Fax: (713) 794-4385
Breast carcinoma in men
A population-based study
Article first published online: 24 MAY 2004
Copyright © 2004 American Cancer Society
Volume 101, Issue 1, pages 51–57, 1 July 2004
How to Cite
Giordano, S. H., Cohen, D. S., Buzdar, A. U., Perkins, G. and Hortobagyi, G. N. (2004), Breast carcinoma in men. Cancer, 101: 51–57. doi: 10.1002/cncr.20312
Fax: (713) 794-4385
- Issue published online: 18 JUN 2004
- Article first published online: 24 MAY 2004
- Manuscript Accepted: 29 MAR 2004
- Manuscript Received: 16 MAR 2004
- National Cancer Institute. Grant Number: 2P30 CA016672 28 (PP-4)
- breast carcinoma;
- prognostic factors;
Male breast carcinoma is an uncommon disease, and most previous studies have been single-institution series that were limited by extremely small sample sizes. The goals of the current study were to fill in the major gaps in knowledge regarding the incidence, presenting characteristics, prognostic factors, and survival rates of male breast carcinoma and to determine how breast carcinoma differs between men and women.
Data from the National Cancer Institute Surveillance, Epidemiology, and End Results 1973–1998 database were used. Age-adjusted incidence rates were calculated. Characteristics of the patients and presenting tumors were compared between men and women. Univariate and multivariate analyses were performed to determine the effect of each variable on overall survival. Survival rates by disease stage were compared for men and women.
Over the years of the study, the incidence of male breast carcinoma increased significantly from 0.86 to 1.08 per 100,000 population (P < 0.001). Men had a higher median age at diagnosis (P < 0.001) and were more likely to have lymph node involvement (P < 0.001), a more advanced stage at diagnosis (P < 0.001), and tumors that were positive for estrogen receptor (ER) (P < 0.001) and progesterone receptor (PR) (P < 0.001). In multivariate analysis, larger tumor size and lymph node involvement were associated with shortened survival. Tumor grade and ER/PR status did not appear to independently influence survival. Relative survival rates by stage of disease for men and women were similar.
Although it remains a rare disease, the incidence of male breast carcinoma is increasing. Breast carcinoma in men has some epidemiologic and biologic differences from breast carcinoma in women. Cancer 2004. © 2004 American Cancer Society.
Breast carcinoma in men is an uncommon disease. In 2003, an estimated 1300 new cases were diagnosed.1 Male breast carcinoma represents 0.6% of all breast carcinomas and < 1% of all malignancies in men. Because of the rarity of breast carcinoma in men, limited information is available regarding the epidemiology, treatment, and prognosis of this disease. To our knowledge, no randomized trials have been performed to date and most published series have been retrospective chart reviews covering many decades and including at most several hundred patients.2–9 Only 2 studies have been published that have contained > 500 male patients with breast carcinoma. One was focused on patterns-of-care and the other, which to our knowledge is the only large population-based series published to date, described variations in survival rates by patient ethnicity.10, 11 The current population-based study, which includes 2537 men with breast carcinoma, was undertaken to fill in the major gaps in knowledge that exist regarding the incidence, presenting characteristics, survival rates, and prognostic factors for male breast carcinoma and to determine how breast carcinoma differs between men and women.
MATERIALS AND METHODS
Data for the current study were obtained from the National Cancer Institute's Surveillance, Epidemiology, and End Results (SEER) program using the nine-registry 1973–1998 data [August 2000 submission]. The SEER database is a population-based cancer registry that is the authoritative source of information concerning cancer incidence and survival in the U.S. The areas included in the registry since 1973 are the states of Connecticut, Hawaii, Iowa, New Mexico, and Utah, as well as the metropolitan areas of San Francisco-Oakland and Detroit. The registries in Seattle and Atlanta were added in 1974. Over the years of the current study, the SEER database included approximately 14% of the U.S. population. The SEER registry routinely collects information regarding patient demographics, tumor characteristics, stage at diagnosis, date of diagnosis, treatment within 4 months of diagnosis, and date and cause of death.
In the current study, all cases of invasive and in situ breast carcinoma diagnosed between 1973–1998 were included. Analyses were based on 2537 cases of male breast carcinoma. All cases of female breast carcinoma (n = 383,146) diagnosed over the same time period were used for comparison. Study variables included patient age at diagnosis, laterality, grade (low, intermediate, high), and tumor histology. Using the International Classification of Disease coding system, tumor histology was classified as ductal (8500–8508), papillary (8050, 8260), mucinous (8480, 8481), medullary (8510–8514), and lobular (8520–8522).12 For cases diagnosed after 1988, American Joint Committee on Cancer (AJCC) stage (AJCC cancer staging manual, 3rd edition), tumor size, and lymph node status were available. In 1990, SEER began to collect data regarding estrogen receptor (ER) and progesterone receptor (PR) status. Within each category, patients with unknown stage, size, lymph node status, grade, laterality, histology, and ER/PR status were excluded from the comparative analysis. “Borderline” ER/PR status were categorized with the unknown cases.
The age-adjusted incidence rates were calculated for breast carcinoma in men and were plotted by age and year at diagnosis. A simple linear regression model was used to determine trends in incidence over time. Patient and disease characteristics at the time of diagnosis were compared between male and female patients using the chi-square test. The association between ER and patient age was tested by the Cochran–Armitage trend test. Overall and relative survival rates were calculated for men and women with breast carcinoma, and stratified by stage of disease. Relative survival rates were calculated using SEER*STAT software. The U.S. 1970, 1980, and 1990 (white, black, other) tables for expected survival rates (which match cases by race, gender, age, and date of coding) were used. Specifically, this expected rate table matches to the cohort cases using the SEER Race Recode A variable and applies the 1970 rates to 1973–1975 diagnosis years, the 1980 rates to the 1976–1985 years, and the 1990 rates to the diagnosis years recorded between 1986–1998. Overall survival was estimated using the Kaplan–Meier product-limit method. Survival times were missing for 13 men and 2290 women. Therefore, 2524 men and 380,856 women were included in the survival analyses. Patient and tumor characteristics were studied to determine the effect of each variable on overall survival. The two-sided log-rank test was used to test the association between patient variables and survival. Multivariate analysis was performed using Cox proportional hazards regression model to determine the independent association between each variable and survival. Variables considered in the multivariate analysis included patient age at diagnosis (dichotomized to age < 65 years and age ≥ 65 years), tumor size, lymph node status, ER/PR status, and tumor grade. Stage was omitted because of its close correlation with size and lymph node status. Histology was not included because there were too few events for the analysis. All P values presented are two sided and P values < 0.05 were statistically significant. Statistical analyses were performed using SAS software (Version 8.02; SAS Institute, Cary, NC), STATA software (Version 7.0; Stata Corporation, College Station, TX), and SEER*STAT (Version 5.0; National Cancer Institute, Bethesda, MD).
Of the 2537 men and 383,146 women included in the current study, the median follow-up was 54 months for the men (range, 0–309 months) and 62 months for the women (range, 0–311 months). There were 1376 deaths reported among the men and 115,290 reported among the women. The common descriptive characteristics of both populations are presented in Table 1.
|Characteristics||Males (2537) No. (%)||Females (383, 146) No. (%)|
|Age at diagnosis (yrs)|
|Median (range)||67 (10–103)||62 (11–108)|
|< 34||22 (0.9)||9504 (2.5)|
|35–44||139 (5.5)||42,768 (11.2)|
|45–54||325 (12.8)||76,943 (20.1)|
|55–64||626 (24.7)||84,986 (22.2)|
|65–74||755 (29.8)||89,685 (23.4)|
|≥ 75||671 (26.4)||79,265 (20.7)|
|White||2171 (85.5)||334,594 (87.3)|
|Black||256 (10.1)||28,215 (7.4)|
|Other||92 (3.6)||18,239 (4.8)|
|Unknown||19 (0.7)||2103 (0.5)|
|Year of diagnosis|
|1973–1978||436 (17.2)||57,459 (15.0)|
|1979–1983||409 (16.1)||55,965 (14.6)|
|1984–1988||468 (18.4)||76,576 (20.0)|
|1989–1993||558 (22.0)||88,978 (23.2)|
|1994–1998||667 (26.3)||104,173 (27.2)|
|San Francisco-Oakland||450 (8.9)||66,857 (8.7)|
|Connecticut||506 (10.0)||63,629 (8.3)|
|Detroit||513 (10.1)||68,958 (9.0)|
|Hawaii||77 (1.5)||14,471 (1.9)|
|Iowa||286 (5.6)||51,011 (6.7)|
|New Mexico||135 (2.7)||18,551 (2.4)|
|Seattle||306 (6.0)||53,499 (7.0)|
|Utah||110 (2.2)||17,707 (2.3)|
|Atlanta||155 (3.1)||28,468 (3.7)|
|Single||244 (9.6)||34,649 (9.0)|
|Married||1787 (70.4)||213,992 (55.9)|
|Separated||38 (1.5)||6974 (1.8)|
|Divorced||136 (5.4)||29,916 (7.8)|
|Widowed||237 (9.3)||86,121 (22.5)|
|Unknown||96 (3.8)||11,499 (3.0)|
The incidence curve for breast carcinoma in men by age at diagnosis is shown in Figure 1. As in women, the incidence of breast carcinoma in men increases with increasing age, with a flattening of the curve in the oldest age groups. Figure 2 shows the age-adjusted incidence over time for male breast carcinoma from 1973–1998. The incidence of breast carcinoma in men has significantly increased over the 26 years included in the study. Male breast carcinoma incidence rates have increased from 0.86 per 100,000 population during the period 1973–1978 (95% confidence interval [95% CI], 0.79–0.95) to 1.08 per 100,000 population during the period 1994–1998 (95% CI, 0.99–1.16) (P < 0.001). However, this 26% increase in incidence is smaller than the 52% increase observed for women over the same time period.
Patient and disease characteristics are displayed in Tables 1 and 2, respectively. The median age at diagnosis for male breast carcinoma is 67 years (range, 10–103 years) compared with 62 years (range, 11–108 years) for women (P < 0.001). There were significant differences noted for stage distribution (P < 0.001), tumor size (P < 0.001), and lymph node status (P < 0.001) between men and women with breast carcinoma. Using the SEER historic staging, 41.7% of men had tumors localized to their breast compared with 50.5% of women. Conversely, 37.7% of men had regional lymph node involvement compared with 29.2% of women. At the time of diagnosis, 6.9% of men versus 5.6% of women were found to have distant metastases. Similar trends were noted with regard to AJCC staging, which included patients diagnosed between 1988–1998. Analysis of tumor size, which also was available in patients diagnosed between 1988–1998, showed that 20.0% of women had tumors measuring < 1 cm compared with only 9.8% of men. Among patients with known lymph node status, men were 1.6 times as likely to have lymph node involvement (P < 0.001).
|Characteristics||Males No. (%)||Females No. (%)||P values|
|SEER historic stage||< 0.0001|
|In situ||228 (9.0)||40,327 (10.5)|
|Localized||1059 (41.7)||193,490 (50.5)|
|Regional||956 (37.7)||111,903 (29.2)|
|Distant||175 (6.9)||21,411 (5.6)|
|Unknown||120 (4.7)||16,020 (4.2)|
|AJCC stage (1988–1998)||< 0.0001|
|0||141 (10.5)||30,283 (14.4)|
|I||394 (29.4)||80,657 (38.4)|
|II||516 (38.5)||63,998 (27.7)|
|III||101 (7.5)||11,148 (5.3)|
|IV||76 (5.7)||8217 (3.9)|
|Unknown||114 (8.5)||15,631 (7.4)|
|Size (1988–1998) (cm)||< 0.0001|
|< 1||131 (9.8)||41,914 (20.0)|
|1–1.9||400 (29.8)||63,663 (30.3)|
|2–4.9||523 (39.0)||59,656 (28.4)|
|≥ 5||71 (5.3)||12,358 (5.9)|
|Unknown||217 (16.2)||32,344 (15.4)|
|Lymph node status (1988–1998)||< 0.0001|
|Negative||574 (42.8)||102,214 (48.7)|
|Positive||421 (31.4)||47,703 (22.7)|
|Unknown||347 (25.9)||60,018 (28.6)|
|1||148 (5.8)||24,114 (6.3)|
|2||512 (20.2)||66,211 (17.3)|
|3||508 (20.0)||76,476 (20.0)|
|Unknown||1370 (54.0)||216,350 (56.5)|
|Right||1220 (48.1)||184,413 (48.1)|
|Left||1271 (50.1)||193,591 (50.5)|
|Bilateral||2 (0.1)||867 (0.2)|
|Unknown||45 (1.8)||4,280 (1.1)|
|Ductal||1861 (76.5)||272,724 (74.3)|
|Papillary||63 (2.6)||2162 (0.6)|
|Carcinoma or adenocarcinoma, NOS||419 (17.2)||34,175 (9.3)|
|Medullary||12 (0.5)||6697 (1.8)|
|Mucinous||43 (1.8)||7977 (2.2)|
|Lobular||36 (1.5)||43,123 (11.8)|
|Estrogen receptor (1990–1998)||< 0.0001|
|Negative||64 (5.7)||28,675 (16.2)|
|Positive||616 (55.3)||91,062 (51.5)|
|Unknown||433 (38.9)||56,969 (32.2)|
|Progesterone receptor (1990–1998)||< 0.0001|
|Negative||124 (11.1)||38,736 (21.9)|
|Positive||536 (48.2)||77,459 (43.8)|
|Unknown||453 (40.7)||60,511 (34.2)|
Pathologic tumor characteristics such as histology and hormone receptor status also differed between men and women with breast carcinoma, as did the distribution of histologic subtypes (P < 0.001). For example, 93.7% of breast tumors in men were classified as either ductal or unclassified carcinomas compared with 83.6% of tumors in women. Lobular carcinomas, which accounted for 11.8% of breast carcinomas reported in women, comprised only 1.5% of male breast tumors. Papillary carcinomas comprised 2.6% of male breast carcinoma compared with 0.6% of female breast carcinoma. It is interesting to note that men had a markedly higher proportion of ER-positive tumors than women (P < 0.001). Among patients with known ER status, 90.6% of men and 76.0% of women had ER-positive disease. Similarly, men were more likely than women to have tumors expressing PR (81.2% vs. 66.7%; P < 0.001). The percentage of men with breast carcinoma who had an ER-positive tumor significantly increased with patient age, from 75% for a man ages 30–34 years to 94% for a man age ≥ 85 years (P = 0.04).
The overall 5-year survival rate was 63%, the 10-year survival rate was 41%, and the median survival period was 92 months (Table 3). The 5-year overall survival rates were 78% for patients with Stage I, 67% for patients with Stage II, 40% for patients with Stage III, and 19% for patients with Stage IV male breast carcinoma. Survival curves by stage are shown in Figure 3. The median survival times were 115 months, 98 months, 50 months, and 16 months, respectively, for Stage I–IV disease. The overall 5-year survival rates were found to be lower stage-by-stage for men versus women with breast carcinoma, with corresponding overall survival rates for women reported to be 88%, 75%, 49%, and 16%, respectively, for Stage I–IV disease. However, the differences disappear when relative survival rates, which adjust for the expected survival rate of the U.S. population for race, gender, and age, were calculated. The relative 5-year survival rates for men with Stage I–IV breast carcinoma were 96%, 84%, 52%, and 24%, respectively. For women, the corresponding rates were 99%, 84%, 55%, and 18%, respectively.
|Characteristics||5-yr survival||10-yr survival||Median survival (mo)||Log-rank, P value||Hazard ratio (95% CI)|
|Overall age (yrs)||0.63||0.41||92||—||—|
|< 65||0.74||0.55||140||< 0.0001||Reference|
|≥ 65||0.55||0.29||71||1.46 (1.38–1.55)|
|Tumor size (cm)|
|< 2||0.74||0.47||104||< 0.0001||Reference|
|> 5||0.37||0.23||35||1.26 (1.12–1.41)|
|Lymph node status|
Patient and disease characteristics were evaluated for prognostic significance in both univariate and multivariate analyses. As expected, patient age ≥ 65 years was associated with a higher risk of death. Larger tumor size was also associated with a higher risk of death as was lymph node involvement (hazard ratio [HR] = 1.41; 95% CI, 1.28–1.57). ER/PR positivity were associated with a nonstatistically significant decreased risk of death. A high tumor grade was found to predict a greater risk of death (Grade 3 vs. Grade 1: HR = 1.17; 95% CI, 1.09–1.25). Multivariate analysis was performed to determine the independent effect of each variable (Table 4). In the adjusted analysis, patient age ≥ 65 years (HR = 1.59), tumor size 2–5 cm (HR = 1.40), and lymph node involvement (HR = 1.50) were found to be independently associated with poorer survival. Tumor grade and ER/PR status were not found to be independent predictors of overall survival for men with breast carcinoma.
|Factors||Hazard ratio (95% CI)||P value|
|age at diagnosis (yrs)|
|≥ 65 +||1.59 (1.35–1.87)||< 0.0001|
|Tumor size (cm)|
|2–5||1.40 (1.20–1.62)||< 0.0001|
|> 5||1.07 (0.89–1.28)||0.46|
|Lymph node status|
|Positive||1.50 (1.29–1.74)||< 0.0001|
We have demonstrated that the age-adjusted incidence rate of male breast carcinoma is increasing. Although the increasing incidence of breast carcinoma in women has been well established, the rates of male breast carcinoma have been reported to be stable.13, 14 We have shown that the incidence of breast carcinoma in men has increased from 1973 to 1998 although, overall, it remains a rare condition. The reasons for this increase in incidence are unclear, but merit further investigation. It is interesting to note that male breast carcinoma incidence has not increased as dramatically as has the incidence of breast carcinoma in women. The larger increase in incidence rates noted in women may be due to improved ascertainment from increasing use of screening mammography. As in women, the incidence of breast carcinoma in men increases with age. However, the rate of increase in incidence does not appear to slow after the age of 50 years as is observed in women. This difference most likely is due to the lack of, or more gradual, change in the hormonal milieu than occurs in women at menopause.
The current study has also shown differences in the presenting characteristics of breast carcinoma in men and women. Men tend to be older at the time of diagnosis compared with women, with a median age at diagnosis of 67 years compared with 62 years for women. Several studies from Sweden, Denmark, and Croatia have suggested that men tended to be older than women at the time of a breast carcinoma diagnosis, but to our knowledge this finding has not been corroborated in a large population-based U.S. cohort.5, 15, 16 In addition, we have demonstrated that men have more advanced disease at presentation compared with women. For example, men present with higher-stage disease, larger tumors, and more frequent lymph node involvement. This finding may reflect that women are more likely to be diagnosed with very small, early-stage tumors by screening mammography. However, it also may reflect a lack of public awareness of breast carcinoma in men and subsequent delays in diagnosis. No guidelines recommend screening mammography at any age for men because of the rarity of the disease. More importantly, there is little public education regarding the existence of male breast carcinoma and there are no recommendations for self-examination or examination of the male breast by physicians in asymptomatic males.
Breast carcinoma in men has biologic differences from breast carcinoma in women. In particular, the distribution of histologic types varies between men and women. Previous small series have reported that 64–87% of men with breast carcinoma have invasive ductal histology.3, 4, 9, 17–19 We also found that the majority of men (93.4%) have invasive ductal or unclassified carcinoma histology. In contrast to women, in whom invasive lobular is the next most frequent histologic type, papillary carcinomas (2.6%) are the second most frequent histology in men. Lobular carcinomas occur in men, but are less common, perhaps due to the rarity of terminal lobules in the male breast. Although true lobules can develop in the male breast as a result of estrogen exposure, most male breast tissue lacks terminal lobules.4
Men also are significantly more likely to have ER/PR-positive disease than women. In this population-based cohort, > 90% of men had ER-positive tumors and 81% had PR-positive tumors, compared with 76% and 67% respectively in women. We have previously reviewed the literature to determine rates of ER/PR positivity in men with breast carcinoma and found that 81% of breast tumors in men expressed ER.20 This estimate was derived from 46 individual series, with what to our knowledge is the largest individual report including only 133 cases.21 In the current population-based series, the percentages are based on 680 cases tested for ER and 660 cases tested for PR. A higher proportion of tumors are ER positive than had been previously demonstrated. In contrast to previous reports,21–23 we have shown that the percentage of ER-positive tumors increases with increasing patient age. The current study findings may differ from that of other investigators simply because our large study population provided sufficient power to detect this association.
Previous studies had reported conflicting data regarding whether the prognosis of male patients with breast carcinoma differed from female patients with breast carcinoma.3, 10, 24, 25 The series from the National Cancer Data Base matched male and female patients with breast carcinoma by stage, age, ethnicity, income, and treating hospital.10 The authors reported that survival rates were equivalent for Stage I and II disease, but a trend was observed for poorer survival for men with Stage III and IV disease. We have found that the relative survival rates, which adjust for the general survival rate of the U.S. population by race, gender, age, and date at diagnosis, are very similar between men and women with Stage I–IV breast carcinoma. It is interesting to note that, in contrast to the findings in the National Cancer Data Base series, survival for men with Stage IV disease was slightly higher than for women (24% vs. 18%). Overall survival rates for men with breast carcinoma, stratified by stage of disease, are lower than for women with breast carcinoma. However, these differences are most likely due to the older age distribution of male patients and the lower life expectancy of men in the general population.
Consistent with previous studies,2–4, 26 we demonstrated that tumor size and axillary lymph node involvement are independent predictors of overall survival. Tumor grade has been reported to be a prognostic factor for men with breast carcinoma when univariate analyses were performed.6, 19 We also found that high-grade tumors were associated with poorer survival in univariate analysis. However, in multivariate analysis, no significant association was observed. Similarly, previous studies had shown that hormone receptor positivity was associated with improved survival for men with breast carcinoma.4, 9 We found a trend for improved survival with hormone receptor-positive disease. However, these effects were no longer present in the adjusted analyses.
The strengths of the current study are the large numbers of male patients, the long-term survival data, and the representativeness of the general male population. However, limitations include the lack of detailed information regarding biologic characteristics of interest such as BRCA1 and BRCA2 mutations, HER-2 status, and p53 status. In addition, the pathologic data are collected from many hospitals and have not been subjected to centralized review. In spite of these limitations, we have reported what to our knowledge is the largest population-based cohort of men with breast carcinoma reported to date. Of particular importance, these data suggest that the incidence of male breast carcinoma is increasing, that men are diagnosed with later-stage disease than women, and that hormone receptor status and tumor grade are not independent prognostic factors for men. We have also shown that the prognosis for men with breast carcinoma is similar to that of women with similar-stage disease. Future research on male breast carcinoma is needed in many areas. The rarity of the disease makes randomized trials impractical, but prospective data can and should be collected. Data are currently sparse regarding risk factors for male breast carcinoma, disease biology, and optimal treatments. Continued clinical research will help to advance the understanding of this rare disease.
- 12International classification of diseases for oncology. Geneva: World Health Organization, 1990., , .