Impact of adding preoperative magnetic resonance imaging to ultrasonography on male breast cancer survival: a matched analysis with female breast cancer

Compliance with Ethical Standards

This retrospective study received approval from the Institutional Review Board of the Ethics Committee, which waived the requirement for obtaining written informed consent (Samsung Medical Center, IRB No. 2022-11-068). The study was conducted in accordance with the guidelines of the Declaration of Helsinki, as revised in 2013.

Study Population

Male and female breast cancer patients who underwent curative breast surgery at the authors’ institution between January 2005 and December 2020, and who did not receive neoadjuvant chemotherapy, were initially screened for inclusion using an electronic medical database. The exclusion criteria included: (1) patients with current bilateral breast cancer; (2) patients with a history of previously treated breast cancer; and (3) patients with other malignancies of non-breast origin, such as lymphoma, metastasis, or sarcoma. Ultimately, 61 male patients were included in this study. Exact matching analysis was conducted between male and female patients at a ratio of up to 1:6, based on age, pathologic T and N stage, and molecular subtypes. This matching ratio was chosen to account for the significantly lower incidence of male breast cancer compared to female breast cancer and to minimize variance within the matched cohort [1,24]. During the screening period, 28,191 female patients underwent curative breast surgery. Of these, 267 were matched with the 61 male breast cancer patients (Fig. 1).

Fig. 1.

Flowchart of patient selection.

Image Acquisition and Interpretation

All patients underwent standard two-view screening or diagnostic mammography at the time of their cancer diagnosis. For preoperative staging, US was performed on all patients, while MRI was conducted only when deemed necessary by the clinician or upon the patient's request. Preoperative US was carried out by one of 12 board-certified radiologists—seven staff radiologists with 10–26 years of experience and five fellow radiologists with 1-2 years of experience in breast imaging. The equipment used included an iU22 machine (Philips Advanced Technology Laboratories, Bothell, WA, USA) with a 12-5-MHz linear array transducer, an RS80A system (Samsung Medison Co., Ltd., Seoul, Korea) with a 3-12-MHz linear high-frequency transducer, or an Aixplorer system (SuperSonic Imagine, Aix-en-Provence, France) with a 15-4 MHz linear array transducer. The choice of machine varied depending on when the patients underwent preoperative US. MRI was performed using either a 1.5T or 3T Philips Achieva MR system (Philips Medical Systems, Bothell, WA, USA) equipped with a dedicated bilateral phased-array breast coil, with the patient in the prone position. The MRI protocol included turbo spin-echo T1- and T2-weighted sequences and a three-dimensional dynamic contrast-enhanced sequence. The axial dynamic contrast-enhanced MRI protocol comprised one pre-contrast and five post-contrast dynamic series. A bolus injection of 0.1 mmol per kg of body weight of gadobutrol (Gadovist, Bayer Healthcare, Leverkusen, Germany) was administered, followed by a 20 mL saline flush. Dynamic images were captured five times, beginning 30 seconds after contrast administration and subsequently every 60 seconds, using a gradient echo sequence. Subtraction images were automatically generated pixel-by-pixel by subtracting the pre-contrast images from the early post-contrast images. The total scan time ranged from 25 to 27 minutes. The MRI examinations were conducted after the US examinations in all cases. The interval between US and MRI varied from 1 to 30 days (mean±standard deviation, 4.7±6.2 days).

US examinations were interpreted by the radiologist who conducted the US, while MRI examinations were interpreted by one of 12 board-certified radiologists with a specialization in breast imaging. Both US and MRI findings were assessed using the Breast Imaging Reporting and Data System [25]. During their evaluations of US and MRI, radiologists identified suspicious additional lesions in both breasts, in addition to the biopsy-confirmed index cancer. They documented the anatomic locations and sizes of the lesions and evaluated the potential for lymph node metastasis in the axillary, internal mammary, and supraclavicular regions. US assessments were based solely on US findings, as US image acquisition and interpretation always preceded MRI acquisition. In contrast, MRI readers had access to US images during their evaluations.

For clinical staging, the clinical T and N stages were determined using either US alone or in combination with MRI. The clinical T stage was assessed by measuring the longest diameter of the index cancer in the images. The clinical N stage was established based on the number and location (axillary, internal mammary, and supraclavicular regions) of suspicious lymph nodes identified in US or MRI scans. Lymph nodes were considered suspicious on imaging if they met at least one of the following criteria: (1) cortical thickness greater than 3 mm, (2) focal cortical bulging or eccentric cortical thickening, (3) rounded shape, (4) complete or partial effacement of the fatty hilum, and (5) complete or partial replacement of the lymph node with an ill-defined or irregular mass [26,27]. In the group assessed with both US and MRI, the clinical T stage was determined by the larger of the longest diameters of the index cancer measured in each modality, and the clinical N stage was determined by the greater number of suspicious lymph nodes counted in each modality.

Data Collection

Clinical characteristics such as age at diagnosis, family history of breast cancer, presence of BRCA1/2 gene mutations, and type of surgery were analyzed. Pathological data, including the size of the index cancer, nodal status, TNM stage according to the American Joint Committee on Cancer (AJCC) 7th edition, histologic type, estrogen receptor (ER) status, progesterone receptor (PR) status, human epidermal growth factor receptor 2 (HER2) status, histologic grade for invasive breast cancer, nuclear grade, and Ki-67 index, were also reviewed. The pathologic index cancer size was defined as the size of the invasive component alone for invasive breast cancer and as the extent of ductal carcinoma in situ (DCIS) for pure DCIS. Molecular subtypes were categorized into four groups according to the St. Gallen 2013 breast cancer classification categories [28]: luminal A (ER+/PR+/HER2-, Ki-67 <20%), luminal B (luminal B HER2 negative: ER+/HER2– and at least one of Ki-67 ≥20% or PR-, luminal B HER2 positive: ER+/HER2+ with any PR, Ki-67), HER2- enriched (ER–/PR–/HER2+), and triple-negative breast cancer (ER–/PR-/HER2-). Imaging data, including preoperative imaging methods (US alone vs. US combined with MRI), the size of the index cancer measured by US and MRI, and clinical T and N stages determined by preoperative imaging methods, were reviewed.

After curative surgery, patients received adjuvant radiation therapy, chemotherapy, and endocrine therapy, tailored according to the pathologic size and molecular subtype of the index cancers, as well as individual circumstances. In the first 2 years post-surgery, follow-up imaging surveillance was conducted every 6 or 12 months using mammography or US. After this period, imaging surveillance intervals were extended to 12 or 24 months. Disease recurrence was defined as the development of a second breast cancer (either ipsilateral or contralateral), lymph node metastasis (including ipsilateral axillary, internal mammary, infraclavicular, or supraclavicular nodes), or distant metastasis (encompassing systemic metastases and contralateral axillary lymph nodes) [29]. The diagnosis of recurrence was confirmed pathologically through biopsy. If a biopsy was not feasible, recurrence was diagnosed using other imaging modalities, such as positron emission tomography or computed tomography.

Statistical Analysis

Statistical analyses were performed according to variable type, using the Wilcoxon rank sum test for continuous variables and either the chi-square test or the Fisher exact test for categorical variables, to compare baseline characteristics between male and female patients.

Kaplan-Meier analysis was used to assess survival outcomes in both male and female patients. DFS was determined as the period from the surgery date to the occurrence of either recurrence or death, whichever came first. Patients who did not experience recurrence and were still alive were censored at their last follow-up. OS was defined as the time from surgery to death, with surviving patients censored at their most recent follow-up. The differences in DFS and OS between male and female breast cancer patients were analyzed using log-rank tests. To identify prognostic factors related to DFS or OS in a matched cohort, Cox proportional hazards regression analysis was performed. The hazard ratio (HR) and 95% confidence interval (CI) were calculated for various clinicopathological factors.

Additionally, weighted kappa statistics were utilized to evaluate the concordance between the pathologic T and N stages and the clinical T and N stages as determined by two preoperative imaging techniques: US alone and US combined with MRI. For cases of DCIS, the T stage was consistently classified as Tis, irrespective of the lesion size, in accordance with the seventh edition of the AJCC [30]. DCIS cases were excluded from the analysis of the agreement between clinical T stage and pathologic T stage. Kappa values of 0.00-0.20 were considered to indicate poor agreement: 0.21-0.40, fair agreement; 0.41-0.60, moderate agreement; 0.61-0.80, good agreement; and 0.81-1.00, excellent agreement.

All statistical tests were two-sided, and a P-value of less than 0.05 was considered statistically significant. The statistical analyses were conducted using SAS version 9.4 (SAS Institute, Cary, NC, USA).

Patient Characteristics

A total of 328 breast cancer patients were included in the study, comprising 61 men (median age, 57 years; interquartile range [IQR], 49 to 63 years) and 267 women (median age, 56 years; IQR, 49 to 63 years). The characteristics of the matched male and female cohorts are summarized in Table 1. In both groups, the majority of patients were classified as pathologic T1 (60.7% of men and 62.2% of women) and N0 stages (50.8% of men and 50.9% of women). Since the male patients presented only with luminal A or B molecular subtypes, the analysis for female patients was restricted to these luminal subtypes.

Table 1.

Characteristics of male and female breast cancer patients

The male patients exhibited a higher frequency of BRCA mutations (9.8% vs. 1.1%, P<0.001) and a smaller pathologic size of the index cancer (median, 1.9 cm vs. 2.1 cm; P=0.001) compared to female patients. Additionally, the use of US combined with MRI was less frequent in male patients than in female patients (75.4% vs. 95.9%, P<0.001), and the rate of breast-conserving surgery was significantly lower in men than in women (4.9% vs. 74.5%, P<0.001). No significant differences were observed in the other variables between the matched cohorts.

Survival Outcomes and Prognostic Factors

During the follow-up period (median, 65 months; IQR, 34 to 101 months), disease recurrence was observed in nine male breast cancer patients (14.8%, consisting of four with ipsilateral lymph node recurrence, one with contralateral breast recurrence, and four with distant metastases) and 13 female breast cancer patients (4.9%, including one with ipsilateral breast recurrence, one with ipsilateral lymph node recurrence, four with contralateral breast recurrence, and seven with distant metastases). There was a significant difference in recurrence rates between the two groups (P=0.001). Kaplan-Meier survival analysis of a matched cohort revealed that DFS in male patients was inferior to that in female patients (P=0.001) (Fig. 2A). In univariable Cox proportional hazards regression analysis, several clinicopathological factors were examined; pathologic index cancer size (HR, 2.262; 95% CI, 1.273 to 4.018; P=0.005) and higher pathologic N stage (N0 vs. N2-3) (HR, 5.739; 95% CI, 1.029 to 31.994; P=0.046) were significantly associated with DFS in male patients (Table 2). Multivariable Cox proportional hazards regression analysis showed that only pathologic index cancer size (HR, 2.013; 95% CI, 1.063 to 3.810; P=0.032) was associated with DFS in male patients. In female patients, only a higher pathologic N stage (N0 vs. N2-3) was significantly associated with worse DFS (HR, 7.867; 95% CI, 1.623 to 38.122; P=0.010) (Table 2).

Fig. 2.

Kaplan-Meier survival curves for disease-free survival (DFS) (A) and overall survival (OS) (B) in male and female breast cancer patients.

CI, confidence interval.

Table 2.

Factors associated with disease-free survival in male and female breast cancer patients

During the same follow-up period, eight male patients died (two from breast cancer-related causes and six from other serious illnesses or unknown causes), and seven female patients died (two from breast cancer-related causes, one from other malignancies, and four from other serious illnesses or unknown causes). Kaplan-Meier survival analysis of a matched cohort revealed that OS in male patients was significantly lower than in female patients (P<0.001) (Fig. 2B). Additionally, the univariable Cox proportional hazards regression analysis identified no significant prognostic factors for OS (Table 3).

Table 3.

Factors associated with overall survival in male and female breast cancer patients

Adding breast MRI to US for preoperative imaging was not associated with either DFS (HR, 1.117; 95% CI, 0.223 to 5.583; P=0.893) or OS (HR, 1.529; 95% CI, 0.3 to 7.781; P=0.609) in male patients (Fig. 3). In female patients, US combined with MRI was also not associated with DFS or OS.

Fig. 3.

A 59-year-old male patient with invasive ductal carcinoma in the left breast.

A. On ultrasonography (US), the index cancer is noted as an oval shape with a non-circumscribed margin in the left subareolar area and measures approximately 1.5 cm along the transverse axis, suggesting clinical T1 stage. US revealed no abnormal lymph nodes in the left axilla, suggesting clinical N0 staging (image not shown). B. On magnetic resonance imaging, the index cancer is noted as an oval mass with heterogeneous enhancement in the axial fat-saturated early phase of contrast-enhanced T1-weighted image. The mass is approximately 1.3 cm on the transverse axis, corresponding to the same clinical T1 stage as the US, despite a smaller side. No abnormal lymph nodes were observed in the left axilla, suggesting clinical N0 staging (image not shown). After left mastectomy, the pathologic index cancer size was confirmed as 1.5 cm and there were no metastatic lymph nodes on the final pathology report, suggestive of pathologic T1N0 disease.

Preoperative T and N Staging of US Alone vs. US Combined with MRI

To investigate the usefulness of MRI in male patients, the concordance was assessed between the pathologic T and N stages and the clinical T and N stages determined by preoperative imaging techniques (US alone or US combined with MRI), as shown in Table 4 and Supplementary Table 1. In male patients, the concordance between the pathologic T stage and the clinical T stage determined by US alone (κ=0.616) indicated good agreement, whereas US combined with MRI (κ=0.585) demonstrated moderate agreement, with US alone showing better concordance than US combined with MRI. The agreement between the pathologic N stage and the clinical N stage, as determined by US alone (κ=0.377) or US combined with MRI (κ=0.337), was fair. In female patients, the agreement between the pathologic T stage and the clinical T stage, as determined by US alone (κ=0.535) or US combined with MRI (κ=0.542), was moderate. The concordance between the pathologic N stage and the clinical N stage was moderate when determined by US alone (κ=0.428) and fair when determined by US combined with MRI (κ=0.365).

Table 4.

Agreement between preoperative T and N staging by US alone and US combined with MRI and final pathologic T and N staging