The incidence of male breast cancer (MBC) continues to rise. The Veterans Affairs (VA) Central Cancer Registry (VACCR) provides a unique source for the study of MBC. The objective of this retrospective analysis was to compare the characteristics and outcome of patients with MBC and patients with female breast cancer (FBC) in the VA population.
VACCR data were used to analyze the database of VA patients who had breast cancer diagnosed between 1995 and 2005. It includes 120 VA medical centers. Primary site codes were identified for breast cancer (500–508). Data were entered and analyzed using biostatistical software.
In total, 3025 patients' records were reviewed, and 612 patients who had MBC were compared with 2413 patients who had FBC. The mean age at diagnosis was 67 years for patients with MBC and 57 years for patients with FBC (P < .005). More patients with MBC were black, and patients with MBC presented with higher disease stage and more lymph node-positive disease. The dominant histology in MBC was ductal carcinoma. No difference in grade or laterality was observed. Estrogen and progesterone receptor-positive tumors were more common in MBC compared with FBC. Overall, patients with MBC received less chemotherapy, whereas no statistical difference was observed in the use of hormone treatment. The median overall survival for patients who had MBC was 7 years compared with 9.8 years for patients who had FBC (log-rank test; P < .005). There was no statistically significant difference in median survival for patients with stage III disease and stage IV disease. However, the median survival differed significantly for patients with stage I disease and stage II disease. In lymph node-negative patients, the median survival was 6.1 years for patients with MBC and 14.6 years for patients with FBC (P < .005), whereas the median survival did not differ significantly in lymph node-positive patients. Using Cox regression analysis age, sex, clinical stage, and lymph node status were independent prognostic factors for survival, whereas race, histology, and grade were not.
Male breast cancer (MBC) is uncommon and accounts for approximately 1% of all breast cancers diagnosed in the United States each year.1–10 The incidence of MBC, however, is rising and continues to increase by 1.1% yearly.10 The American Cancer Society estimates that, in 2006, 1760 men will be diagnosed with breast cancer, and 460 men will die of the disease. The National Cancer Institute (NCI) Surveillance, Epidemiology, and End Results (SEER) Program noted that the incidence of MBC increased by 26% from 1973 to 1998.7 In the SEER data, the median age at diagnosis of breast cancer is 68 years for men compared with 63 years for women.2 The disease has been reported in men at ages ranging from 5 years to 93 years.3
Greater than 40% of patients with MBC present with stage III or IV disease.11, 12 Unlike the declining mortality of female breast cancer (FBC), death rates from MBC have not decreased since 1975.9 The rarity of MBC has precluded the development of prospective, randomized clinical trials. The management of MBC, therefore, has been based on treatment for FBC.
Significant differences in the biology and clinicopathologic characteristics have been described in MBC and FBC.6, 7, 9 A better understanding of MBC is desirable to influence this disease of rising incidence. For the current study, we evaluated patients with MBC and patients with FBC who were diagnosed and/or treated in the Veterans Affairs (VA) medical centers throughout the United States, comparing their baseline characteristics, outcomes, and prognostic factors.
MATERIALS AND METHODS
Data for this study were obtained from the VA Central Cancer Registry (VACCR) between 1995 and 2005. The VACCR is a function of the Chief, Program Office for Oncology at VA Headquarters in Washington, DC. The reference date for data collection and reporting was January 1, 1995. Approximately 120 VA medical centers diagnose and/or treat patients with cancer. The VACCR aggregates the data collected by the medical centers' cancer registries. The primary site codes for breast cancer in this database were 500 through 508.
In total, 3025 veterans throughout the 23 regional Veterans Integrated Service Networks, which form the VA health system, were included (612 male veterans and 2413 female veterans). The variables studied were patient age at diagnosis, sex, race, clinical stage, histology, grade, lymph node status, and hormone status of the tumor. Tumor histology was classified by using the International Classification of Diseases coding system: carcinoma not otherwise specified (8010), papillary carcinoma (8050), ductal carcinoma (8500), lobular carcinoma (8520), mucinous carcinoma (8480), medullary carcinoma (8510), inflammatory carcinoma (8530), ductal and lobular carcinoma (8522), phyllodes carcinoma (90203), lymphoma (9680), sarcoma (8850, 9120, and 8980), and carcinoid (82403).8
Data were entered and analyzed using SPSS biostatistical software (version 10.0.05). Patient and disease characteristics among men and women with breast cancer were compared using the chi-square test. Overall survival (OS) was estimated with the Kaplan-Meier product-limit method. The 2-sided log-rank test was used to compare survival between men and women with breast cancer. Multivariate analysis was performed using the Cox proportional hazards regression model to determine the independent association between each variable and survival. Variables that were considered in the multivariate analysis included patient age at diagnosis, sex, disease stage, lymph node status, tumor grade, and histology. All P values presented are 2-sided, and P values <.05 were considered statistically significant.
Records of 612 men and 2413 women with breast cancer were analyzed. Patient characteristics of both populations are presented in Table 1. The mean age at diagnosis for patients with MBC was 67 years compared with 57 years for patients with FBC (P < .005). More patients with MBC were black (23% vs 16%; P = .013). Alcohol (35% vs 22%; P < .005) and smoking (47% vs 38%; P < .005) were more common in patients with MBC.
Table 1. Patient Characteristics
No. of patients (%)
Male, n = 612
Female, n = 2413
Mean age at diagnosis, y
Current alcohol user
History of alcohol
Alcohol history unknown
No tobacco history
Current tobacco user
History of tobacco
Family history of cancer
No family history
Positive family history
Disease characteristics for both populations are presented in Table 2. More patients with FBC were diagnosed at an early stage (stage I or II; 47% vs 29%; P < .0001), whereas more patients with MBC presented with disease at higher stages (stage III or IV; 62% vs 42%; P < .0001). More men than women had tumors that measured >2 cm (49% vs 33%; P < .0001). Regarding lymph node status and grade, more patients with MBC had lymph node-positive tumors at presentation (42% vs 28%; P < .0001), and men also had higher grade tumors than women (57% vs 49%; P < .0001). Laterality (distribution of breast cancer between the right and left breasts) did not differ significantly between men and women. In men with breast cancer, ductal histology was dominant and was observed more commonly than in women with breast cancer (74% vs 67%). Lobular carcinoma was infrequent in patients with MBC. More men had estrogen receptor (ER)-positive or unknown tumors (96.7% vs 89%; P < .005) and more progesterone receptor (PR)-positive tumors compared with women (53% vs 47%; P < .004). Therapeutic approaches to both men and women with breast cancer were assessed; more patients with FBC received chemotherapy (36% vs 29%; P < .013) and radiotherapy (34% vs 20%; P < .001). But endocrine hormone therapy was provided equally for both MBC and FBC.
Table 2. Disease Characteristics
No. of patients (%)
Male, n = 612
Female, n = 2413
NOS indicates not otherwise specified.
Tumor size, cm
Any size with chest wall invasion
Lymph node status
Ductal and lobular
Of the 612 men and 2413 women with breast cancer in this study, there were 120 deaths reported among men and 273 deaths reported among women. Four hundred ninety-two men (80%) and 2140 women (88%) remained alive. The median OS was lower in men (7 years vs 9.6 years; P < .0001) (Fig. 1, Table 3). When the patients were stratified by disease stage, the median OS was lower in men with stage I disease (7 years vs not reached; P = .005) and stage II disease (6 years vs 8.6 years; P = .001) (Fig. 2A,B). However, the median OS did not differ significantly between men and women with stage III and IV disease (Fig. 2C,D). With regard to lymph node status, OS was inferior in men compared with women who had lymph node-negative tumors (6.1 years vs 14.6 years; P < .005) but did not differ statistically for men and women with lymph node-positive tumors (Fig. 3A,B). In patients with ER-positive tumors, the OS was inferior in MBC (7.3 years vs 9.8 years; P < .005) (Fig. 4). For men with ER-negative tumors, the median OS could not be analyzed adequately because of the small number with documented ER-negative tumors (22 men), and 222 patients had unknown hormone receptor status.
Table 3. Comparison of Overall Survival for Male and Female Breast Cancer
OS indicates overall survival.
Tumor size, cm
Any size with chest wall invasion
Lymph node status
Causes of death were evaluated and available information was recorded for 110 men and 259 women with breast cancer. Regarding disease-specific mortality, 21 men and 91 women reportedly died from breast cancer. Other causes of death included lung cancer (8 men and 6 women), gastrointestinal cancers (4 men and 4 women), other cancers (men: 1 with bladder cancer, 2 with hepatic cancer, 1 with leukemia, and 2 with skin cancer; women: 1 with ovarian cancer, 2 with hepatic cancer, 1 with lymphoma, and 1 with bone cancer), severe infections (7 men and 6 women), cardiac (63 men and 140 women), respiratory failure (1 man and 2 women), renal failure (1 man and 2 women), gastrointestinal bleeding or obstruction (1 man and 1 woman), and stroke (0 men and 2 women). The causes of death for the remaining patients were unknown. Finally, Cox regression analyses showed that sex, age at diagnosis, stage, and lymph node status were independent predictors of OS, whereas tumor grade, race, and histology were not.
The results from this study suggest that there are differences in the characteristics of patients with breast cancer between men and women in the VA population. Patients who had early-stage MBC (stage I and II) had worse outcomes than patients who had FBC with the same stage. Specifically, the OS of patients with lymph node-negative MBC was significantly inferior to that of patients with lymph node-negative FBC. OS was not statistically different, however, between lymph node-positive MBC and lymph node-positive FBC. This significant difference at early stages may be attributable to the intrinsic differences in the biology of the tumor.
In addition, patients who had ER-positive MBC had inferior survival compared with patients who had FBC, although there were no differences in the use of endocrine hormone therapy between MBC and FBC. This may imply that, in MBC, ER-positive tumors do not respond as well to standard hormone therapy with tamoxifen (standard therapy at the VA during the study period). This is consistent with reports suggesting that the expression of ER and PR in MBC does not appear to translate necessarily into a better prognosis compared with ER-negative/PR-negative tumors5, 13, 14; and the benefit from tamoxifen also was uncertain.13, 14 Proliferative activity reportedly was higher in ER-positive/PR-positive MBC tumors than in ER-negative/PR-negative MBC tumors.15 It also has been suggested that ER status in MBC does not have the same function as ER status in FBC.16 It is possible that, based on the estrogen-dependent pathway model, there is up-regulation of ER in a low estrogen environment like the male system.6 This may translate downstream into an increased response of estrogen targets, such as Bcl-2 expression, which inhibits apoptosis and has a higher incidence of expression in MBC.6 Other approaches to hormone manipulation in breast cancer include the use of aromatase inhibitors. Third-generation aromatase inhibitors have not been studied adequately in MBC. The first-generation aromatase inhibitor aminoglutethimide, however, demonstrated poor results when it was used in patients with advanced MBC. However, it was effective in the treatment of men who previously had undergone orchiectomy.9 Endocrine manipulation for patients with MBC probably should be distinguished from that for the treatment of patients with FBC. If the male tumors, in fact, are dependent hormonally, then alternative pathways and hormones have yet to be explored.7, 9 There were only a few patients with MBC in the VACCR database who had ER-negative tumors, so the current analyses did not allow us to make conclusions regarding the survival of patients with ER-negative MBC. This study did not assess HER-2/neu status because of the lack of collected data in the VA database; however, it is likely that HER-2/neu status confers the same prognostic implication for MBC and FBC. A small study that evaluated the survival of patients with MBC in the VA population showed that, stage for stage, HER-2/neu overexpression was correlated with a shorter disease-free survival.14
In the current study, we also detected racial differences between MBC and FBC. The incidence rate of breast cancer appears higher in black men than in black women; whereas we observed the opposite in white patients, who had a higher incidence in women. Again, this is consistent with the racial differences described in the literature between MBC and FBC.17
In patients with MBC, the Cox regression analysis indicated that age at diagnosis, stage, and lymph node status were predictors of OS, similar to FBC (Table 4). Grade, race, and histology, however, were not predictors in patients with MBC. This is also differed from FBC, in which grade and race are known as prognostic parameters. The Cox regression analysis indicated that sex is an independent predictor of survival.
Table 4. Cox Regression Analysis
Our results indicated that patients who have MBC tend to be older at the time of diagnosis compared with patients who have FBC, as previously reported.1, 18 Patients with MBC also present with higher stage disease, larger tumors, and more lymph node-positive disease. These findings reflect that women are more likely to be diagnosed with small, early-stage tumors by screening mammography.1 In addition, general public awareness about MBC essentially is minimal, leading to possible delays in diagnosis.1 In our study more ER-positive tumors were observed in MBC compared with FBC (60% vs 52%), whereas more PR-positive tumors were observed in MBC (53% vs 47%). The percentages of ER-positive and PR-positive MBC were higher in the SEER data,1 in the range of 90%. Similarly, the percentages of ER-positive and PR-positive FBC were higher in the SEER data, in the range of 76%. This probably is because 37% of MBC patients were registered with unknown ER status and, 38% were registered with unknown PR status in the VA database. Similarly, 37% of FBC patients had unknown ER/PR status.
Regarding lymph node status, results from both our study and the SEER database indicate that patients with lymph node-positive MBC have worse survival compared to their female counterparts. Unlike the NCI SEER database, our study showed that patients who had stage I and stage II MBC had a significantly worse OS compared with patients who had FBC in the same stage. Furthermore, the current results indicated that patients who had lymph node-negative MBC had inferior survival compared with their female counterparts (6 years vs 14 years).
The SEER database indicated that survival was equivalent among patients who had early-stage MBC and early-stage FBC. In MBC patients with stage III and IV disease, our study indicated that OS was similar between the 2 groups and generally was low (stage III: 2.56 years for MBC vs 5.51 years for FBC; stage IV: 1.47 years for MBC vs 2.22 years for FBC). This is consistent with the National Cancer Database18 and the SEER data.
To our knowledge, this is the first study to evaluate MBC in the veteran population with long-term survival data. It allows comparison with FBC in a relatively homogenous population of patients within the VA system. Limitations include the retrospective design of the study; the lack of detailed information regarding HER-2/neu overexpression and BRCA1 and BRCA2 mutations; the inclusion of many patients with unknown hormone status; and an absence of detailed treatment records. In addition, coding for cause of death is nonspecific, because most deaths in the MBC group reportedly were secondary to infection (sepsis, pneumonia). It is unclear whether this was related to advanced-stage disease. We cannot draw firm conclusions regarding the disease- specific mortality and survival of patients who had MBC compared with patients who had FBC. Although the worse OS in the MBC group could be attributed to older age and increased mortality in men in the general population, the current study highlights important differences between MBC and FBC. A better understanding of this disease is needed so that new opportunities for therapeutic intervention may be developed.
We acknowledge Dr. Stanley Troup for his valuable editorial comments