Breast cancer is the most common cancer in women with a lifetime probability of one of seven according to Surveillance, Epidemiology, and End Results (SEER) program [13]. Due to the overwhelming relevance and frequency of this disease, current treatment of breast cancer in women is well defined and optimized. A comprehensive framework of guidelines for diagnosis and treatment exists for this entity [3, 12, 15, 31, 33], which is based on a large body of data from randomized, controlled trials. Beneficial and adverse effects of standard treatment are well characterized [10, 35, 36]. In contrast, male breast cancer is a rare and often overlooked disease which represents < 1% of all breast cancer cases [1]. Despite possible differences in pathogenesis, biology and genetics between both sexes, the treatment strategies for male breast cancer have been extrapolated from the large body of knowledge for female breast cancer [7, 37].

However, recent studies using large cancer registries or multi-institutional databases have only marginally increased knowledge about breast cancer in men. It has been assumed that male breast cancer carried a worse prognosis compared to breast cancer in women [1, 9]. These differences in prognosis have mainly been attributed to delays in diagnosis, higher stages and later ages associated with increased co-morbidity rather than to differences in tumor biology or to systematic differences in patterns of care between both groups [2, 9, 14].

Therefore, the aim of the present study was to analyze disease characteristics and treatment strategies of male breast cancer in a single institution over the last 25 years. In this regard, prognosis in terms of overall survival (OS), disease-free survival (DFS), and distant metastasis-free survival (DMFS) as well as adverse effects of treatment were evaluated. Furthermore, an estimated outcome was calculated for a “virtually” matched pair cohort of women by the Adjuvant!Online® 8.0 algorithm and compared to outcome of men in this series with no, same, and optimal adjuvant treatment from current point of view.

Patients and methods

Patient characteristics

To be included in this retrospective analysis, the following conditions had to be fulfilled: all men had to suffer from invasive breast cancer curatively treated with (modified radical) mastectomy and postoperative radiotherapy. Patients presenting with metastases, recurrences or second cancers were excluded. Men fulfilling these criteria were identified by the departmental database. Therefore, 21 out of an initial cohort of 61 patients were excluded due to the following reasons: palliative treatment (n = 9), no postoperative radiotherapy (n = 7), recurrence (n = 1) or second cancers (n = 4). After discussing the intended analysis, the institutional review board (ethics committee of our institution) had no objections (No. 395/2011A). Thus, 40 men—treated between 1982 and 2007—fulfilled the conditions and were analyzed. The median age of the cohort was 62 years (range 35–80 years). Tumor stages of all patients were converted to the seventh edition of AJCC classification [6] for easier comparison of prognosis. According to tumor and nodal stages most patients were classified as AJCC stage II and IIIb. Details of patients’ characteristics considering tumor and treatment-related factors are displayed in Tab. 1.

Tab. 1 Patients’ characteristics
Full size table

Surgery

A total of 37 patients (92.5%) were treated with a modified radical mastectomy, the remaining 3 patients underwent a simple mastectomy. No breast conserving surgeries were performed. Mastectomy provided clear margins in almost all patients (n = 38). Only 2 patients were classified having microscopic residual tumor (R1). Thirty-seven men (92.5%) received an axillary lymphadenectomy and 3 patients a sentinel lymphadenectomy (7.5%). A mean of 13 lymph nodes (range 1–33) were removed, while four (range 0–24) were affected. Of these patients, 65% (n = 26) presented with pathologically confirmed positive lymph nodes (Tab. 1).

Radiotherapy

Postoperative radiotherapy to the chest wall was fractionated with single doses of 1.8–2.5 Gy to a median dose of 50 Gy (range 45–50.4 Gy) either with tangential photon fields (n = 19) or by an electron beam rotational technique (n = 21) as described elsewhere [19]. Eleven patients (27.5%) received an additional boost to the tumor region of median 10 Gy (range 5–16 Gy). In 22 men (55%) regional lymphatics were included to a total dose of 50 Gy (Tab. 1). Supra-/infraclavicular fossa (n = 22) was treated by an anterior oblique field. Axillary nodes (n = 8) were included in the supra-/infraclavicular field or tangential fields. Additional treatment of parasternal nodes (n = 19) was performed using the mixed beam technique by extension of tangential fields and subsequently irradiated by a smaller tangential field plus a separate anterior electron field. Acute toxicity (within 90 days) and late toxicity after 90 days was assessed using RTOG criteria [4].

Systemic treatment

Further adjuvant treatment consisted of chemotherapy in 22.5% of patients (n = 9). Three different schedules were administered: CMF (cyclophosphamide, methotrexate, 5-fluoruracil), (F)EC (5-fluoruracil, epirubicin, cisplatinum), TAC (docetaxel, adriamycin, cyclophosphamide). Adjuvant antihormonal therapy was given in 55% (n = 22) of men and mainly performed with tamoxifen alone (n = 17). A few patients received aromatase inhibitors (n = 4) or a switch from tamoxifen to an aromatase inhibitor (n = 1).

Generating virtual matched pairs by Adjuvant!Online® 8.0

To compare outcome of men with that of women, an online calculator for overall survival of female breast cancer was employed (http://www.adjuvantonline.com). This evidence-based calculator was developed to evaluate benefits of receiving adjuvant therapy, i.e., chemotherapy, antihormonal therapy, or both. The origins of Adjuvant!Online® 8.0 were San Antonio Data Base, SEER database, and clinical trials as described elsewhere [32]. Matched pair survival data for female breast cancer patients were generated according to patient characteristics defined by age, receptor status, grading, size of tumor, and number of positive nodes. The 10-year overall survival was calculated for no/same like men/optimal adjuvant treatment.

Statistical analyses

OS, DFS, and DMFS were calculated using the Kaplan–Meier method from time of diagnosis [23]. Local failure was defined as any recurrence of tumor in the ipsilateral chest wall or in mastectomy scars. Recurrence at any other site was considered as distant failure (DMFS). Any event, i.e., local or distant failure, defined DFS.

Actuarial curves were compared by the two-tailed log-rank test. Statistical comparison of female matched pair survival data was performed with the paired t-test. A p value ≤ 0.05 was considered significant for both tests. The statistical analysis was done with the software package PASW 18.0 (SPSS Inc., Chicago, IL, USA).

Results

Acute and late radiation toxicity

Radiation treatment was completed as planned in all patients. Higher graded acute toxicities did not occur. Acute radiation dermatitis of grade 1 and 2 was detected in 32.5% (n = 13) and 45.0% (n = 18) patients, respectively. Nine patients developed no acute side effects. Esophagitis of grade 1 was detected in 2 men receiving radiotherapy of supraclavicular fossa. After a median follow-up of 47 months no late toxicity of G3 + was observed. Radiation atrophy and moderate teleangiectasia were detected in 30.0% (n = 12) and 7.5% (n = 3) of patients, respectively. Radiation fibrosis of grade 1 was determined in 5% (n = 2) of men.

Local control and survival figures

Local failure occurred in 1 patient resulting in an estimated 5-year local control rate of 96.2%. Local recurrence was treated by excision and adjuvant tamoxifen leading to no further evidence of disease in this man. At last follow-up, 8 patients experienced distant metastasis with predominance of bone (n = 7). In 3 patients, bony lesions were combined with lung or liver metastases. One additional man experienced lung metastases only.

The estimated OS, DMFS, and DFS at 5 years were 66%, 82%, and 79%, respectively (Fig. 1). The median follow-up of 47 months applies for toxicity, information on recurrences and metastases. Information on overall survival was updated by the tumor registry of our Comprehensive Cancer Center. Therefore, mean overall survival reached 120 months with a median of 84 months. Estimated 10-year overall survival was 43%.

Fig. 1
figure 1

Survival parameters. a Overall survival, b distant metastasis-free survival, and c disease-free survival were estimated using Kaplan–Meier method after a median follow-up of 47 months. Data on overall survival were updated with the tumor registry leading to a mean overall survival of 120 months (median 84 months) and an estimated 10-year overall survival of 43%

Full size image

Subanalyses of factors potentially influencing overall survival (OS)

To further investigate relevance of tumor- and treatment-related factors on OS the following subanalyses were performed. Impact of T-stage on OS was significant. Estimated 5-year OS for T1 and T2 were 81.8% and 80.4%, respectively, versus 30.3% for T4 stages (Fig. 2 a). Presence of node-negative disease resulted in an estimated 5-year OS of 68.8% vs. 63.1% for node-positive breast cancer (p = 0.08; Fig. 2 b). Given chemotherapy was associated with 100% 5-year OS compared to 58.2% without chemotherapy reaching a trend with p = 0.06 (Fig. 2 c). In contrast, administration of antihormonal treatment had no impact on OS with p = 0.516 (Fig. 2 d).

Fig. 2
figure 2

Relevance of tumor- and treatment-related factors on overall survival (OS). Five-year OS rates were estimated using Kaplan–Meier method. a Lower T-stage was associated with superior survival (T1: 81.8%, T2: 80.4% vs. T4:30.3%; p < 0.05), while b node-negativity reached only a trend to better outcome (pN0: 68.8% vs. pN +: 63.1%; p = 0.08). c Regarding treatment related factors, there was a trend to improved overall survival at 5 years for adjuvant chemotherapy (100% vs. 58.2%; p = 0.06). d Adjuvant antihormonal treatment (AHT) had no impact on overall survival with 60.4% vs. 71.1% (p = 0.516)

Full size image

Matched pairs by Adjuvant!Online® 8.0

To evaluate, whether male and female breast cancer had a similar outcome, matched pairs of males with virtual females were generated. The 10-year OS was calculated for the 40 male characteristics (defined by five items as described above) of this series with Adjuvant!Online® 8.0 in combination with two variables, i.e., adjuvant chemotherapy and/or antihormonal therapy. We used three constellations of both variables: no adjuvant systemic treatment, treatment as performed in this series, and optimal adjuvant treatment. For each characteristic (n = 40) of this series, three rates depending on adjuvant therapy were generated. Mean 10-year OS for virtual women without adjuvant treatment reached 55%, for virtual women with a treatment as performed in this male cohort 59% and for virtual women with optimal treatment 68%. Comparison of these OS rates by paired t-test was significant (p < 0.001). Regarding the 10-year OS of our cohort with estimated 43%, same postulated treatment in women according to Adjuvant!Online® 8.0 resulted in a 16% better OS.

Discussion

The outcome of the present male cohort of breast cancer patients in terms of OS was inferior to women as expected from some literature data [14] and confirmed by our matched pairs with virtual women. The main reason for treatment failure was distant metastases. The main prognostic factor for DFS and OS was T-stage. Unlike in larger series [1, 5, 11], the effect of T-stage was much more pronounced than the effect of nodal status. We only observed a trend toward inferior outcome for node-positive men. However, lower impact of nodal stage might also be caused by low number (35%) of node-negative men. Possibly due to anatomical differences in men, larger tumors tend to metastasize more frequently directly by a hematogenic pathway [22]. The relevance of control of micrometastases is further emphasized by the fact that given chemotherapy was a strong positive prognosticator in this group.

This historical cohort was definitely undertreated when current state-of-the-art adjuvant treatment for breast cancer is used as a measure for optimal therapy [3, 12, 15, 31, 33]. But the 25% inferior survival of men (43% versus 68%) in this series compared to women with optimal treatment in randomized trials according to the Adjuvant!Online® 8.0 algorithm might not only be attributable to suboptimal adjuvant treatment as further indicated by the fact that there was a difference of 12% in survival between the sexes with same adjuvant treatment. However, this comparison has to be used with caution since direct statistical comparison of real matched pairs was not available. Furthermore, all adjuvant assessment tools for prognosis including Adjuvant!Online®, the St Gallen Consensus, Oncotype DX®, and MammaPrint® have limitations [29]. Consideration of biological markers like Her2-neu, Mitotic Index, and Ki67 were not possible leading to an overestimation of survival in patients < 40 years old [18]. But, only 1 patient was < 40 years in this series, limiting the possible relevance of this confounder. Another limitation of the comparison is the given radiotherapy in all male patients, while radiotherapy is not taken into consideration in Adjuvant!Online®. However, as all of our male patients had standard indication for radiotherapy, treatment without radiation in the respective female groups would have to be rated as undertreatment and would suggest a worse outcome of the female comparison. As Adjuvant!Online® comprises randomized data published from the early 1970s [16, 17, 32] to the present, the time period of this study is completely matched compensating for potential influence of historical radiation treatment techniques which could be associated with increased cardiotoxicity.

A population-based comparison of SEER data also demonstrated superior survival of 10% for women with breast cancer compared to men [1]. These SEER data may reflect differences in pattern of care between men and women, but it may also point to systematic and relevant biological differences between both groups. A recently published matched pair analysis from Sweden further corroborates this theory with a 14% inferior survival in men [28]. But, radiation treatment was significantly less often delivered in male patients. Therefore, the current series clearly substantiates that even with maximal local treatment (all patients received mastectomy and radiotherapy) outcome of men with breast cancer is inferior compared to virtual matched women treated in randomized trials. We first used the evidence-based calculator Adjuvant!Online® 8.0 for comparison of survival of men and women with breast cancer. Even considering the known limitations of such a tool, the fact that calculated OS of women of same stage and treatment was clearly superior compared to men and the predominant influence of T-stage point to more aggressive tumor biology or other biological differences between the two groups.

Recent studies documented improved survival rates over recent decades for both male and female breast cancer, but progress for men has lagged behind that for women [1]. In this series, the same effect was seen mainly attributable to application of chemotherapy which predominately was administered in the last decade. An additional reason for slower increase of survival figures or different biological features in men could be a limited benefit from antihormonal treatment [8, 30]. Thus, it is uncertain whether hormone receptor positivity has the same prognostic implication in male as in the female disease [27, 37].

In female breast cancer, durable local control is the precondition for long-term survival and leads to a reduction of distant metastases [34]. Excellent local control in men was achieved in this intermediate to high-risk population comparable to other trials of both genders [25, 26, 38]. Moreover, local treatment was as well tolerated as in women [1, 20, 21]. Therefore, intensification of systemic treatment is a viable option to further increase cure rates for men.

Conclusion

Compared to historical and virtual matched cohorts of women our male breast cancer patients had inferior outcome suggesting that not only undertreatment, but also biological differences either of tumor or of population may contribute to the worse prognosis. This assumption is supported by the predominant impact of T-stage on prognosis, while nodal stage did influence outcome to a much lesser extent and the limited effect of adjuvant endocrine therapy. It is unclear whether inferior prognosis could be completely compensated by the indicated more frequent use of chemotherapy. Therefore, further treatment- and outcome-related investigations including multi-institutional clinical trials are necessary [24].