Kaplan-Meier curve for overall survival of male and matched female patients with breast cancer.
Kaplan-Meier curve for disease-specific survival of male and matched female patients with breast cancer.
Causes of mortality of male patients with breast cancer.
El-Tamer MB, Komenaka IK, Troxel A, Li H, Joseph K, Ditkoff B, Schnabel FR, Kinne DW. Men With Breast Cancer Have Better Disease-Specific Survival Than Women. Arch Surg. 2004;139(10):1079–1082. doi:10.1001/archsurg.139.10.1079
Male breast cancer patients have better disease-specific survival than carefully matched female breast cancer patients.
Patients and Methods
Each man in the breast cancer database at Columbia-Presbyterian Medical Center (New York, NY) between the years 1980 and 1998 was matched with a woman. Matching was done based on age and date of diagnosis, stage, and primary histologic findings.
Main Outcome Measures
The overall survivals and disease-specific survivals of the male breast cancer group and female breast cancer group were compared.
Fifty-three male patients were matched with an equal number of female breast cancer patients. The Kaplan-Meier curves demonstrated that there was no significant difference in overall survival. The 5- and 10-year survivals for women were 0.77 and 0.51, and for men 0.77 and 0.56. When the Kaplan-Meier curves for breast cancer–specific survival were compared, however, there was a significant difference in the 5- and 10-year survivals (P = .05, log-rank test). For women, the 5- and 10-year disease-specific survival was 0.81 and 0.7, respectively, while for men it was 0.9 and 0.9, respectively. In a Cox regression analysis for time to death from breast cancer, stage was the only predictor of death that approached significance (P = .06).
While the overall survivals were equivalent, male breast cancer patients had significantly better disease-specific survivals compared with their female counterparts. Male patients were 4 times more likely to die of other causes than their breast cancer.
Male breast cancer is a relatively uncommon disease. It accounts for only 1% of all cases of breast cancer and less than 1% of all malignancies in men. In 2004, approximately 1450 new cases of male breast cancer are anticipated in the United States, and 470 men will die of this disease.1 In contrast to the increasing age-adjusted incidence of breast cancer in women, the incidence of breast cancer in men has remained stable over the past 4 decades.2- 4 Because of this low incidence, unlike carcinoma of the breast in women, a limited number of patients are available for study. Even large oncologic centers have relatively small numbers. The resulting paucity of studies has contributed to a lack of knowledge and understanding of the disease.
Recent reviews using large databases or multi-institutional data, however, have increased the understanding of the natural history of breast cancer in men. It was long considered that male breast cancer carried a worse prognosis than breast cancer in women.5- 7 In studies where male patients were matched with similar female patients by age and stage, however, the survivals were equivalent.8,9 The previously observed differences in prognosis were thought to result from delays in diagnosis rather than a different biological profile or tumor aggressiveness.10,11 There are no data available comparing disease-specific survival in these 2 groups. Our objective is to compare disease-specific survival in a group of carefully matched men and women with breast cancer.
Each male breast cancer patient in the prospective breast cancer database at Columbia-Presbyterian Medical Center (New York, NY) between the years 1980 and 1998 was carefully matched with a woman from the same database. Male patients with invasive breast cancer, carcinoma in situ, or components of both were included in the study.
The matching was done based on age at diagnosis (within 1 year), date of diagnosis (within 7 years), and stage and histologic findings of the primary cancer. The women were chosen from a total of 7676 consecutive patients who underwent treatment at our institution from 1980 to 1998. When more than 1 woman met the criteria for matching, the female patient whose age at diagnosis was closest to the male patient was chosen.
The database was reviewed to obtain the demographics, clinical history, stage, and histologic findings of the diagnoses. Follow-up information was obtained from the database, the medical records, and contact with the patients or their family. Mortality due to breast cancer, other cancers, and non–cancer-related causes were the end points considered.
We determined and compared the 5- and 10-year survival for male and female matched groups. Both the overall survival time and breast cancer–specific survival times were determined. Survival curves were calculated by the Kaplan-Meier method for time to death overall and for time to death from breast cancer.12 Cox regression analysis was then used to determine predictors of death.13
The 53 male breast cancer patients were matched to an equal number of female patients during the same period. The median age at diagnosis was 70 years (range, 37-90 years) for women and 70 years (range, 38 -91 years) for men. The median follow-up was 5.2 years for women and 5.5 years for men. Table 1 displays the demographic and pathologic data of the matched male and female breast cancer patients. Invasive ductal carcinoma without an in situ component was found in 31 (58%) of the tumors. In addition, 1 invasive cancer was the papillary subtype and another was the tubular subtype. Two (6%) of these patients also had findings of Paget's disease. Eleven (21%) of the patients had pure ductal carcinoma in situ (DCIS). Nine (17%) had components of both invasive ductal carcinoma as well as areas of DCIS in the specimen. Of those who underwent an axillary node dissection, 20 (44%) of 46 had positive nodes (range, 1-29 positive nodes). There were no cases of invasive or in situ lobular carcinoma identified.
The treatments of both men and women are presented in Table 2. In men, modified radical mastectomy was performed in 36 patients (67%), radical mastectomy in 9 (17%), and total mastectomy in 5 (9%). Two (4%) had lumpectomy, and 1 had lumpectomy with axillary dissection. The treatment of matched women was similar (Table 2). Although it appeared that a greater percentage of women received chemotherapy and men more frequently received hormonal therapy, these differences were not significant by χ2 test.
The Kaplan-Meier curves for overall survival (Figure 1) demonstrated no significant difference between the male patients and the matched female patients (P = .94, log-rank test). The 5- and 10-year overall survivals for women were 0.77 (95% confidence intervals [CI], 0.65-0.92) and 0.51 (95% CI, 0.34-0.75), with a median survival of 11.6 years. The 5- and 10-year survivals for men were 0.77 (95% CI, 0.66-0.91) and 0.57 (95% CI, 0.42-0.78), with a median survival of 11.1 years. In the Cox regression analysis of overall survival, age at diagnosis was the only significant predictor of death (P = .03). Sex was not significant after adjusting for other matching factors.
The Kaplan-Meier curves for breast cancer–specific survival (Figure 2) demonstrated a significant difference between men and women (P = .05, log-rank test). The 5- and 10-year breast cancer–specific survivals were 0.81 (95% CI, 0.7-0.91) and 0.7 (95% CI, 0.56-0.88), respectively, for women, and 0.9 (95% CI, 0.81-0.99) and 0.9 (95% CI, 0.81-0.99), respectively, for men. The outcomes of the 2 groups are presented in Table 3. Of the 16 men who died, only 3 patients (19%) died of breast cancer. Nine (56%) of 16 men died of causes other than cancer (Figure 3). In the Cox regression analysis for time to death from breast cancer, stage was the only predictor of death that approached significance (P = .06).
It was long thought that male breast cancer carried a poor prognosis compared with breast cancer in women. Many hypothesized that a different tumor biological profile in men was responsible for the poor outcomes.5,14 With time, and as more cases became available for review, retrospective studies showed similar predictors of outcome for men and women with breast cancer. Older age at diagnosis and more advanced stage of disease at diagnosis have consistently been associated with poor overall survivals in both men and women. Previous series10,15,16 have reported the median age at diagnosis of male breast cancer to range from 59 years to 62 years. This makes the age at diagnosis of breast cancer in men 5 to 11 years later than in women.15,17,18 The median age at diagnosis for male patients in our series was 70 years. This is significantly older than the median age for females (58 years; mean 58.1 years) with breast cancer at our institution during the same period. Thus, the difference in age of 12 years between male breast cancer patients and female breast cancer patients was comparable to the previously reported figures.
The overall survival of the matched male and female breast cancer patients from our institution was similar (Figure 1). This finding is supported by previous studies of matched male and female patients.8,9 The review by O'Malley et al19 using the data from the Surveillance, Epidemiology, and End Results program (SEER) reported 5-year overall survival rates for white patients, black patients, and other men as 66%, 57%, and 75%, respectively. The 5-year overall survival rate of our patients was slightly above their range at 77%. One reason for this difference may have been an earlier stage of diagnosis in our population. This earlier diagnosis may be related to greater knowledge of breast cancer seen in the more contemporary period (starting in 1980) of our data. Further, the higher socioeconomic status of our patient population may have contributed to better access to medical care.
When Kaplan-Meier survival curves were created for breast cancer-specific survival, however, men had significantly better disease-specific survival (Figure 2). The male patients in our review were more likely to die of other causes rather than of breast cancer (Figure 3). With increasing age, other diseases and malignancies become more prevalent and result in more deaths in male than in female patients of similar age at diagnosis and stage of disease. In the female group, 11 of 12 reported deaths were from breast cancer. By contrast, 3 times as many male patients died of non–cancer-related causes (9/16 [56%]) than from breast cancer (3/16 [19%]). Other causes of mortality were primarily related to cardiac and vascular morbidities. Further, male patients were equally as likely to succumb to other malignancies as they were to breast cancer. The other cancers resulting in mortality in our series were colon, prostate, and lung. Statistically, in the general male population, the most common causes of cancer-related mortality are lung, colon, and prostate.1 In the review by Cituli et al20, the 5- and 10-year overall survival rates of male breast cancer patients were 65% and 38%, respectively, while the disease-specific survival rates were 74% and 51%, respectively. O'Malley et al19 also reported a seemingly higher rate of breast cancer–specific survival in all races of male patients, 83%, 72%, and 89% for white patients, black patients, and other men, respectively, compared with overall survival. These reviews demonstrated lower absolute percentages than found in our data. Some reasons for the higher percentages in our data may be the smaller data sample from a single institution. The racial composition of our patient group appeared similar to that reported with 80% white patients, 14% black patients, and 5% other races. However, our patient group had a higher proportion of patients older than 70 years. This older age at diagnosis may have influenced a higher breast cancer–specific survival. Our population may be at increased risk for death from other comorbidities, which accumulate with advancing age and therefore become more likely to kill a patient before breast cancer.
Previous reviews comparing matched male and female patients, however, have not looked at the distinction between overall survival and breast cancer–specific survival. Our finding of significantly better disease-specific survival is interesting because it suggests that overall survival may not be the most accurate end point of prognosis in male breast cancer. Our statistical analysis demonstrated that age was the only significant prognostic indicator of overall survival. This finding suggests that overall outcomes are related to the age of the patient rather than the investigated disease. Therefore, in male breast cancer, as in other cancers that appear at older ages with a relatively good prognosis, one must specify the end points and causes of mortality.
Previous reviews have reported that axillary node involvement was an important predictor of prognosis, and in some studies it has been the most important factor.10,21,22 Other studies have found that the previously reported poor outcomes in male breast cancer patients were thought to arise from a delay in diagnosis, later treatment, and advanced stage at diagnosis.10,15 A delay in diagnosis of 18 months has been previously reported.15 More recently the time to diagnosis has decreased to a median delay of approximately 18 weeks.10 A review from Austria reported that 10% of patients had metastatic disease at the time of first treatment.23 Another series from France reported 56% of patients had axillary nodal involvement.20 Our series is more contemporary with less than half (44%) of the patients having positive lymph nodes and 21% of patients with purely DCIS, demonstrating a trend toward earlier diagnosis. This incidence of DCIS is higher than the 2.3% to 4.8% of previously reported older series.20,23- 25 The earlier stage at diagnosis in our review probably explains the higher disease-specific survival rates. Both male and matched female groups, however, had similar numbers of DCIS; yet the male patients still demonstrated a significantly better breast cancer–specific survival (Figure 2).
The Cox regression analysis of factors influencing disease-specific survival revealed that stage of disease was the only predictor of death that approached significance (P = .06). Although men demonstrated a significantly better disease-specific survival compared with their female counterparts on univariate analysis, sex was not as important as stage at diagnosis. This confirms the expected finding that regardless of sex, those who are diagnosed with advanced disease have the worst prognosis.
In conclusion, our finding of significantly better breast cancer–specific survival in male breast cancer patients is interesting because it suggests that overall survival may not be the most accurate end point in male breast cancer outcomes. Similarly, when comparing elderly patients with cancer, particularly in uncommon types of cancer for a specific sex, then disease-specific survival must be studied. This is one of the first studies to compare breast cancer–specific survival based on sex. Given the small number of subjects, further research from large population-based samples is encouraged.
Correspondence: Mahmoud El-Tamer, MD, Atchley Pavilion, 10th Floor, 161 Fort Washington Ave, New York, NY 10032 (firstname.lastname@example.org).
Accepted for publication March 9, 2004.