
1. Breast cancer was more frequent than the ex- pected rate of 29% of all tumors in the popula- tion in the mothers of the index patients (42%),
| Patients with | n | Total % | Familial cases % | Sporadic cases % | Study |
|---|---|---|---|---|---|
| Invasive lobular carcinoma | |||||
| 138 | 23.9* | 22.6 | 24.7 | this study | |
| 273 | 11.0 | 9.6 | 11.4 | Haggie [10] | |
| 129 | 18.6* | 17.3 | 20.4 | Mulcahy [8] | |
| 1458 | 8.7 | McDivitt [11] | |||
| Invasive ductal carcinoma | |||||
| 138 | 67.4 | 71.7 | 64.7 | this study | |
| 273 | 69.6 | 71.0 | 69.2 | Haggie [10] | |
| 129 | 50.4* | 48.0 | 53.7 | Mulcahy [8] | |
| 1458 | 78.1* | McDivitt [11] | |||
| Tubular carcinoma | |||||
| 138 | 3.6 | 1.9 | 4.7 | this study | |
| 273 | 4.8 | 8.1* | 3.8 | Haggie [10] | |
| 211 | 7.6* | 16.2* | 5.7 | Lagios [12] | |
| Medullary carcinoma | |||||
| 138 | 2.9 | 1.9 | 3.5 | this study | |
| 273 | 1.5 | 3.2* | 0.9 | Haggie [10] | |
| 129 | 10.9* | 16.0* | 3.7 | Mulcahy [8] | |
| 1458 | 4.3 | McDivitt [11] |
* Discussed percentages in the text. The “familial” cases have at least one relative of the nuclear pedigree with breast cancer.
in the sisters (40%), in the maternal aunts (38%), and in the paternal aunts (42%). The apparent higher number than expected was not, however, statistically significant.
2. Although percentages of gynecological tumors (without breast cancer) were higher in the mothers (26%), sisters (24%), and maternal aunts (26.5%) than in the population (16.5%), they also were not statistically significant.
3. Among the gastrointestinal malignancies the number of esophageal carcinomas in the fathers showed a significantly higher excess by compari- son with the anticipated number (11% in com- parison to 1.6% in the population; p<0.0001), as did that of stomach cancer in the uncles (19.5%; p<0.001) and grandfathers (29%; p<0.0001), compared to the control group (6.2%). The proportion of colorectal carcino- mas, however, was not higher than expected. (It should be noted, that only one third of the stomach cancers could be histologically ver- ified.)
4. An association with neoplasms belonging to the
spectrum of the Li-Fraumeni/SBLA-Syndrome was documented by a statistically significant ex- cess of brain tumors in the mothers (6.5% against 1.5% in the population; p<0.05) and sarcomas in the brothers (11.8% in comparison to 0.9%; p<0.0001).
It has been reported that there has been an associ- ation of familial breast cancer with the histologies of tubular [10, 12] and medullary [8, 10] carcinomas. We were unable to demonstrate a correlation be- tween a family history of breast cancer and the occurrence of tubular and medullary carcinomas. Therefore, the pedigrees of these patients were analyzed in more detail and many tumors other than breast cancer were found. Table 6 summarizes the findings among the 8 families of index patients with tubular carcinoma and Table 7 those among the probands having medullary carcinoma of the breast. It is apparent that particularly those malig- nancies with more than the expected frequency in the 138 families compared to the frequency in the population, very often occurred in families of index patients with tubular or medullary breast cancer.
| Pedigree | Histological combination | Family member suffering from a malignancy | Type of malignancy |
|---|---|---|---|
| A | tubular | - | - |
| B | tubular | son | sarcoma |
| C | tubular | mother | uterine carcinoma |
| grandmother | skin tumor | ||
| D | tubular | sister | breast cancer (muc.) |
| half-aunt | leukemia | ||
| *E | tubular | mother | colonic cancer |
| father | esophageal cancer | ||
| uncle | liver carcinoma | ||
| *F | tub., i. lob. | mother | breast cancer (i. duc) |
| father aunt | liver carcinoma sarcoma | ||
| G | tub., i. duc. | mother | colonic cancer |
| *H | tub., i. duc. | mother | uterine carcinoma |
| father | stomach cancer | ||
| uncle | stomach cancer |
muc. = mucinous, tub. = tubular, i. lob. = invasive lobular, i. duc. = invasive ductal.
* The pedigrees of these families are drawn in the appendix (Fig. 1 to 3).
| Pedigree | Histology | Family member suffering from a malignancy | Type of malignancy |
|---|---|---|---|
| *I | medullary | female cousin | breast cancer (i. duc) |
| grand-aunt | breast cancer | ||
| sister | Hodgkins lymphoma | ||
| mother | colonic cancer | ||
| father | leukemia | ||
| J | medullary | father | prostatic carcinoma |
| uncle | prostatic carcinoma | ||
| female cousin | Hodgkins lymphoma | ||
| K | medullary | grandfather | stomach cancer |
| *L | medullary | mother | stomach cancer |
| aunt | brain tumor | ||
| uncle | stomach cancer | ||
| grandfather | stomach cancer |
i. duc. = invasive ductal carcinoma.
* The pedigrees of these families are drawn in the appendix (Figs 4 and 5).
In conclusion, in approximately 56.5%, the his- tological analysis revealed a heterogeneity of the breast carcinomas with the same tumor showing more than one histological type. To facilitate fur- ther comparisons between familial and sporadic cases, a simplified classification system which con- sisted of four groups was adopted. When patients were categorized according to the degree of kinship and the number of relatives with breast cancer, no specific association with the histological types was found. Familial tumors also encompassed a wide spectrum of histopathological diagnoses. This may suggest the absence of a histological marker in familial breast cancer. Among the relatives of the patients, the number of esophageal carcinomas in the fathers, of stomach cancers in the uncles and grandfathers, of brain tumors in the mothers, and of sarcomas in the brothers, was found to be signifi- cantly higher than expected from the tumor spectra in the population. An accumulation of the same tumors, especially stomach cancer and tumors re- lated to the SBLA syndrome, was observed in families of index patients with tubular or medullary breast carcinoma. Therefore there may be a cor- relation between tubular or medullary breast can- cer and an elevated risk of familial malignancies.
This possibility is of sufficient interest to warrant further investigations.
Acknowledgements
The authors are grateful to Dr. M. Buser, Pro- fessor L.R. Emmons, Dr. T. Kuttapa and Mr. B. Temminck for their untiring help. This work was supported by the Swiss National Fund No. 3.818.0.84.
Appendix
14
Li
E 51
C 56
+60
+51
+ 56
B 41 tub.
Fig. 1. Pedigree E
The index patient is suffering from a premenopausal tubular breast cancer on the right side. Both of her parents also have developed a malignancy at an early age; her mother had a colonic cancer and her father an esophageal cancer. Both died in the year of diagnosis of their tumor. The proband also had a paternal uncle who died of a liver carcinoma at age 60. It is not known how far non-genetic factors, such as alcohol, played a role in the development of cancer on the paternal side of this family.
Sa 41 Li 57 + 41 +57
B 48 duc.
B 30 tub.
Fig. 2. Pedigree F The index patient herself is suffering from bilateral premeno- pausal breast cancer with tubular histology. Her mother also had a breast cancer, but of invasive ductal histology, at the age of 48 years. The earlier age at diagnosis in the daughter than in the mother is seen here very clearly. The father of the proband had a liver carcinoma diagnosed at the age of 57 years of which he died the same year. His sister also suffered from a malig- nancy, namely a sarcoma, of which she died at 41. It seems likely that in this family a genetic factor plays a role in the develop- ment of malignancy.
St 82 St 71 + 82 + 71
Ut 70 + 70
B 46 tub.
and her father as well as her paternal uncle died of stomach cancer. The malignancies in these relatives occurred relatively late, but there is nevertheless a substantially increased incidence of tumors in this family.
B
Le +73
C
CSU +80 +62
B
B 56 MH 23 med. + 23
B 32 duc.
The index patient has right-sided postmenopausal breast cancer of medullary histology. Her sister died very early of a Hodgkins lymphoma at the age of 23. Her mother died at 80 of a colonic cancer and her father at 73 of leukemia. The twin brother of her mother also died of a malignancy, but the site is unknown. There are still other relatives suffering from breast cancer in this family: a maternal grand-aunt (histology not available) and a maternal female cousin. The latter developed her invasive ductal breast cancer very early at the age of 32. It is to be noted that her mother, who is a relative of the proband only through marriage, also has breast cancer. This family, with many dif- ferent and in some relatives uncommon malignancies, seems to have a high probability of a familial cancer syndrome.
82 St 22
Br 52 + 52
Stól
St +68
2
2
2
B 45 med.
Pedigrees of probands with tubular or medullary carcinoma of the breast and a pertinent accumulation of other tumors among their relatives.
The circles stand for women and the squares for men. The crossed family members have already died. The circle marked with an arrow points out our index patient. Solid circles and squares mark persons with a histologically proved malignancy and the signs with a bar only designate family members with a tumor of undetermined histology. In the case of breast cancer patients the small spots indicate the side of the tumor. The letter below persons with a malignancy is an abbreviation for the kind of tumor, followed by a number indicating the age at diagnosis. If the patient has already died, this is noted a line below with a ”+” followed by the age at death. The abbreviations of the malignancies are the following: B = breast cancer, Br = brain carcinoma, C = colonic cancer, CSU = cancer site unknown, E = esophageal cancer, Le = leukemia, Li = liver carcinoma, MH = Hodgkins lymphoma, Sa = sarcoma, St = stomach can- cer, Ut = uterine carcinoma.
With histologically verified breast cancers, the histological type is also mentioned: lob. = invasive lobular carcinoma, duc. = invasive ductal carcinoma, tub. = tubular carcinoma and med. = medullary carcinoma.
References
1. Anderson DE: Breast cancer in families. Cancer 40: 1855- 1860, 1977
2. Anderson DE: Breast cancer. In: Müller HJ, Weber W, Kuttapa T (eds) Familial Cancer. Karger, New York, 1985, pp 16-19
3. Gardner EJ, Stephens FE: Breast cancer in one family group. Am J Hum Genet 2: 30-40, 1950
4. Qualheim RE, Gall EA: Breast carcinoma with multiple sites of origin. Cancer 10: 460-468, 1957
5. Robbins GF, Berg JW: Bilateral primary breast cancers. A prospective clinicopathological study. Cancer 17: 1501- 1527, 1964
6. Bürki N: Familiäres Mamma-Karzinom unter spezieller Berücksichtigung der Histologie. Thesis, Basel, 1987
7. WHO: Histological typing of breast tumors. WHO, Ge- neva, 1981
8. Mulcahy GM, Platt R: Pathologic aspects of familial car- cinoma of breast. In: Lynch HT (ed) Genetics and Breast Cancer. Van Nostrand Reinhold Co., New York, 1981, pp 65-97
9. Fisher ER, Gregorio RM, Fisher B, et al .: The pathology of invasive breast cancer. A syllabus derived from findings of the National Surgical Adjuvant Breast Project (Protocol No. 4). Cancer 36: 1-85, 1975
10. Haggie JA, Birch JM, Howell A, Sellwood RA: Epidemiol- ogy of premenopausal familial mammary cancer in North West England (Poster). IVth International Congress on Senology and Breast Diseases, Paris, 1986
11. McDivitt RW, Stewart FW, Berg JW: Tumors of the breast. In: Atlas of Tumor Pathology, Second Series, Fasci- cle 2. Armed Forces Institute of Pathology, Washington, D.C., 1968, pp 156
12. Lagios MD, Rose MR, Margolin FR: Tubular carcinoma of the breast. Association with multicentricity, bilaterality, and family history of mammary carcinoma. Am J Clin Pathol 73: 25-30, 1980
13. Lynch HT, Albano WA, Danes S, Layton MA, Kimberling WJ, Lynch JF, Cheng SC, Costello KA, Mulcahy GM, Wagner CA, Tindall SL: Genetic predisposition to breast cancer. Cancer 53: 612-622, 1984
Address for offprints: Nicole Bürki, Laboratory of Human Ge- netics, Department of Research at the University Hospital, Hebelstrasse 20, CH-4031 Basel, Switzerland