Clinical and ultrasonographic features of male breast tumors: A retrospective analysis

Objective The purpose of this study was to determine clinical and ultrasonographic characteristics of male breast tumors. Methods The medical records of male patients with breast lesions were retrieved from an electronic medical record database and a pathology database and retrospectively reviewed. A total of 112 men (125 breast masses) with preoperative breast ultrasonography (US) were included (median age, 59.50 years; age range, 15–96 years). Data extracted included patient age, if the lesions were bilateral, palpable, and tender, and the presence of nipple discharge. Breast lesion features on static US images were reviewed by three experienced radiologists without knowledge of physical examination or pathology results, original breast US image interpretations, or surgical outcomes. The US features were documented according to the BI-RADS (Breast Imaging-Reporting and Data System) US lexicons. A forth radiologist compiled the data for analysis. Results Of the 125 breast masses, palpable tender lumps and bilateral synchronous masses were more likely to be benign than malignant (both, 100% vs 0%, P < 0.05). Advanced age and bloody discharge from nipples were common in malignant lesions (P <0.05). A mass eccentric to a nipple, irregular shape, the presence of an echogenic halo, predominantly internal vascularity, and rich color flow signal on color Doppler ultrasound were significantly related to malignancy (all, P < 0.05). An echogenic halo and the presence of rich color flow signal were independent predictors of malignancy. Conclusion Specific clinical and US characteristics of male breast tumors may help guide treatment, and determine if surgery or conservative treatment is preferable.

Introduction Inclusion criteria for this study were patients with preoperative breast sonographic images and biopsy and/or postoperative pathology of male breast lesions. Once patients received operation in TVGH, surgeon always submit their biopsy specimens or excised tissue to the pathology department in TVGH for examination. Patients with only imaging results or clinical diagnosis were excluded. Biopsy or postoperative pathological examination results served as the final diagnostic standard.
A research assistant blind to the study hypothesis abstracted medical findings and pathology results from the electronic medical record database and pathology database at TVGH using the index terms "male" and "breast". Over an 8-year period from January 1, 2007 to December 31, 2014, 1,319 male patients had records matching the search terms.
Of the 1,319 patients, 1,110 who did not receive a biopsy or surgery were excluded. Of the remaining 209 patients who received either a biopsy or surgery, 97 were excluded: 11 with inadequate specimens, 19 who were actually females, and 67 without preoperative breast US. Thus, a total of 112 males who had breast US and a postoperative pathological diagnosis were included in the study.

Sonographic examination
Ultrasound was performed from 1 day to 1 week of the physical examination. Technicians with at least 10 years' experience at the departments of radiology performed the bilateral breast US examinations. Examinations were performed with the patient in the supine position with a scanner using high-resolution linear probes (L12-5 MHz) (S2000 Siemens Healthcare, Mountain View, CA; HDI 5000 and IU 22; Philips Healthcare, Bothell, WA; and LOGIQ E9; GE Healthcare, Milwaukee, WI). The technicians also performed color Doppler US (CDU) if a breast mass was detected, using an appropriate Doppler setting after the gray scale US study. The Doppler gain setting was adjusted to a level associated with minimal noise. The technicians examined a breast in four planes (sagittal, transverse, radial, and orthogonal), and obtained images at the 12 o'clock, 2 o'clock, 4 o'clock, 6 o'clock, 8 o'clock, and 10 o'clock positions, and of the nipple and axilla. If something suspicious was detected, multiple images in different planes were obtained on gray-scale US and CDU. The size, site, and distant to the nipple of lesions were recorded. After the bilateral examination was completed, all images were uploaded to a picture archiving and communication system (PACS) for storage. Revision of static US images was not allowed.

Sonography interpretation
Two experienced radiologists (H-CH, with 20 years of experience, and Y-YC with 10 years of experience), without knowledge of the physical examination findings, the previous preoperative breast image interpretation, or the histopathological results, reinterpreted the preoperative static breast US images of all 125 masses. Final results were determined by consensus. Disagreement were resolved by a third radiologist (Y-HC with 35 years of experience). The following 13 features were recorded for each of the images: maximum diameter of the lesion, site, relationship to the nipple, echotexture, echo pattern, homogeneity, shape, margin, boundary, orientation, posterior acoustic features on gray-scale US, and vascularity and grading on CDU. The lesion diameter (mm) was determined by the maximum diameter shown on US. The lesion site was the right or left breast. For the relationship to the nipple, we divided a mass into five parts of equal width (Fig 1). When the center of the mass (part A) was located under the center of a nipple, the mass was classified as concentric to the nipple. If it was somewhat off center, with part B under the nipple, it was classified as mildly eccentric to the nipple. If either part C or none of the mass was under a nipple, it was classified as markedly eccentric to the nipple.
The echotexture of each mass was characterized as cystic (cystic changes > 75% of a mass), solid (cystic changes < 25% of a mass), or solid and cyst (cystic changes 25-75% of a mass). Color flow distribution was defined as absent, vessels predominantly in the rim, or predominantly internal vascularity. Color flow grading of a mass on CDU was classified as low color flow signal or rich color flow signal. Low color flow signal was defined as the absence of color signals, the presence of occasional transient color pixels, or the presence of only a single color flow signal in or around the mass. Rich color flow signal was defined as multiple color pixels, multiple pedicles of blood supply, or the presence of multiple well-defined vessels in or around the mass. The image with the maximum color flow signals for each mass was chosen for grading.
Other US findings, including echo pattern, heterogeneity, shape, margin, boundary, orientation, and posterior acoustic features, were recorded according to BI-RADS sonographic lexicon classification. The echo pattern relative to fat was classified as anechoic, hypoechoic, isoechoic, hyperechoic, or complex. Heterogeneity was reported as predominantly homogeneous or heterogeneous. Shape was classified as oval, round, or irregular. The margin was defined as circumscribed or noncircumscribed. A circumscribed margin indicated a well- The relationship of a mass to a nipple. The mass (arrow) is divided into five parts of equal width. When part A is located below the center of a nipple, it is classified as concentric to the nipple, as in this example. When part B is below the nipple, it is classified as mildly eccentric to the nipple. If part C is below the nipple or the mass is in area D and not below the nipple at all, it is classified as markedly eccentric to the nipple.
https://doi.org/10.1371/journal.pone.0194651.g001 defined or sharp border with an abrupt transition between the lesion and surrounding tissue. The margin of a breast mass was classified as noncircumscribed if the margin had one or more the following features: an indistinct, angular, microlobulated, or spiculated margin. The mass boundary was defined as an abrupt interface or echogenic halo. Orientation was classified as parallel or nonparallel. Parallel orientation indicated that the long axis of the mass was parallel to the surface of the skin or the anterior-posterior diameter of the mass was shorter than the transverse diameter. Nonparallel orientation indicated that the long axis of the mass was perpendicular to the surface of skin or the anterior-posterior diameter of the mass was equal to or larger than the transverse diameter. The posterior acoustic appearance was categorized as normal, enhancement, or shadowing. Because microcalcifications on conventional US are not easily determined during the examination [12], microcalcifications were not recorded on the reevaluation of the images.

Histopathologic evaluation
Within 1 day to 5 months of the US examination, all 112 patients underwent a needle biopsy, excisional biopsy, or mastectomy, yielding a total 125 specimens for histopathological examination. This included tissue from one lesion in 100 patients, two lesions in 11 patients, and three lesions in 1 patient. One experienced pathologist (A-FL) with 25 years of experience and without knowledge of the preoperative US findings, retrospectively reviewed all the microscope slides from the 112 patients and confirmed the histopathological results for all 125 lesions.
A fourth radiologist (W-HY) integrated the clinical findings, histopathological results, and the US interpretations of all 125 masses to analyze the clinical and US characteristics, and their association with pathological results.

Statistical analysis
Data were analyzed with SPSS version 19.0 software (SPSS Inc., Chicago, IL). The level of significance was set at P < 0.05. An unpaired t test was used to compare continuous variables (age and maximum diameter), and a χ 2 or Fisher exact test for categorical variables. Statistically significant US variables were further assessed by multivariate logistic regression analysis, with calculation of odds ratios (ORs) and 95% confidence intervals (CIs). Table 1 summarizes the preoperative clinical findings of the 125 breast lesions in 112 male patients. The mean age of the 87 patients with breast benign masses was 56.68 ± 20.51 years (range 15-88 years), and that of 25 patients with breast malignancies was 73.16 ± 12.63 years (range 50-96 years) (P < 0.001). In addition to advanced age, bloody discharge from the nipple was also significantly associated with malignancy (P = 0.0387, Fisher's exact test). In contrast, bilateral synchronous masses and palpable tender lumps were more likely to be benign (both, 100% vs 0%, P = 0.0037 and P = 0.0019, respectively, Fisher's exact test). Of the 125 masses, 50 (40%) that were bilateral and/or palpable tender lumps were benign, and 2 (1.6%) with bloody discharge from the nipple were malignant. Table 2 summarizes the gray-scale US findings of all 125 breast lesions. The features of the masses on color Doppler ultrasonography are summarized in Table 3. Mildly or markedly eccentric to nipple (P = 0.044), round or irregular shape (P = 0.044), presence of an echogenic halo (P < 0.001), predominantly internal vascularity (P = 0.006), and rich color flow signal (P < 0.001) were significantly associated with malignancy (Figs 2-7). At least 1 of 5 statistically significant US features was present in each malignancy. In addition, the mean maximum diameter, echotexture, echo pattern, heterogeneity, margins, orientation, and posterior acoustic features were not helpful for distinguishing malignant (Figs 2-7) from benign masses (Figs 8-11).
Of the 12 patients with synchronous bilateral benign masses, 1 had bilateral pseudogynecomastia in the subareolar area, 1 had epidermal cysts in each breast, and the other 10 had bilateral gynecomastia, of whom 1 also had a benign palpable lymph node in the left breast near the axilla.
As shown in Table 1, being palpable was not different for benign versus malignant masses (79% vs 21%, P = 1.0, Fisher's exact test). The 3 nonpalpable malignant masses included: 1 ductal carcinoma in situ with bloody nipple discharge, 1 intraductal papilloma with ductal carcinoma in situ presenting with bloody nipple discharge, and 1 invasive ductal carcinoma presenting with left breast with tenderness. Furthermore, 1 palpable tender mass with serous nipple discharge was chronically inflamed tissue. A nonpalpable lipoma presented with serous discharge from the right nipple, while the other nonpalpable benign masses presented with focal breast swelling and/or tenderness on physical examination. As shown in Table 2, relative to the position of the nipple, of the 50 benign lesions with concentric features 38 (76%) were gynecomastia (Fig 8) and 12 (24%) were not. Of 16 benign conditions with mildly eccentric features, 10 (63%) were gynecomastia (Fig 9) and 6 (38%) were not. Of 34 conditions with markedly eccentric features, 5 (15%) were gynecomastia (Fig 10) and 29 (85%) were not (Fig 11). One cystic mass was a benign lymphangioma with an anechoic pattern. Of the 5 lesions with solid and cystic echotexture, 3 (60%) were benign masses including 2 abscesses and 1 area of chronic inflammation, and 2 (40%) were malignant and both were invasive ductal carcinomas ( Fig 5). Five masses with a complex echo pattern were nongynecomastia lesions with solid and cystic texture.

Discussion
Male breast cancer is much less common than gynecomastia and other benign breast pathologies [4], which makes it challenging for clinicians to differentiate malignancies from benign conditions in men with breast masses. Benign conditions of the male breast usually need only conservative therapy, but breast malignancies require timely surgical intervention. Clinical examination is the first step towards evaluating male breast masses. In our study, advanced age (mean age 73.16 vs. 56.68 years, P < 0.001) and bloody discharge from the nipple (P = 0.0387) were significant clinical features associated with malignancy. Synchronous bilateral lesions and palpable tender lumps were more likely to be benign than malignant. Of the 125 mases, 50 (40%) that were bilateral and/or palpable tender lumps were benign; 2 (1.6%) with bloody discharge from nipple were malignant. Among US features, being eccentric to a nipple, having an irregular shape, the presence of an echogenic halo, the presence of internal vascularity, and a rich color flow signal were all significantly associated with malignancy. Moreover, the presence of an echogenic halo and a rich color flow signal were independent predictors of breast malignancy. Six (4.8%) of 125 masses with anechoic cystic echotexture or complex solid and cyst features were all non-gynecomastia lesions. Clinical presentations in concert with US assessment may help the evaluation and management of male breast masses. Gynecomastia or non-gynecomastia benign masses usually  present as palpable tender lesions [1,4]. Synchronous, bilateral nodules in the subareolar areas are present in approximately half of patients with gynecomastia [7]. Our research results are compatible to those of prior studies that suggested US might not be necessary for bilateral and/ or palpable tender lumps (50/125, 40%) in male breasts (P < 0.05). Correcting underlying cause of gynecomastia and close follow-up might be a good strategy for managing bilateral and/or palpable tender lumps [4]. In converse, male breast cancer usually presents as an unilateral, palpable, and painless lump [1]. The incidence of synchronous bilateral breast cancers is in men reported to be only 1.5% to 2% of all male breast cancers [13]. Moreover, bloody nipple discharge or retraction and axillary lymphadenopathy might be present in the case of malignancy [1]. Bloody nipple discharge is found in 25% of men with breast cancers [5]. In our study, 2 of 2 masses presenting as bloody nipple discharge were malignant (P = 0.0387). Previous reports have shown that the mean age at diagnosis of male breast cancer is 67-69 years [6,14,15,16], which is younger than the mean age of our patients (73.16 years). Timely biopsy might be performed when the male breast masses (75/125, 60%) exhibit significantly malignant clinical and US features, such as advanced age, bloody discharge from the nipple, presence of an echogenic halo, and a rich blood flow signal. However, when conservative treatment for any mass is not effective, biopsy or excision should be performed [4].
The relationship between US features of malignant breast lesions and pathological characteristics of breast cancers have been reported [5,6,8,[17][18][19]]. An irregular shape and  noncircumscribed margins on US suggest tumor invasion into surrounding tissue. Nonparallel orientation may represent tumor spread through tissue-plane boundaries [20]. Posterior acoustic shadowing may indicate a desmoplastic reaction in breast cancers. More highly vascular lesions and intralesional color flow signals on CDU are compatible with increased angiogenesis in breast malignancies [18].
Gynecomastia in its early stage is characterized histologically by proliferation of intraductal epithelium, periductal inflammation, and surrounding edema, which clinically corresponds to a painful breast mass and usually appears on US as a hypoechoic oval nodule with a well circumscribed margin in the subareolar region [14]. However, late stage gynecomastia has dilated ducts surrounding stromal fibrosis, and usually presents as a hypoechoic lesion with an irregular, ill-defined, or macrolobulated margin [5,14]. In our study, some masses in cases of gynecomastia (Figs 8-10) had various projections, which could develop a nonparallel orientation, and noncircumscribed margins mimicking breast cancer. These particular US features, therefore, were not helpful in our series for differentiating malignant from benign conditions. The situation is different in women, where the 3 most common breast masses are simple cysts, fibrocystic changes, and fibroadenoma [21,22]. These usually present with typical benign US features with high predictive value: an oval shape, circumscribed margins, an abrupt interface, and a homogeneous echo texture.
Previous studies have demonstrated that heterogeneity and hypoechoic features are not useful in differentiating benign from malignant male breast lesions [2,6,23]. Hyperechogenicity is thought to be a reliable predictor of benignity, but there was a malignant mass in our series Fig 7. A palpable lump in the left breast of an 80-year-old man. (a) Radial sonogram shows an irregular shape, a noncircumscribed margin, a heterogeneously hypoechoic, subareolar, solid nodule (arrow). The nodule has a parallel orientation (arrow) and a concentric relationship to the nipple (N). There is an abrupt interface. (b) Color Doppler ultrasound shows multiple color pixels and pedicles of blood supply (arrows) pooling in the nodule, classified as a rich color flow signal with predominantly internal vascularity. (c) Pathology specimen shows an invasive papillary carcinoma with tissue invasion and fibrous reaction. There is no tumor necrosis. (Hematoxylin-eosin, original magnification ×100).
https://doi.org/10.1371/journal.pone.0194651.g007 with heterogeneous hyperechogenicity. Posterior acoustic enhancement has been reported to be associated with relative preservation of the transmitted ultrasound beam distal to a mass [24]. Fluid, mucin, hemorrhage, or necrosis usually results in the appearance of posterior acoustic enhancement [25,26]. Conversely, Maturen et al. considered that fluid within a mass was not required for posterior acoustic enhancement [24]. Any mass with a simple internal architecture might have posterior enhancement on US, which may be why in our study the appearance of posterior features did not differ significantly between malignant and benign lesions.
In terms of the relationship of male breast masses to the nipple, we found that lesions mildly or markedly eccentric from nipple were significantly associated with cancer; in contrast, masses concentric to the nipple were significantly more likely to be benign (P = 0.044). Prior studies have also suggested that gynecomastia usually appears as a concentric subareolar mass, while other benign masses and cancers in men are usually eccentric to the nipple-areolar complex [14,27]. However, eccentricity to the nipple was present in some cases of gynecomastia in our study, as well as with a number of other benign lesions, so the relationship of the lesion to the nipple was not a statistically significant independent predictor of malignancy.  Chen at el. recommended that a mass in the male breast with a cystic component on US should be considered suspicious for malignancy [6]. Chau at al. reported that invasive or in situ papillary cancers in the male breast often had complex mixed cystic and solid composition on US [15]. In our study, 2 (40%) of 5 male breast lesions with cystic and solid components were invasive ductal carcinomas. The only invasive papillary carcinoma in our series did not have mixed components. A small number of benign conditions (3/100, 3%), 1 case of chronic inflammation and 2 abscesses also had cystic and solid US features. Our findings thus suggest that mixed components on US are not a feature specific for male breast cancers, and are not present in gynecomastia.
The histopathological findings associated with an echogenic halo indicate the invasion of malignant cells into fat tissue, where they exist along with adipocytes and elastic fibers [19]. Abscesses can also display an echogenic halo, which may be related to an inflammatory reaction and edema seen on histopathological examination. An echogenic halo is considered to be of no diagnostic significance for breast malignancy in the 2013 BI-RADS Atlas (5 th edition).  However, our findings suggest that an echogenic halo was not only a significant feature, but also an independent US predictor for male breast malignancy. In our study, 8 (47%) of 17 masses were benign lesions, including 2 cases of gynecomastia, and had an echogenic halo. Conversely, there was no obvious echogenic halo around 2 abscesses in the current series. Some studies [28,29] have found that vascular characteristics were the best diagnostic determinant in women to distinguish malignant from benign breast lesions. Benign breast lesions either had no vascularization, or had vessels around the margin of the lesions, while breast cancers had multiple vessels penetrating into, or blood pooling in the lesions [28][29][30]. However, our study showed that absent vascularization, predominantly in the rim or predominantly internal vascularity, could be present in benign male breast masses. The feature of predominantly internal vascularity was significantly more common in malignancies than in benign masses, but on multivariate analysis it was still not an independent predictor of cancer. On the other hand, a rich color flow signal was an independent predictor of malignancy (OR = 8.330, CI: 1.505-46.102, P < 0.001).
This retrospective study has some bias. Ultrasound is a dynamic study, and very much operative dependent. Information bias is introduced in our study when using the static US images saved by technicians that performed the examination. Selection bias occurred when the question selecting patients with the preoperative ultrasonic examination and the pathological results thus missing many examinations. However, it is very difficult to accumulate enough cases of a rare condition like male breast cancer. This study also has several limitations. First, it is a retrospective study with a relatively small number of patients. Only biopsy-proven diagnoses of masses in patients who had had preoperative US were included. Had cases without US had been included, the findings might have differed somewhat. Second, because of the retrospective nature data recorded in the charts may have been incomplete. For this reason, we did not evaluate other risk factors for male breast cancer including prior radiation therapy of the chest, smoking and alcohol history, liver disease, hyperestrogenism, androgen deficiency, undescended testes, orchitis, congenital inguinal herniation, Klinefelter's syndrome, or family history of breast cancer [6,27,31]. Third, evaluation of microcalcifications and axillary lymph nodes were not included in this study. Fourth, we did not compare US to mammography. Last, because of incomplete data a detailed correlation between the US findings and clinical presentation of benign and malignant lesions, including onset, duration, skin changes, and other associated symptoms and signs could not be investigated.

Conclusions
In conclusion, clinicians might use an integration of clinical and US findings to evaluate and manage a male breast lesion. Advanced age, bloody discharge from the nipple, a mass eccentric to a nipple, irregular shape, the presence of an echogenic halo, predominantly internal vascularity, and rich color flow signal on CDU were features more common in cancers than in benign lesions. The presence of an echogenic halo and rich color flow signal were independent US predictors of male breast cancers on multivariate analysis. The abovementioned specific clinical and US features of male breast cancers suggest that a biopsy or surgical excision be performed. Conversely, synchronous bilateral subareolar and/or palpable tender masses were almost always benign, and conservative follow-up may be appropriate. However, when conservative management is not effective, surgery should be considered.