629 Male Breast Carcinoma A Review of 229 Patients Who Presented to the Princess Margaret Hospital during 40 Years: 1955–1996 Paul E. Goss, M.D., Ph.D.1 Caroline Reid, M.Sc.1 Melania Pintilie, M.Sc.2 Ruth Lim, B.Sc.1 Naomi Miller, M.D.3 1 Department of Medical Oncology and Hematology, The Princess Margaret Hospital, Toronto, Ontario, Canada. 2 Department of Biostatistics, The Princess Margaret Hospital, Toronto, Ontario, Canada. 3 Department of Pathology, The Princess Margaret Hospital, Toronto, Ontario, Canada. BACKGROUND. A single-institution review of clinical presentation, treatment, and outcome of male breast carcinoma was conducted. METHODS. Data obtained by chart review of 229 cases were analyzed with respect to clinical presentation, treatment choice, significant prognostic factors, and survival. The patients were analyzed both as a single cohort and as four cohorts grouped according to decade of diagnosis. RESULTS. Presentation occurred at a median age of 63 years, most often with a self-detected lump. Pathology consisted of subtypes similar to those of female breast carcinoma. The majority of tumors were larger than 2 cm in greatest dimension. Lymph node status, hormone receptors, and histologic and nuclear grade were underreported. Primary, adjuvant, and advanced disease treatment practices were reviewed over time. The 5-year disease free survival (DFS), overall survival (OS), and local control were 47%, 53%, and 91%, respectively. No difference in outcome by decade of diagnosis was observed. Negative lymph nodes and adjuvant hormone treatment predicted for better DFS and OS. Younger age and Stage 0 also predicted for better OS. CONCLUSIONS. Compared with data from female breast carcinoma patients, 5-year OS for this series was low; however, when these patients were separated by lymph node status, survival was similar for those with axillary lymph node metastases. Despite a change in standard primary surgical treatment and an increased use of chemotherapy and hormone therapy over the study period, no difference in outcome was observed among these males. In the absence of prospective, randomized clinical trials, collection of comprehensive data on the presentation and management of male breast carcinoma may help to optimize clinical care. Cancer 1999;85:629 –39. © 1999 American Cancer Society. KEYWORDS: male breast carcinoma, clinical presentation, treatment, survival. R Presented in poster form at the May 1997 meeting of the American Society of Clinical Oncology, Denver, Colorado. Address for reprints: Paul E. Goss, M.D., Ph.D., Princess Margaret Hospital, 610 University Avenue, 5-303, Toronto, Ontario, Canada M5G 2M9. Received April 1, 1998; revision received July 17, 1998; accepted July 17, 1998. © 1999 American Cancer Society elative to the disease in women, male breast carcinoma is rare, accounting for less than 1% of all cases of breast carcinoma and with an incidence of 1 in 100,000 men.1 Each year approximately 1000 cases are diagnosed in the U.S., resulting in roughly 300 deaths annually.2 Previous reports have drawn similarities between breast carcinoma in men and women with regard to presentation and sites of distant metastases.3,4 In addition, there is clinical evidence that, as in females, tumor size and axillary lymph node involvement are important prognostic factors in male breast carcinoma.5 Histopathologic subtypes are also similar in men who have infiltrating ductal carcinomas with or without an intraductal component observed most frequently.6 Lobular carcinoma, previously not thought to occur in men due to lack of lobular differentiation, has now been reported in a few cases.7 630 CANCER February 1, 1999 / Volume 85 / Number 3 A number of factors have been considered to be related to increased risk for breast carcinoma in men. A family history of breast carcinoma,1,8 –10 clinical conditions causing hypoandrogenism (including Klinefelter’s syndrome, testicular trauma, or infertility),11,12 liver disease causing hyperestrogenism,13 and gynecomastia have all been implicated as possible risk factors. Unfortunately, these factors are not consistently reported; therefore, their relation to the disease is unclear. Development of optimal treatment strategies specific to male breast carcinoma is limited by a lack of traditional large scale clinical trial data. As a result, current clinical management is generally extrapolated from principles established for the treatment of female breast carcinoma. In the absence of sufficient cases necessary to conduct informative prospective clinical trials, reviews of groups of male breast carcinoma patients, accumulated over time, provide a means to study the clinical presentation of the disease and to assess the efficacy of various treatments. We report here on a large series of male breast carcinoma cases with long term follow-up presenting to our institution over a period of 42 years. Patient history, including recorded risk factors, clinical presentation, treatment, and outcome, are described for the group as a whole. In addition, we have examined our patient population for changes in clinical presentation over time and for any relation between disease outcome and patient survival with changes in management over the 4 decades. METHODS Men presenting to the Princess Margaret Hospital (PMH) with breast carcinoma were ascertained from patient registration records from 1955–1996. Some of these cases have been described previously.14,15 Data regarding patient history, presenting signs and symptoms, risk factors, primary tumor pathology, initial treatment (surgery, chemotherapy, and radiation), details of recurrence, treatment for recurrence, and survival were obtained by chart review. Data extraction was designed and analyzed by the department of biostatistics at PMH. The data was analyzed both as a single cohort and as 4 cohorts grouped according to decade of diagnosis (1955– 1965, 1966 –1975, 1976 –1985, and 1986 –1996). The Kruskal–Wallis test was used to test for differences in age, weight, and height between decades. Most clinical characteristics of the population were described using tables for categoric variables. Age was considered a continuous variable and was described by the mean, median, and range. Associations between clinical factors and patient history or treatment choice were reviewed in cross-tabulation form and tested for significance with the chi-square test. The Wilcoxon rank sum test was used to compare the duration of symptoms between patient groups with or without a family history of cancer. Influence of clinical factors on treatment choice was tested in a logistic model16 using a stepwise selection technique. Cases missing clinical factor information were excluded from this analysis. Relapse or death due to any cause were considered factors in disease free survival (DFS). Death due to any cause was considered a factor in overall survival (OS). Local relapse was the only event considered in determining the local control rate (LCR). Survival curves were drawn based on Kaplan– Meier estimates,17 and differences were assessed with the log rank test.18 Fifteen patients with breast tumor pathologies other than adenocarcinoma were omitted from outcome analysis. These included pure ductal carcinoma in situ (DCIS), pure Paget’s disease, sarcoma, medullary and gelatinous carcinoma, and papillary adenocarcinoma. The 14 patients with unknown pathologies were included in the outcome analysis results presented here, because no major differences were observed when they were excluded. In addition to 4 patients for whom disease status at last follow-up was unknown, 26 patients with metastatic disease at diagnosis and 15 patients for whom M stage at diagnosis was unknown were excluded from disease free survival and local control analysis. Prognostic factors were tested with the Cox proportional hazards model19 using the stepwise selection technique. Cases with any missing prognostic factor information were excluded from this multivariate analysis. The group excluded due to missing data were not significantly different in terms of OS, DFS, or LCR by the log rank test. Hormone receptor status and family cancer history were not included in the multivariate analysis because this information was not reported in a large number of cases. The variables describing whether surgery was performed and type of surgery were also excluded, because those who did not have surgery or had lumpectomies were often missing data on either lymph node status or tumor size. Inclusion of these factors in the analysis did not result in any major differences in the results. RESULTS Two hundred twenty-nine male breast carcinoma patients were treated over the 42-year period. The incidence in our patients of factors previously reported to be associated with an increased risk for male breast carcinoma are reported in Table 1. Clinical gynecomastia was recorded in 15 cases (6.6%). Family cancer information, either positive or negative, was reported in only 51.5% of cases overall. However, the frequency of family history reporting increased from only 19.4% in the first decade of the study period to 62.1% by the most recent decade. There were no patients with clin- A Review of 229 Male Breast Carcinoma Cases/Goss et al. TABLE 1 Occurrence of Putative Risk Factors in This Series 631 TABLE 2 Patient Characteristics at Presentation Risk factor No. of cases (%)a History of physical trauma History of alcohol abuse History of smoking (ⱖ10 pack-years) Prior breast radiation therapy No offspring Unilateral clinical gynecomastia Bilateral clinical gynecomastia Other malignancy (prior or subsequent) Family history of breast carcinoma 21 (9.2%) 22 (9.6%) 62 (27.1%) 7 (3.1%) 21 (9.2%) 10 (4.4%) 5 (2.2%) 56 (24.5%)b 23 (19.5%)c a Percentage of 229 cases, unless otherwise indicated. b Ten colon, 4 head and neck, 14 skin, and 28 unspecified. Twenty-five cases occurred before breast cancer diagnosis. c Out of 118 cases. A first-, second-, or third-degree relative with breast carcinoma was considered a positive family history. ical Klinefelter’s syndrome or a history of unusual exposure to radiation (e.g., occupational). Patient characteristics at presentation are summarized for the group as a whole in Table 2. The median age at diagnosis for the entire group was 63 years. When patients were grouped according to decade of diagnosis, age and height did not vary significantly. An increase in average weight was observed over the study period (Kruskal–Wallis test, P ⫽ 0.0041). Initial detection of breast carcinoma occurred in most cases due to a self-detected lump (60.7%) or the appearance of breast symptoms, such as nipple discharge (29.3%). The manner of detection did not vary significantly over the study period. Only 27 cases (11.8%) were investigated mammographically. The majority of these mammograms (24 cases) were performed on cases diagnosed within the last two decades of the study. The most common sign or symptom at first examination was a palpable breast mass. All but 3 patients had some sign or symptom (Table 2), with 62.9% presenting with 2 or more symptoms. Delay from first appearance of symptoms to first physician contact was ⬎6 months for 36.0% of the 203 patients for whom this information was available (median, 4 months). The duration of symptoms was found to be significantly shorter (Wilcoxon’s rank sum test, P ⫽ 0.0025) for patients with a family history of cancer (median, 1 month) compared with patients who did not have a family history of cancer (median, 4 months). No significant difference in age at diagnosis was observed when patients were separated according to family cancer history. Tumors occurred most frequently (44.1% of 152 cases with tumor location information) in the central subareolar region. The second most common site was Median yrs of age at diagnosis (range) Median height (range), n ⫽ 155 Median weight (range), n ⫽ 191 Ethnic origin, n ⫽ 168 White Black Asian Other Initial detection of breast carcinoma Self-detected lump Symptoms (i.e., pain, nipple discharge) Physician-detected lump Other or unknown Presenting signs and symptoms Breast mass Nipple ulceration Nipple bleeding Nipple discharge Nipple retraction Local pain Inflammatory skin Other No symptoms Breast carcinoma laterality Right Left Bilateralb a b No.a Range or % 63 171.5 cm 78.8 kg (23–97) (150.0–200.7) (48.8–120.0) 156 3 4 5 (92.8%) (1.8%) (2.4%) (3.0%) 139 67 12 11 (60.7%) (29.3%) (5.2%) (4.8%) 196 18 20 13 60 22 11 70 3 (85.6%) (7.9%) (8.7%) (5.7%) (26.2%) (9.6%) (4.8%) (30.6%) (1.3%) 101 126 2 (44.1%) (55.0%) (0.9%) Unless otherwise indicated, n ⫽ 229. Of these 2 cases, 1 was synchronous and 1 was metachronous. the upper outer quadrant (25.7%). Other tumor characteristics are shown in Table 3. Infiltrating ductal carcinoma with or without other associated histologies was the most common tumor type, occurring in 147 cases (64.2%). Lobular carcinoma was identified in 6 cases (2.6%). From 1955–1965, 36.1% of cases had no known tumor size and 13.9% of cases had tumors larger than 5 cm. By 1986 –1996, the number of cases with no reported tumor size had decreased to 6.1%, and only 4.5% of tumors were larger than 5 cm. Axillary lymph nodes were not assessed in 68 cases. The ratio of cases found to have negative versus positive lymph nodes remained fairly constant over the study period. Importantly, the proportion of cases with no pathologic lymph node assessment decreased with each decade, from 44.4% for 1955–1965 to 15.1% for 1986 –1996. Hormone receptor status was not reported at all until the last two decades of the study period (28.1% of cases for 1976 –1985 and 69.7% of cases for 1986 –1996). Over the entire period histologic grade and nuclear grade were reported in only 31 and 28 cases, respectively. Twenty-seven patients (12.7%) had Stage IV disease at presentation with bone as the most 632 CANCER February 1, 1999 / Volume 85 / Number 3 TABLE 3 Tumor Characteristics Characteristic Histologic tumor type (n ⫽ 215) Infiltrating ductal Infiltrating ductal with DCIS Infiltrating lobular Mixed infiltrating ductal and lobular Infiltrating lobular with DCIS, LCIS, and papillary DCIS Inflammatory Paget’s disease of the nipplea Papillary adenocarcinoma Papillary with DCIS Otherb Unspecified adenocarcinoma Tumor Size (n ⫽ 185) ⬍1 cm 1–2 cm 2–5 cm ⬎5 cm Multifocality (n ⫽ 148) Marginsc (n ⫽ 160) Negative Positive (margins reexcised) Positive (no further surgery) Lymph node status (n ⫽ 161) Positive Negative Positive receptor status Estrogen (n ⫽ 66) Progesterone (n ⫽ 65) No. % 105 38 4 1 1 4 1 6 3 1 9 42 48.8% 17.7% 1.9% 0.5% 0.5% 1.9% 0.5% 2.8% 1.4% 0.5% 4.2% 19.5% 6 63 98 18 10 3.2% 34.0% 53.0% 9.7% 6.8% 113 29 18 70.6% 18.1% 11.3% 92 69 57.1% 42.9% 60 49 90.9% 75.4% a Paget’s associated with unspecified adenocarcinoma in 2 of 6 cases; associated with infiltrating ductal in 3 of 6 cases; pure Paget’s in 1 of 6 cases. b Other pathologies include sarcoma, epithelioid angiosarcoma, fibrosarcoma or stromal sarcoma, infiltrating polygonal carcinoma, medullary carcinoma, trabecular duct carcinoma, gelatinous carcinoma, anaplasia. c In 57 cases pathology reports did not comment on margin status, and in 12 cases surgery was not performed. common site of metastases (occurring in 16 cases). The proportion of cases with metastatic disease at diagnosis remained constant for each decade of diagnosis (data not shown). Primary treatment consisted of surgical resection in 94.8% of cases. The majority (73.3%) underwent a simple or modified radical mastectomy. Almost half of all patients (47.2%) diagnosed during the period 1955– 1965 had a radical mastectomy. By the third decade (1976 –1985), however only 5.6% had a radical mastectomy, and during the period 1986 –1996 no patients were treated with radical mastectomy. Following the first decade, simple or modified radical mastectomy became the most popular surgical technique and was used with increasing frequency over the last 2 decades of the study period. Lumpectomy with or without axillary dissection was uncommon throughout the study period, with only 8 cases in the entire group (3.5%) treated with lumpectomy with axillary dissection and 12 (5.2%) treated with lumpectomy alone. Twelve patients (5.2%) did not have surgery, and in 1 case the type of surgical intervention was not described. The 12 patients who did not undergo surgery as part of their primary treatment either had Stage IV disease at diagnosis or were older than 71 years. Initial treatment regimens included radiation for 131 patients (57.2%), chemotherapy for 21 (9.2%), or additive hormone therapy (i.e., tamoxifen) for 57 patients (24.9%). Nine patients, all of whom were diagnosed in the first three decades of the study period, underwent orchiectomy. In 7 cases the orchiectomy was performed following their breast carcinoma diagnosis. Two patients, both with previous cancer diagnosis (prostate and testicular), were orchiectomized prior to their breast carcinoma diagnosis. Primary and adjuvant treatment combinations are presented per decade and for the group as a whole in Table 4. For the first 3 decades, primary treatment most frequently consisted of surgery plus radiation. Surgery alone was the next most common primary treatment choice. From 1986 to 1996, however, the first and second most popular treatment regimens were surgery plus radiation plus hormone therapy and surgery plus hormone therapy. Of the 57 patients receiving additive hormone therapy, 75.4% of these were diagnosed between 1986 and 1996. Approximately half of patients receiving additive hormone therapy (50.9%) continued treatment for less than 2 years. Most of the 21 patients treated with chemotherapy at this stage were diagnosed during the last two decades of the study period (90.5% from 1976 to 1996). Thirteen patients received regimens in which cyclophosphamide, methotrexate, and 5-fluorouracil were combined. Only two patients (both diagnosed between 1986 and 1996) received anthracyclines. The type of chemotherapy was unknown in four cases. Age, lymph node status, tumor size, presence or absence of metastatic disease at diagnosis, and decade of diagnosis were shown to influence significantly the choice of primary treatment modality by logistic regression analysis. Patients were more likely to be treated with radiotherapy if they were diagnosed before 1976 (P ⫽ 0.0001), if their tumor was large (P ⫽ 0.0003) or if their disease status was less advanced than Stage IV (P ⫽ 0.0002). Because most chemotherapy use occurred in the last 2 decades of the study period, analysis of the influence of clinical factors on chemotherapy use was based on the subset of patients diagnosed between 1976 and 1996. Patients were more likely to be treated with chemotherapy if they had positive lymph nodes (P ⫽ 0.002) or if they were A Review of 229 Male Breast Carcinoma Cases/Goss et al. 633 TABLE 4 Primary and Adjuvant Treatment Combinations Decade of diagnosis Treatment combinationa 1955–1965 (n ⴝ 36) 1966–1975 (n ⴝ 56) 1976–1985 (n ⴝ 71) 1986–1996 (n ⴝ 66) All decades (n ⴝ 229) Surgery alone Surgery and radiation Surgery and hormone therapyb Surgery and chemotherapy Surgery, radiation, and hormone therapy Surgery, chemotherapy, and hormone therapy Radiation and hormone therapy 11.1% 69.4% 5.5% 2.8% 5.5% 0% 2.8% 17.9% 71.4% 3.6% 0% 0% 0% 1.8% 32.4% 39.4% 8.4% 11.3% 0% 2.8% 4.2% 18.2% 7.6% 27.3% 6.1% 31.8% 3.0% 0% 21.4% 44.8% 12.2% 5.7% 10.0% 1.7% 2.2% a Does not include the following treatment combinations, which were used in ⱕ2 cases: surgery, radiation and chemotherapy; radiation, chemotherapy, and hormone therapy; chemotherapy and hormone therapy; radiation, chemotherapy, or hormone therapy alone. Two patients received no treatment. b Includes additive hormone treatment with tamoxifen, diethylstilbestrol, or medroxyprogesterone and ablative hormone therapy (orchiectomy). younger (P ⫽ 0.01). Because most hormone therapy use occurred between 1986 and 1996, analysis of the influence of clinical factors on hormone therapy use was based on the subset of patients diagnosed in that decade. Patients were more likely to receive hormone therapy if they had large tumors (P ⫽ 0.02) or if they had a positive family history (P ⫽ 0.01). At last follow-up, 89 patients had relapsed and 30 patients had no appreciable disease free period. The site of first recurrence was local in 13 cases. Six of these cases were lymph node positive at diagnosis and 1 was lymph node negative. Lymph node status was unknown in 6 of these cases. In 7 cases locally recurrent disease was treated by surgical excision either alone (5 of 13 cases) or in combination with radiation (2 of 13 cases). Others were treated by radiation alone (1 of 13 cases), hormone therapy alone (1 of 13 cases), or a combination of radiation, chemotherapy, and hormone therapy (1 of 13 cases). Three cases were not treated for their local recurrence. The site of first recurrence was regional or distant in 76 cases. Treatment for these included hormone therapy in 59 cases, radiation therapy in 51 cases, and chemotherapy in 23 cases. Most cases of regional recurrence that occurred during follow-up involved the axilla (22 of 30 cases, 73.3%). The first site of distant metastases was most often bone (44 of 78 cases, 56.4%) or lung (18 of 78 cases, 23.1%). Median follow-up was 5.3 years, ranging from 3 weeks to 30 years. Survival curves for the entire group are shown in Figure 1. The estimated 5-year DFS, OS, LCR, and cause specific survival (CSS) for all patients are 47%, 53%, 91%, and 63%, respectively. Median survival from first relapse was 4.2 years. No significant difference was observed in outcome for cases grouped according to decade of diagnosis. As shown in Table 5, univariate analysis (log rank test) indicated that DFS was significantly longer for patients who had primary surgery, smaller tumor sizes, negative lymph node status, or positive estrogen receptor (ER) status. Adjuvant chemotherapy was associated with shorter DFS. Younger age, smaller tumor size, negative lymph node status, less advanced disease (⬍ Stage IV) at diagnosis, primary surgical resection, or a positive family cancer history had a positive influence on OS. Again, adjuvant chemotherapy was associated with poorer OS. Factors that were associated with better local control were ER positive tumors, primary surgical resection (vs. no primary surgery), and mastectomy (vs. lumpectomy). Local control was worse in patients treated with adjuvant chemotherapy. Only 6 local relapses occurred within the group of patients known to be free of metastatic disease at diagnosis. Any interpretation of local control data must take this small number of relevant events into account. No significant difference in DFS, OS, or LCR was observed in comparing patients with ⱕ1 or ⬎1 risk factor or patients with a delay from first appearance of symptoms to presentation to a physician of either ⱕ6 months or ⬎6 months. Multiple regression analysis of prognostic factors found that negative lymph node status (P ⫽ 0.0001, Fig. 2) and adjuvant hormone treatment (P ⫽ 0.0368) have a significant positive influence on DFS. Younger age (P ⫽ 0.0003), negative lymph node status (P ⫽ 0.0070, Fig. 2), absence of Stage IV disease at diagnosis (P ⫽ 0.0001), and adjuvant hormone treatment (P ⫽ 0.0401) had a positive influence on OS. Number of risk factors, tumor size, radiotherapy, and chemotherapy had no significant effect on either OS or DFS according to multiple regression analysis. A multivariate analysis of factors affecting local control was not found to be informative because only 6 local relapses 634 CANCER February 1, 1999 / Volume 85 / Number 3 FIGURE 1. A Kaplan–Meier plot of disease free survival, overall survival, and cause specific survival is shown for the group overall (5-year disease free survival: 47%; 5-year overall survival: 53%; 5-year cause specific survival: 63%). occurred within the relevant group of patients (i.e., known to be free of metastatic disease at diagnosis). Data analysis for associations between variables in the group as a whole found only a trend for larger tumors in older patients. Sixteen percent of patients older than 63 years had tumors larger than 5 cm, whereas only 3% of patients age 63 years or younger had such large tumors (chi-square test, P ⫽ 0.014). There was no association between age at diagnosis and family cancer history or number of risk factors, between lymph node status and family cancer history or ER status, or between tumor size and ER status. DISCUSSION The proportion of men with clinical gynecomastia reported in our series is within the range reported in the literature on male breast carcinoma.20 However, the incidence of clinical gynecomastia in healthy adult cohorts has been reported to be as high as 36%.21 In an unselected series of 100 adult male autopsies, microscopic gynecomastia was reported in 55 cases. Clinical gynecomastia was absent in all cases.22 Microscopic examination of histologic specimens from 79 male breast carcinoma cases23 found only 21 with gynecomastia. These numbers suggest that histologic gynecomastia may in fact occur less frequently in men with breast carcinoma. Male breast carcinoma and gynecomastia may be unrelated except that both may be caused by abnormal hormonal environments.24 A longitudinal, prospective histologic or mammographic study of men with gynecomastia may be the only way to resolve this question. The median age at diagnosis in our series of men was similar to that in previous studies,25–27 further supporting the theory that breast carcinoma generally presents almost a decade later in men than in women. Unlike in women, for whom familial breast carcinoma is often associated with an earlier age of onset,28 the median age at diagnosis was similar for patients with and without a family history of cancer. However, underreporting of family cancer history in our series must be considered in any interpretation of these data. The relatively infrequent use of mammography in our patient population, even within the last 2 decades A Review of 229 Male Breast Carcinoma Cases/Goss et al. 635 TABLE 5 Prognostic Factors for Disease Free Survival, Overall Survival, and Local Control Rate: Results of the Log Rank Test and Multiple Regressiona Analysis DFS OS Factor Log rank P value Multiple regression P value (risk ratio) n ⴝ 108 Log rank P value Age (yrs)b ⱖ63, ⬎63 Tumor size (cm) ⬍2, 2–5, ⬎5 Lymph node status Negative, positive ER status Positive, negative Stage IV disease at diagnosis Absent, present Primary surgery Any surgery, no surgery Type of surgery Mastectomy, lumpectomy Adjuvant chemotherapy No chemo, chemo Adjuvant hormone therapy No hormone therapy hormone/orchiectomy Family cancer history Present, absent NS NS 0.0088 0.0349 NS 0.0003 0.0001 0.0011 (2.37) 0.0023 0.0413 Multiple regression P value (risk ratio) n ⴝ 120 0.0003 (1.05) NS 0.0070 (2.03) NS 0.0001 LCR Log rank P value NS NS NS 0.0061 0.0001 (5.98) 0.0057 0.0001 0.0027 NS NS 0.0012 0.0110 NS 0.0493 NS 0.0252 NS 0.0368 (0.36) NS 0.0401 (0.38) NS 0.0483 NS NS DFS: disease free survival; OS: overall survival; LCR: local control rate. a Adjuvant radiotherapy and the presence of ⱖ1 risk factors had no significant influence on outcome by either univariate or multivariate analysis. Blank sections indicate exclusion of the factor from multivariate analysis, as described in the “Methods” section. b Age was considered continuous for the Cox proportional hazards model. of the study period, likely reflects both the large proportion of palpable tumors at presentation and the absence of routine breast carcinoma screening in men. The relative ease of diagnosis by physical exam and the rarity of the disease does not support widespread mammographic screening. However, mammography is useful for differentiating between gynecomastia and malignancy in the male patient. Previous descriptions of male breast carcinoma series have reported a longer median delay between the onset of symptoms and diagnosis, ranging from 6 to 18 months.4,29 Our data indicate a shorter median (4 months), although the delay did range up to 2 years. The significant difference in symptom duration detected between patients with (median, 1 month) or without (median, 4 months) a family cancer history may reflect a greater disease awareness in families affected by cancer or more rapid tumor growth in familial cases. Previously, duration of symptoms has been correlated with decreased survival rates.30 –31 Analysis of our data, however, did not find that the length of delay significantly influenced any of the survival variables. In view of this, it is difficult to explain the positive association (P ⫽ 0.0483) between family cancer history and DFS indicated by our univariate analysis. It is likely that this result is unreliable due to inconsistent reporting of family cancer history in our series. Consistent with previous reports,25,32 the central subareolar region is the most common tumor site in our series. In women the most frequent location is the upper outer quadrant of the breast. In keeping with data from the Surveillance, Epidemiology, and End Results Program (SEER),33 there was no significant excess in laterality (P ⫽ 0.097) observed in our series. The SEER data did report an excess in left-sided breast carcinoma in women. These gender differences in tumor laterality and location are likely related to the relatively rudimentary nature of the male breast. Our low incidence of bilateral male breast carcinomas, as compared with that among females, is consistent with other series. Reports of low disease bilaterality in men have suggested that this is due, in part, to the later age of diagnosis of male patients resulting in death due to other causes prior to contralateral disease onset.32 Based on this point, primary con- 636 CANCER February 1, 1999 / Volume 85 / Number 3 FIGURE 2. Kaplan–Meier plots of male breast carcinoma patients according to lymph node status are shown. (A) Disease free survival at 5 years: lymph node negative, 65%: lymph node positive, 35%. (B) Overall survival at 5 years: lymph node negative, 68%; lymph node positive, 47%. A Review of 229 Male Breast Carcinoma Cases/Goss et al. tralateral mammographic breast surveillance is unlikely to be important in clinical management of the male breast carcinoma patient. Our finding of lobular carcinoma in 2.6% of cases, though higher than reported in previous series (SEER),34 may still be an underestimate of the frequency of this histopathologic subtype. Historically, histologic evaluation of male breast carcinomas has not been more specific than a general diagnosis of adenocarcinoma.35 The earlier assumption that lobular carcinoma could not occur in men due to the absence of lobules in normal male breast tissue has been dispelled. It is now known that true acinii and lobules can develop in the male breast during exposure to increased estrogen.36 It is possible that some cases of lobular carcinoma are also included in the 19.5% of cases with unspecified adenocarcinoma in our series. Despite a greater proportion of cases with large tumors (⬎5 cm) during the first 2 decades (1955–1975) of the study period, the proportion of cases with either metastatic disease or positive lymph nodes at diagnosis was constant for each decade of diagnosis. Furthermore, tumor size was not found to influence OS by multivariate analysis, and survival was similar for each decade of diagnosis. Thus, though it appears that previously undetected, smaller male breast tumors are now being identified, no significant improvement has been made in terms of diagnosing prior to initiation of the metastatic process. The transition in primary surgical procedure in our series from the radical mastectomy to the simple or modified radical mastectomy mirrors the elimination of the radical mastectomy from standard treatment of female breast carcinoma in the 1970s.37 Lumpectomy was not performed often, even within the most recent decade of our series. As is indicated by the negative association between local control and the use of lumpectomy in our series, the paucity of breast tissue in men may often render adequate tumor excision by lumpectomy impossible. It is not likely that lumpectomy with or without axillary dissection will be performed as often on men as on women. Treatment data from our group of male breast carcinoma patients suggests that, despite an increase in use during the last 2 decades, hormone therapy and chemotherapy are used less often in the adjuvant treatment of male breast carcinoma than female breast carcinoma. Furthermore, the duration of adjuvant hormone therapy in our series was often less than 2 years. Studies of women with ER positive disease have conclusively shown that adjuvant treatment with tamoxifen improves overall survival,38 and that 5 years of adjuvant tamoxifen is more beneficial than 2 637 years.39 It is possible then that more frequent use of adjuvant tamoxifen prescribed for 5 years may improve the survival of ER positive male breast carcinoma patients. In addition, recent studies by the International Breast Cancer Study Group,40 the National Surgical Adjuvant Breast and Bowel Project,41 and the Southwestern Oncology Group42 of postmenopausal women with breast carcinoma have shown that the addition of chemotherapy to adjuvant tamoxifen may be of benefit for both lymph node negative and lymph node positive patients with either ER positive or ER negative disease. These findings suggest that more aggressive treatment, which includes adjuvant chemotherapy, may be beneficial to men. Chemotherapy use may, however, be limited by the older average age of male breast carcinoma patients. It is important to note that the negative association between survival variables and adjuvant chemotherapy use found in our univariate analysis was not confirmed by multivariate analysis. This likely reflects administration of adjuvant chemotherapy to patients with more advanced disease, as is indicated by our finding that chemotherapy was more likely to be given to patients with positive lymph nodes. Indeed, when disease stage and other factors were included in the analysis, this negative effect was no longer found. Reported overall 5-year survival rates for male breast carcinoma vary widely from 40% to 65%.27,43,44 The 53% OS at 5 years found in our series falls in the middle of this range. It is important to note that the Kaplan–Meier method for analyzing survival in our series gives a biased, often overestimated CSS.45 Therefore, though the CSS has been estimated for our series (shown in Fig. 1), the DFS and OS are more informative. Our finding that men diagnosed with breast carcinoma at a younger age have a survival advantage over older patients has been observed previously in a large, population-based study of 1429 patients from Nordic countries.46 The larger sample size of the Nordic study enabled survival estimates to be more accurately controlled for the fact that men are diagnosed with breast carcinoma at an older average age than women therefore are more likely to die of causes other than breast carcinoma. In female breast carcinoma patients, age-associated mortality risk does not appear to be linear. Women diagnosed when younger than 40 years or older than 50 years have been shown to have a worse prognosis than those age 40 – 49 years.47– 49 Increased survival in this age group has been attributed to the increased likelihood of natural menopause occurring during this time period, thereby depriving newly diagnosed breast tumors of hormonal stimulation.49 The linear decrease in survival observed with 638 CANCER February 1, 1999 / Volume 85 / Number 3 increasing age among male breast carcinoma patients may reflect the lack of significant age-dependent endocrine changes. Lymph node involvement has often been cited as a significant prognostic indicator in men.5,26,27,32,50 Reported 5-year survival rates for lymph node negative and lymph node positive patients vary. The 68% overall survival estimate from our lymph node negative group of patients is lower than previous reports, ranging from 77% to 100%.14,27,32 This difference is partially attributable to the exclusion of patients with pathologies associated with a better prognosis (i.e., pure DCIS) from our survival analysis. However, the 47% overall survival of lymph node positive patients is in keeping with previous reports, ranging from 37% to 61%.14,27,32 Comparison of the overall prognoses for male and female breast carcinoma patients is controversial. Studies reporting a worse prognosis for males have suggested that the anatomy of the male breast may provide less of a barrier to metastases or that more aggressive tumor biology may be the basis of survival variation.26,51,52 However, others have found that, once separated according to stage or lymph node involvement, the prognosis is the same.5,25,27 Our 53% OS at 5 years for the entire group is lower than the 64% OS found for female breast carcinoma patients reported by the National Surgical Adjuvant Breast and Bowel Project in the 1970s.53 When separated according to negative or positive lymph node status, the 5-year OS for the same group of women is 78% and 47%, respectively. Though OS for the lymph node negative male patients in this study is lower, the OS for the lymph node positive male patients in our study is the same. However, the lack of controlled, large male breast carcinoma cohorts make a definitive prognostic comparison between genders difficult. Some modifications in clinical management were observed over the 4 decades of the study period. Smaller tumors were detected and axillary lymph nodes were assessed more often by the end of the study period. The standard primary surgical technique evolved from the radical mastectomy to the simple or modified mastectomy, and chemotherapy and hormone therapy were used slightly more often in the last 2 decades. However, these changes in management had no detected impact on the outcomes of the patients in our series. As suggested previously, increased use of adjuvant tamoxifen and adjuvant chemotherapy similar to that indicated in emerging guidelines for female breast carcinoma may improve outcome for this group with long term poor prognosis. Clearly, these statements regarding treatment may only be considered as suggestions at this point. The data presented in this retrospective study is not appropriate for accurately assessing the efficacy of any given treatment; however, prospective clinical trials of male breast carcinoma have yet to be performed. Though the rarity of the disease limits the feasibility of clinical trials, comprehensive, prospective data collection alone would augment our understanding of risk factors and prognostic factors as well as improve clinical management specific for the male breast carcinoma patient. Though some improvement was seen over the study period, the insufficient reporting of risk factors, such as family history and pathology data (including histologic grade, nuclear grade, and hormone receptor and axillary lymph node status), in our series further emphasizes the need for rigorous, thoughtful data collection. Thus further research is required to determine the molecular biologic properties of male breast carcinoma, particularly in relation to BRCA154 and BRCA2,55,56 to characterize the role of adjuvant chemotherapy; and to determine the optimal treatment of metastatic disease. REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. Sasco AJ, Lowenfels AB, Pasker-De Jong P. Epidemiology of male breast cancer: a meta-analysis of published case– control studies and discussion of selected aetiological factors [review article]. Int J Cancer 1993;53:538 – 49. Boring CC, Squires TS, Tong T, Montegomery S. Cancer statistics, 1994. CA Cancer J Clin 1994;44:7–26. Izquierdo MA, Alonso C, DeAndres L, Ojeda B. Male breast cancer: report of a series of 50 cases. Acta Oncol 1994;33: 767–71. Crichlow RW. Carcinoma of the male breast. Surg Gynecol Obstet 1972;134:1011–9. Guinee VF, Olsson H, Moller T, Shallenberger RC, van den Blink JW, Peter Z, et al. The prognosis of breast cancer in males. Cancer 1993;71:154 – 61. Hecht RJ, Winchester DJ. Male breast cancer. Am J Clin Pathol 1994;102(S1):S25–S30. Michaels BM, Nunn CR, Roses DF. Lobular carcinoma of the male breast. Surgery 1994;115:402–5. Everson RB, Li FP, Fraumeni JF, Wilson RE, Fishman J, Stout D, et al. Familial male breast cancer. Lancet 1976;I:9. Kozak FK, Hall JG, Baird PA. Familial breast cancer in males: a case report and review of the literature. Cancer 1986;58: 2736 –9. Anderson DE, Badzioch MD. Breast cancer risks in relatives of male breast cancer patients. J Natl Cancer Inst 1992;84: 1111–7. Evans DB, Crichlow RW. Carcinoma of the male breast and Klinefelter’s syndrome. CA Cancer J Clin 1987;37:246 –51. Kelsey JL, Gammon MD. Epidemiology of breast cancer. Epidemiol Rev 1990;12:228 – 40. Thomas DB. Male breast cancer. Epidemiol Rev 1993;15: 220 –31. Erlichman C, Murphy KC, Elhakim T. Male breast cancer: a 13-year review of 89 patients. J Clin Oncol 1984;2:903–9. McLachlan SA, Erlichman C, Liu F-F, Miller N, Pintilie M. Male breast cancer: an 11-year review of 66 patients. Breast Cancer Res Treat 1996;40:225–30. A Review of 229 Male Breast Carcinoma Cases/Goss et al. 16. Hosmer DW, Lameshow S. Applied logistical regression. New York: John Wiley & Sons, 1989. 17. Kaplan EL, Meier P. Nonparametric estimation from incomplete observations. J Am Stat Assoc 1958;53:457– 81. 18. Mantel N. Evaluation of survival data and two new rank order statistics arising in its consideration. Cancer Chemother Rep 1966;50:163–70. 19. Cox DR. Regression models and lifetables. J R Stat Soc Series B 1972;34:187–202. 20. Korenman SG. The endocrinology of the abnormal male breast. Ann N Y Acad Sci 1986;464:400 – 8. 21. Braustein GD. Gynecomastia. N Engl J Med 1993;328:490 –5. 22. Anderson JA, Gram JB. Male breast at autopsy. Acta Pathol Microbiol Immunol Scand (A) 1982;90:191–7. 23. Scheike O, Visfeldt J. Male breast cancer 4: gynecomastia in patients with male breast cancer. Acta Pathol Microbiol Scand (A) 1973;81:359 – 65. 24. Donegan WL. Cancer of the breast in men. CA Cancer J Clin 1991;41:339 –54. 25. Stierer M, Rosen H, Weitensfelder W, Hausmaninger H, Teleky B, Jakesz R, et al. Male breast cancer: Austrian experience. World J Surg 1995;19:687–93. 26. Salvadori B, Saccozzi R, Manzari A, Andreola S, Conti RA, Cusumano F, et al. Prognosis of breast cancer in males: an analysis of 170 cases. Eur J Cancer 1994;30A:930 –5. 27. Cutuli B, Lacroze M, Dilhuydy JM, Velten M, De Lafontan B, Marchal C, et al. Male breast cancer: results of the treatments and prognostic factors in 397 cases. Eur J Cancer 1995;31A:1960 – 4. 28. Lynch HT. Breast cancer genetics: family history, heterogeneity, molecular genetic diagnosis, and genetic counseling. Curr Probl Cancer 1996;Nov/Dec:331– 65. 29. Scheike O. Male breast cancer 5: clinical manifestations in 257 cases in Denmark. Br J Cancer 1973;28:552– 62. 30. Appleqvist P, Salmo M. Prognosis in carcinoma of the male breast. Acta Chir Scand 1982;148:499 –502. 31. Roswit B, Edlis H. Carcinoma of the male breast: a thirtyyear experience and literature review. Int J Radiat Oncol Biol Phys 1978;4:711– 6. 32. Borgen PI, Wong GY, Vlamis V, Potter C, Hoffmann B, Kinne DW, et al. Current management of male breast cancer: a review of 104 cases. Ann Surg 1992;215:451–9. 33. Weiss HA, Devesa SS, Brinton LA. Laterality of breast cancer in the United States. Cancer Causes Control 1996;7:539 – 43. 34. Stalsberg H, Thomas DB, Rosenblatt KA, Jimenez LM, McTiernan A, Stemhagen A, et al. Histologic types and hormone receptors in breast cancer in men: a population-based study in 282 United States men. Cancer Causes Control 1993;4:143–51. 35. Michaels BM, Nunn CR, Roses DF. Lobular carcinoma of the male breast. Surgery 1994;115:402–5. 36. Schwartz IS. Small-cell breast carcinoma in men [letter]. Hum Pathol 1982;13:185. 37. Wilson RE, Donegan WL, Mettlin C, Smart CR, Murphy GP. The 1982 national survey of carcinoma of the breast in the United States by the American College of Surgeons. Surg Gynecol Obstet 1982;159:309 –18. 38. Early Breast Cancer Trialists’ Cooperative Group (EBCTCG). Treatment of early breast cancer: worldwide evidence in 1985–1990. A systematic overview of all available randomized trials in early breast cancer of adjuvant endocrine and cytotoxic therapy. Treatment of early breast cancer. Volume 1. Oxford, UK: Oxford University Press, 1990. 639 39. Swedish Breast Cancer Cooperative Group. Randomized trial of two versus five years of adjuvant tamoxifen for postmenopausal early stage breast cancer. J Natl Cancer Inst 1996;88:1543–9. 40. International Breast Study Group. Effectiveness of adjuvant chemotherapy in combination with tamoxifen for nodepositive postmenopausal breast cancer patients. J Clin Oncol 1997;15:1385–94. 41. Fisher B, Dignam J, Wolmark N, DeCillis A, Emir B, Wickerham DL, et al. Tamoxifen and chemotherapy for lymph node-negative, estrogen receptor positive breast cancer. J Natl Cancer Inst 1997;89:1673– 82. 42. Albain K, Green S, Osborne K, Cobau C, Levine E, Ingle J, et al. Tamoxifen (T) vs. cyclophosphamide, adriamycin, and 5-FU (CAF) plus either concurrent or sequential T in postmenopausal, receptor (⫹), node (⫹) breast cancer: a Southwest Oncology Group Phase III Intergroup Trial (SWOG 8814) [Abstract 450]. Proc Am Soc Clin Oncol 1997;16:128a. 43. Wagner JL, Thomas CR, Wui-jin K, Rudolph RH. Carcinoma of the male breast: update 1994. Med Pediatr Oncol 1995;24: 123–32. 44. Ribiero G. Male breast carcinoma: a review of 301 cases from the Christie Hospital and Holt Radium Institute, Manchester. Br J Cancer 1985;51:115–9. 45. Caplan RJ, Pajak TF, Cox JD. Analysis of probability and risk of cause-specific failure. Int J Radiat Oncol Biol Phys 1994; 29:1183– 6. 46. Adami HO, Hakulinen T, Ewertz M, Tretli S, Holmberg L, Karjalainen S. The survival pattern in male breast cancer: an analysis of 1429 patients from the Nordic countries. Cancer 1989;64:1177– 82. 47. Adami HO, Malker B, Holmberg L, Persson I, Stone B. The relation between survival and age at diagnosis in breast cancer. N Engl J Med 1986;315:559 – 63. 48. Palmer MK, Lythgoe JP, Smith A. Prognostic factors in breast cancer. Br J Surg 1982;69:697– 8. 49. Mohle-Boetani J, Grosser S, Malec M, Whittemore AS. Survival advantage among patients with breast cancer diagnosed at 45 to 49 years of age. N Engl J Med 1986;315:587. 50. Crichlow RW. Breast cancer in men. Semin Oncol 1974;1: 145–52. 51. Joshi MG, Lee AK, Loda M, Camus MG, Pedersen C, Heatley GJ, et al. Male breast carcinoma: an evaluation of prognostic factors contributing to a poorer outcome. Cancer 1996;77: 490 – 8. 52. Ciatto S, Iossa A, Bonardi R. Male breast carcinoma: review of a multicenter series of 150 cases. Tumori 1990;76:555– 8. 53. Fisher B, Slack N, Katrych D, Wolmark N. Ten year follow-up results of patients with carcinoma of the breast in a cooperative clinical trial evaluating surgical adjuvant chemotherapy. Surg Gynecol Obstet 1975;140:528 –34. 54. Struewing JP, Brody LC, Erdos MR, Kase RG, Giamberresit R, Smith SA, et al. Detection of eight BRCA1 mutations in 10 breast/ovarian cancer families, including 1 family with male breast cancer. Am J Hum Genet 1995;57:1–7. 55. Tavtigian SV, Simard J, Rommens J, Shattuck-Eidens D, Neuhausen S, Merajver S, et al. The complete BRCA2 gene and mutations in chromosome 13q-linked kindreds. Nat Genet 1996;12:333–7. 56. Couch FJ, Farid L, DeShano ML, Tavtigian SV, Calzone K, Campeau L, et al. BRCA2 germ line mutations in male breast cancer cases and breast cancer families. Nat Genet 1996;13: 123–5.