- Email: [email protected]
Metastases to the Breast from Non-mammary Malignancies
Metastases to the Breast from Non-mammary Malignancies: Primary Tumors, Prevalence, Clinical Signs, and Radiological Features €rgen Holzhausen, MD, Kathrin Ruschke, MD, Alexey Surov, MD, Eckhard Fiedler, MD, Hans-Ju Hans-Joachim Schmoll, MD, Rolf-Peter Spielmann, MD Rationale and Objectives: Most secondary intramammary tumors occur as metastatic involvement from the contralateral breast. Breast metastases (BM) from nonmammary malignancies are very rare. The aims of this study were to estimate retrospectively the prevalence of BM from nonmammary malignancies and to describe their radiologic appearance. Materials and Methods: BM were identified in 51 patients, including 43 women and eight men with a median age of 61 years (range, 24–84 years). Computed tomography of the thoracic region identified 108 lesions in 38 patients. Mammography was available for 37 patients (54 lesions). Ultrasound evaluation was performed in 43 patients (71 lesions). In 24 patients (93 lesions), magnetic resonance imaging of the breast was done. Images were reviewed in consensus by two radiologists according to the Breast Imaging Reporting and Data System lexicon. Results: The prevalence of BM in several tumors ranged from 0.12% to 4.92%. On computed tomography, most metastases were round or oval in shape with marked or moderate enhancement. On mammography, solitary or multiple round or oval masses with circumscribed margins were the most common pattern of BM. Ten percent showed microcalcifications. On ultrasound, most BM were hypoechoic, oval or round in shape, with microlobulated or circumscribed margins, and posterior acoustic enhancement. Doppler imaging showed hypervascularity in 39% of BM. On magnetic resonance imaging, most lesions demonstrated marked homogenous contrast enhancement. Type 1 kinetic curve was seen in 18%, type 2 in 52%, and type 3 in 30%. Conclusions: The radiologic features reported in this study should be taken into consideration in the differential diagnosis of breast lesions. Key Words: Breast; metastases; extramammary malignancies; radiologic features. ªAUR, 2011
M
ost secondary intramammary tumors occur as metastatic involvement from the contralateral breast (1–3). Breast metastases (BM) from nonmammary malignancies are very rare (4–6). This may related to the fact that the breast contains large areas of fibrous tissue with a relatively poor blood supply (1,2). Alva and Shetty-Alva (7) found in their review of the literature that most BM arise from malignant melanomas, sarcomas, lung cancers, ovarian tumors, renal carcinomas, and thyroid tumors in decreased order of frequency (7). Some single-institution studies have suggested, however, that BM from ovarian and renal carcinomas are very rare (8). The reported data regarding the prevalence of BM in nonmammary malignancies are also controversial. Abrams et al (9) detected BM in 5% of 1,000 autopsied cases of carcinoma. Hejmadi et al (10), however, reported a prevalence of BM of 0.11%.
Most reports regarding BM are clinical observations or pathologic series and provide either no or inconsistent radiologic information (11–17). Radiologic publications consist predominantly of isolated case reports or small series (18–23). Furthermore, most radiologic studies were conducted from 1970 to 1990, before the Breast Imaging Reporting and Data System (BI-RADS) classification had been introduced (18–22). In addition, ultrasound and magnetic resonance imaging (MRI) of BM have been reported only sporadically in the literature (2,23). Therefore, the purpose of this study was to estimate the prevalence of BM and to describe their radiologic appearance.
MATERIALS AND METHODS This retrospective study was approved by the institutional ethics committee.
Acad Radiol 2011; 18:565–574 From the Departments of Radiology (A.S., K.R., R.-P.S.), Dermatology (E.F.), Pathology (H.-J.H.), and Oncology (H.-J.S.), Martin-Luther-University HalleWittenberg, Ernst-Grube-Straße 40, 06097 Halle, Germany. Received October 17, 2010; accepted December 14, 2010. Address correspondence to: A.S. e-mail: [email protected] ªAUR, 2011 doi:10.1016/j.acra.2010.12.009
Patients
A retrospective search in the statistical database of our institution from January 2000 to December 2009 revealed 6,668 patients with metastasizing solid malignancies who were examined and treated at our institution. In addition, 2,896 565
SUROV ET AL
patients with breast carcinoma were treated at our breast center in the investigated time period. Furthermore, we retrospectively reviewed the radiologic database for metastatic breast involvement in oncologic diseases (through a computerized search for key words in the reports). BM were identified in 51 patients, including 43 women and eight men with a median age of 61 years (range, 24–84 years). In all of the identified patients, the diagnosis of metastatic disease was proven histopathologically. In 36 patients (71%), the diagnosis was confirmed on breast biopsy specimens. In the other 15 patients (29%) with multiple metastases, soft tissue or cutaneous lesions of another localization had been investigated pathologically. All radiologic images of the identified patients were reevaluated. Patients with direct invasion of tumors into the breast from the thoracic wall were excluded from the study. Furthermore, patients with primary breast sarcoma, primary or secondary breast lymphoma, and breast involvement in hematologic oncologic diseases were also not included in the study. Additionally, we validated the radiologically identified BM cases by comparison with the database of the department of pathology of our hospital. Imaging and Review
Computed Tomography. Computed tomography (Somatom Plus 4 VZ and Somatom Sensation 64; Siemens Medical Systems, Erlangen, Germany) of the thoracic region was performed as a staging investigation in 38 patients. In all cases, 1.5 mL of iodinated intravenous contrast medium (Solutrast; Bracco Altana GmbH, Konstanz, Germany) per body mass was given at a rate of 1.5 to 3.5 mL/s by a power injector (Medtron GmbH, Germany), with a scan delay of 60 seconds after the onset of injection. Typical imaging parameters were 120 kVp, 150 to 300 mAs, and a 0.6-mm to 6-mm slice thickness, with a pitch of 0.8 to 1.2. Mammography. Mammography was performed in 37 patients using dedicated mammographic equipment (Mammomat 3000 Nova and Mammomat Novation DR; Siemens Medical Systems). Standard mediolateral oblique and craniocaudal mammograms were obtained, with additional views if necessary. Breast findings were classified according to the American College of Radiology’s BI-RADS lexicon (24).
Academic Radiology, Vol 18, No 5, May 2011
TABLE 1. Prevalence of BM in Metastasizing Oncologic Diseases Malignant Disease
Prevalence of BM (%)
Malignant melanoma of skin Carcinoma of ovary Carcinoma of stomach Renal cell carcinoma Sarcoma Carcinoma of bronchus and lung Carcinoma of larynx Carcinoma of cervix uteri Seminoma of testis Carcinoma of prostate Carcinoma of thyroid gland Carcinoma of colon
4.92 4.43 1.47 1.17 1.13 0.89 0.77 0.76 0.52 0.24 0.22 0.12
BM, breast metastases.
thickness using a T1-weighted (T1W) three-dimensional fast low-angle shot sequence (repetition time, 14 ms; echo time, 7 ms; flip angle, 25 ). Measurement time was 87 seconds, once before and five times after intravenous bolus injection of 0.2 mmol gadolinium diethylenetriamine penta-acetic acid per kilogram body weight (Schering AG, Berlin, Germany). The flow rate was 2 mL/s, and the bolus was followed by a 20-mL saline flush. Central k-space data were acquired at 43, 130, 217, 304, and 391 seconds after injection. Automatic image subtraction was performed routinely. T2-weighted images included a turbo spin-echo sequence (repetition time, 4,200 ms; echo time, 90 ms; flip angle, 180 ) with a 252 256 matrix, a 350-mm field of view, and a slice thickness of 5 mm. Kinetic analysis of contrast enhancement was performed additionally. Time–signal intensity curves were drawn using operator-defined regions of interest. The region of interest was smaller than the lesion size. The time to peak enhancement (ie, the time between administration of contrast agent and the maximum signal intensity value in the postcontrast phase) was estimated. Furthermore, the initial signal increase (SI) from the precontrast value (SIp) to the maximum peak within the first 3 minutes after the administration of contrast medium (SI1–3 min) was calculated (26–29): initial SI ð%Þ ¼ ðSI13min SIP Þ=SIP 100%:
Ultrasound. Ultrasound evaluation was performed in 43 patients using a 10.5-MHz linear-array transducer (Sonoline Elegra and Acuson Antares; Siemens Medical Systems). Additionally, power and color Doppler ultrasound scanning was performed in 20 patients. Sonographic findings were classified according to the BI-RADS ultrasound lexicon (25).
The postinitial behavior of the signal curve from the maximum peak (SIpeak) to the end of the examination (SIend) was also analyzed (26–29):
MRI. In 24 patients, MRI of the breast was performed additionally using a dedicated receive-only breast coil on a 1.0-T scanner (Impact Expert; Siemens Medical Systems). Both breasts were imaged in the coronal plane with 2.5-mm slice
MRI findings were classified according to the American College of Radiology’s BI-RADS lexicon (30,31). All available images were interpreted by two radiologists (A.S. and K.R., with 7 and 11 years of experience,
566
postinitial SI ð%Þ ¼ SIpeak SIend
SIpeak 100%:
Academic Radiology, Vol 18, No 5, May 2011
METASTASES TO THE BREAST
Figure 1. Imaging findings in a 48-year old woman with a known history of bronchial carcinoma. (a) Staging computed tomographic examination detecting an oval mass with moderate homogenous enhancement within the left breast (arrow). (b) Bilateral mediolateral oblique mammograms show multiple round and oval dense lesions with circumscribed and microlobulated margins (arrows). (c) Ultrasound of the left breast documents an oval mixed isoechoic to hyperechoic mass with small boundary posterior shadowing. (d) Transverse power Doppler ultrasound scan shows no hypervascularity of the mass. (e) On T1W images before contrast administration, the mass is hypointense to isointense compared to breast tissue (arrow). (f) Subtracted images after intravenous administration of contrast medium demonstrated marked homogenous enhancement of the mass (arrow). Time-intensity curve of the lesion documenting rapid enhancement in the first minute associated with wash out (type 3 curve). Time to peak enhancement was 90 seconds, initial increase of signal intensity was 132%, and washout ratio was 22%. (g) Histologic analysis after ultrasound-guided biopsy revealed a metastasis of bronchial carcinoma (hematoxylin and eosin stain, 100).
respectively). Consensus of the investigators was obtained on the following features of the identified breast lesions: number, shape, localization, size, margin, attenuation, homogeneity, and contrast enhancement patterns. Lesion size was determined by measuring the maximum diameter. Furthermore, case histories were reviewed retrospectively to determine clinical signs at presentation. Histologic Analysis
Biopsy specimens were evaluated by one pathologist (H.-J.H., with 37 years of experience). All available histologic materials (sections stained by hematoxylin and eosin and by immunohistochemistry) were analyzed.
Statistical Analysis
For statistical analysis, SPSS (SPSS, Inc, Chicago, IL) was used. Collected data were evaluated by means of descriptive statistics (absolute and relative frequencies). Continuous variables are expressed as mean standard deviation and categorical variables as percentages. RESULTS Primary Tumors and Number and Prevalence of BM
One hundred eighty-four metastases were diagnosed in the breast in 51 of 6,668 oncology patients (0.76%) who met 567
SUROV ET AL
Academic Radiology, Vol 18, No 5, May 2011
Figure 2. Radiologic findings in a 67-year-old woman with a known history of small cell lung cancer. (a) Staging computed tomographic examination detecting a round mass with moderate inhomogenous enhancement within the right breast (arrow). Additionally, right pulmonary masses are seen. (b) On mammography, the mass was round with microlobulated margins (arrow). (c) Transverse ultrasound documents irregular lesions with microlobulated margins and lightly posterior acoustic enhancement. (d) Transverse power Doppler shows marked internal hypervascularity. Histologic examination of the mass revealed a metastasis of the known bronchial carcinoma (not shown).
the inclusion criteria. The number of BM varied from one to 26. In 24 patients, BM were localized in the right breast (47 lesions [25.5%]) and in 13 patients (28 lesions [15.2%]) in the left breast, and in 14 patients (109 lesions [59.3%]), BM were bilateral. The size of metastases varied from 2 to 60 mm (median size, 8 mm). Systemic metastases outside the breast were diagnosed in 40 these 51 patients (78.4%). In 11 patients (21.6%), BM were the only manifestation of metastatic disease. Furthermore, in three of the 11 patients, BM were the first sign of malignancy. The intramammary metastases were derived from the following malignancies: malignant melanoma (31%), genital tumors (18%), pulmonary tumors (16%), urologic tumors (12%), gastrointestinal tumors (7%), and sarcomas (4%). Other neoplasia were rare. In 4% of the cases, no primary tumor could be identified. The intramammary lesions showed here no typical immunohistochemical patterns for primary breast cancer. The prevalence of detected BM associated with the different tumors is given in Table 1. Clinical Findings
Clinically, 16 patients (31%) presented with solitary or multiple painless breast lumps. In five patients (10%), unilateral axillary lymphadenopathy occurred additionally. Skin edema was seen in two patients (3.9%). In 30 patients (59%), however, breast involvement was diagnosed incidentally during staging examination on computed tomography. Furthermore, in five patients with primary unknown malignancies (10%), BM were found during a thorough tumor search on mammography or breast ultrasound. Radiologic Features
Computed Tomography. On computed tomography, which was available for 38 patients (108 lesions), most metastases 568
were round or oval in shape with circumscribed margins (Figs 1–3). They showed a marked enhancement in 52% and moderate and light enhancement in 37% and 11%, respectively (Table 2). Mammography. Mammography was performed in 37 patients. Mammographic findings are summarized in Table 3. Solitary or multiple breast masses (Figs 1–3) were the most common pattern of BM (27 patients [73%]). In two patients (5%), architectural distortion (diffuse unilateral or bilateral breast opacity) was seen. In eight patients (22%), however, no abnormalities could be identified on mammography. Twenty percent of the mammographic findings were interpreted as BI-RADS category 0 (American College of Radiology 4), 14% as BI-RADS category 3, 37% as BI-RADS category 4, and 29% as BI-RADS category 5. Ultrasound. Ultrasound images were available for 43 patients (71 lesions). The ultrasound features are shown in Table 4. Most of the intramammary masses were hypoechoic, oval or round in shape, with microlobulated or circumscribed margins and posterior acoustic enhancement (Figs 1–3). Color and power Doppler imaging showed hypervascularity in 16 of the identified lesions (39%; Table 5). US findings were categorized as BI RADS 2 in 1 (2%), BI RADS 3 in 8 (19%), BI RADS 4 in 29 (67%) and BI RADS 5 in 5 (12%) cases. MRI. MRI was performed in 24 patients (93 lesions). On T2weighted images, most lesions were isointense in comparison to the normal breast parenchyma (Table 6). On T1W images, most BM were isointense to the normal breast tissue and demonstrated marked homogenous contrast enhancement after intravenous administration of contrast medium (Figs 1, 3, and 4). Kinetic analysis of contrast enhancement was performed for 62 lesions. It showed in most cases (89%) a rapid initial SI >100% in comparison with precontrast signal intensity. The median initial
Academic Radiology, Vol 18, No 5, May 2011
METASTASES TO THE BREAST
Figure 3. Imaging findings in a 46-year old woman with a known history of medullary thyroid carcinoma. (a) Staging CT examination detecting an oval mass with circumscribed margins and marked enhancement within the left breast (arrow). On mammography (not shown), there was no evidence of malignancy (American College of Radiology 4). (b) Ultrasound documents a round hypoechoic to hyperechoic mass with indistinct margins without posterior acoustic phenomena. (c) Transverse power Doppler shows no hypervascularity of the lesion. (d) On T1W images before contrast administration, the mass is hypointense to isointense compared to breast tissue (arrow). (e) Subtracted images after intravenous administration of contrast medium demonstrated marked homogenous enhancement of the lesion (arrow). Time-intensity curve of the lesion showing rapid enhancement in the first minute associated with washout (type 3 curve). Time to peak enhancement was 90 seconds, initial increase of signal intensity was 506%, and washout ratio was 14.2%. (f) Histologic examination after ultrasound-guided biopsy revealed metastases of medullary thyroid carcinoma (hematoxylin and eosin stain, 50).
SI was 167% (range, 20%–672%). The median time to peak enhancement was 180 seconds (range, 90–270 seconds).
MRI findings were categorized as BI-RADS category 4 in 11 patients (46%) and BI-RADS category 5 in 13 patients (54%). 569
SUROV ET AL
Academic Radiology, Vol 18, No 5, May 2011
TABLE 2. Computed Tomographic Findings of Breast Metastases
TABLE 4. Ultrasound Findings of Breast Metastases Characteristic
Characteristics Lesions Shape Oval Round Lobulated Margin Circumscribed Lobulated Enhancement High Moderate Low
n (%) 108 (100) 18 (17) 82 (76) 8 (7) 101 (94) 7 (6) 56 (52) 40 (37) 12 (11)
TABLE 3. Mammographic Findings in 37 Patients with Breast Metastases Characteristic Pattern Mass Architectural distortion No abnormalities Mass features (n = 52) Shape Round Oval Lobular Margin Circumscribed Microlobulated Density High Isodense Calcifications Yes No ACR 1 2 3 4 BI-RADS 0 3 4 5
Patients 27 (73%) 2 (5%) 8 (22%) Lesions
41 (95%) 2 (5%) Lesions 30 (44%) 28 (41%) 10 (15%) 34 (50%) 18 (26%) 16 (24%) 13 (19%) 35 (52%) 20 (29%) 52 (76%) 11 (16%) 4 (6%) 1 (2%)
24 (46%) 28 (54%)
needle was performed in 26 patients. In two patients, stereotactic biopsy (11-gauge needle) was performed. In eight patients, incisional breast biopsy was performed. In the other 15 patients (29%) with multiple metastases, soft tissue or cutaneous lesions of another localization had been investigated pathologically. In 11 patients (22%), in whom BM were the only manifestation of metastatic disease, metastases were resected surgically. Patients with systemic metastases were treated with chemotherapy. Thirty patients (59%) died during therapy within 1 year after the diagnosis of BM. The remaining 10 patients are alive to date under chemotherapy.
5 (10%) 47 (90%)
DISCUSSION
37 (71%) 9 (17%) 6 (12%) 46 (88%) 6 (12%)
31% 31% 12% 26% 20% 14% 37% 29%
ACR, American College of Radiology; BI-RADS, Breast Imaging Reporting and Data System.
Diagnosis and Follow-Up
In all of the identified cases, the diagnosis of metastatic disease was proven histopathologically. In 36 patients (71%), the diagnosis was confirmed on breast biopsy specimens. Ultrasound-guided core-needle biopsy using a 14-gauge 570
Pattern Mass Architectural distortion (diffuse infiltration) Mass features (n = 68) Shape Round Oval Lobular Margin Microlobulated Circumscribed Indistinct Echo pattern Anechoic Hypoechoic Mixed hypoechoic to hyperechoic Posterior phenomena Enhancement None Shadowing Mixed
Patients
The prevalence of BM varies significantly in several studies, depending on how the prevalence is calculated. In relation to all breast biopsies performed in several single-institution studies, it ranged from 0.11% to 1.93% (10,32). Furthermore, it was 0.33% to 6.3% of cases with identified breast malignancies (8,33). The frequency of metastatic involvement of the breast in several tumors has been reported in only one series previously (34). Here, it ranged from 0.019% in hepatocellular carcinoma to 0.621% in thymus carcinoma (34). The investigators estimated the prevalence for each tumor with BM. However, it was calculated related to the total number of malignancies, regardless whether they had metastasized or not. In our study, the prevalence of BM was generated in relation to metastasizing tumors only and ranged from 0.12% to 4.92% (Table 1). Furthermore, the frequency of BM from all breast malignancies was 1.76% in our study. According to previous reports, BM are more common solitary than bilateral or multiple lesions (1–5,12–16). In the
Academic Radiology, Vol 18, No 5, May 2011
METASTASES TO THE BREAST
TABLE 5. Doppler Findings of 42 Breast Metastases in 20 Patients Hypervascularity None Light Moderate High
n (%) 25 (61) 10 (24) 2 (5) 4 (10)
TABLE 6. Magnetic Resonance Findings of Breast Metastases Characteristics
Figure 4. Magnetic resonance imaging of the breasts in a 28-yearold woman with a known history of ovarian cancer and right breast lumps. Mammography (not shown) demonstrated an architectural distortion (diffuse dense opacity) of the right breast. (a) Subtracted images after intravenous administration of contrast medium showing diffuse infiltration of the right breast with marked inhomogenous enhancement (arrow). (b) Transverse reconstruction of the breasts (subtracted images) reveals also diffuse infiltration of the right breast (arrow). Time-intensity curves document continuous increase in signal intensity on each successive contrast-enhanced image >10% in comparison with initial enhancement (type 1). (c) Histologic analysis after magnetic resonance–guided biopsy showing diffuse metastatic infiltration of the breast (hematoxylin and eosin stain, 100).
present study, however, 59.3% of BM were bilateral. Furthermore, most BM were multiple. Clinically, most intramammary metastases have been reported to be rapidly growing masses without pain (5,12– 16). Retraction of the skin or nipple was typically not seen (1,13–15). In our study, in 59% of patients, breast involvement was asymptomatic and diagnosed incidentally during staging examination. Only 31% of the patients presented clinically with solitary or multiple painless breast lumps with or without unilateral axillary lymphadenopathy. Radiologic series have described relatively small numbers of patients, ranging from six to 33 (18–22,35). Furthermore, patients were often investigated only by a single
Pattern Mass Diffuse infiltration Mass features (n = 91) Shape Round Oval Lobular Margin Smooth Irregular Signal intensity (T1W images) Hypointense Isointense Mixed isointense to hyperintense Signal intensity (T2W images) Isointense Hyperintense Internal mass enhancement Homogenous Inhomogenous Rim enhancement Time–signal intensity curve pattern Type 1 Type 2 Type 3
Patients 22 (92%) 2 (8%) Lesions 54 (59%) 32 (35%) 5 (6%) 84 (92%) 7 (8%) 10 (11%) 79 (87%) 2 (2%) 76 (84%) 15 (16%) 73 (80%) 7 (8%) 11 (12%) 11 (18%) 32 (52%) 19 (30%)
examination. Therefore, our study, with 184 findings in 51 patients, is the largest to date (Table 7). Computed tomographic appearances of BM have been described previously in rare cases (35). According to the present study, most metastases were round or oval in shape, with circumscribed margins and high or moderate enhancement. This finding is unspecific and could equally represent a cyst, fibroadenoma, or breast carcinoma. Therefore, further radiologic investigations are needed to determine these lesions. Mammographic appearances of BM have been reported as well known (1,5,15,16,18–22,35–37). According to previous reports, the most common mammographic appearances are circumscribed masses, located in the upper outer quadrants without spiculation (5,18–22,35–37). However, other findings are also documented. For example, in the largest mammographic study (n = 18), by Noguera et al (35), the 571
SUROV ET AL
Academic Radiology, Vol 18, No 5, May 2011
TABLE 7. Comparison of Reported Studies Regarding Radiologic Findings of Breast Metastases Radiologic Investigations (Patients) Study Bohman et al (18)* McCrea et al (19)* Derchi et al (20) Hebert et al (21) Vergier et al (5) Muttarak et al (36) Yeh et al (2) Noguera et al (35)* Present study
Patients
CT
Mammography
Ultrasound
Doppler
MRI
11 8 6 21 8 7 15 22 51
— — — — — — — 4 38
11 8 — 5 8 7 — 18 37
— — 6 — 8 6 11 11 43
— — — — — — — — 20
— — — — — — — 1 24
CT, computed tomography; MRI, magnetic resonance imaging. *Cases of lymphoma, leukemia, and intramammary metastases from contralateral breast cancer were excluded from the analysis.
findings of three BM cases (17%) resembled those of inflammatory breast carcinoma. In addition, asymmetric density occurred in one case (6%). Furthermore, the identified masses showed circumscribed margins in 58% only, and in 42%, indistinct margins were seen (35). According to the literature, some metastases present microcalcifications (1,5,18–22). Typically, they occur in mucinous ovarian carcinoma (18–22). However, metastases of hepatocellular carcinoma, gastric cancer, and medullary thyroid carcinoma can also develop calcifications (34,35,38). In our series we analyzed 37 mammograms (56 findings). Therefore, this is the largest study to date. In 5% of the cases, BM manifested as diffuse unilateral or bilateral breast opacity. Most BM, which presented as breast masses, were round or oval in shape, with circumscribed margins. Lobular lesions occurred in 12%. Calcifications were seen in 10%. There were cases with metastatic ovarian cancer. Ultrasound findings of BM have been described in small series ranging from six to 11 cases (2,5,20,35,36,39). According to the literature, intramammary metastases were hypoechoic or isoechoic in comparison to the breast parenchyma, with circumscribed margins (2,20,35). In addition, most showed no posterior acoustic phenomena. In a study by Yeh et al (2), however, five of 11 lesions had irregular margins. In our study, ultrasound investigation was performed in 43 patients (71 lesions). In contrast to the previous reports, several patterns could be identified here. Fifteen percent of intramammary masses were lobular, 41% oval, and 44% round. Twenty-six percent of lesions had circumscribed margins, 50% were microlobulated, and 24% had indistinct margins. Most lesions were homogenously hypoechoic with posterior acoustic enhancement. Interestingly, many BM were anechoic and demonstrated posterior acoustic enhancement. This pattern is typical of breast cysts, and these lesions can be misdiagnosed. Therefore, further investigations, such as Doppler imaging, are needed here. Doppler imaging findings in BM have been reported in only eight lesions in the literature (40–47). In most cases, metastatic lesions were hypervascular (40,42,44–47). The published 572
metastases arise from sarcomas, carcinoid tumors, melanoma, and renal cell carcinoma (ie, from hypervascularized tumors). Our study, however, showed that in 61% of BM, no hypervascularity was seen. Light vascularity was documented in 24%, moderate in 5%, and high in 10%. MRI findings of BM have been reported in only 12 patients previously (23,35,43,48–54). There were isolated case reports describing MRI findings in several malignancies. Furthermore, in most reports, no information of enhancement kinetics was provided. In our series, 24 patients with 93 lesions were analyzed. On T2-weighted images, most lesions were isointense in comparison to the normal breast parenchyma and hypointense compared to fat. On T1W images, most BM (87%) were isointense or lightly hypointense compared to the normal breast tissue. However, 2% of the metastases were mixed isointense to hyperintense. High signal intensity of metastases on T1W images was observed in our study in metastases from melanoma. Metastases of malignant melanoma contain melanin, which shows a hyperintense signal on T1W images (55). On the other hand, melanoma metastases often bleed (56). However, SI on T1W images was seen in only two of 10 patients with melanoma, who were investigated using MRI. After intravenous contrast administration, several patterns of BM were identified in our series. The diagnosed metastases presented as solitary or multiple lesions, round or oval in shape with smooth margins and marked homogenous contrast enhancement. Inhomogenous and rim enhancement were rarely present. Furthermore, in 8%, diffuse infiltration of breast parenchyma was identified. As seen, MRI appearances of BM may mimic benign breast masses, such as fibroadenoma. Therefore, it is important to use further analysis of breast MRI findings, particularly kinetic analysis. In our series, kinetic analysis of contrast enhancement showed in most cases a rapid initial SI >100% in comparison to precontrast signal intensity. This pattern is typical for breast malignancies (26,29). However, in 11% of the lesions, initial SI was <100%, ranging from 20% to 92%.
Academic Radiology, Vol 18, No 5, May 2011
Type 2 and 3 kinetics were present in most cases (82%). According to the literature, both types of curves suggest malignancy (26,29). In 18% of BM, type 1 kinetic curve was seen. This pattern is usually associated with benign findings (26,29). As seen, BM presented with different radiologic appearances and may be misdiagnosed as primary breast carcinoma or as benign lesions. Therefore, the diagnosis of BM should be confirmed histopathologically. Some reports have suggested that fine-needle biopsy can definitively confirm the diagnosis of BM (10,11,17). However, several metastatic malignancies, such as endometrial or prostatic cancers, and some variants of malignant melanomas have cytologic features resembling mammary carcinoma (33). In our study, core needle and incision biopsy was used in all cases. There were several limitations to our study. First, this was a retrospective analysis. Second, not each patient underwent all investigations (mammography, ultrasound, Doppler, and MRI). Third, in only 71% of the patients was the diagnosis established by breast biopsy. In the other 29% of patients, soft tissue or cutaneous lesions of another localization had been investigated pathologically. CONCLUSIONS BM manifest most frequently as round or oval masses with circumscribed margins on mammography and as hypoechoic masses with microlobulated or circumscribed margins and posterior acoustic enhancement on ultrasound. On MRI, most BM present as circumscribed masses with marked or moderate homogenous enhancement. These radiologic features should be taken into consideration in the differential diagnosis of breast lesions. REFERENCES 1. Bartella L, Kaye J, Perry NM, et al. Metastases to the breast revisited: radiological-histopathological correlation. Clin Radiol 2003; 58:524–531. 2. Yeh CN, Lin CH, Chen MF. Clinical and ultrasonographic characteristics of breast metastases from extramammary malignancies. Am Surg 2004; 70: 287–290. 3. Hajdu SI, Urban JA. Cancers metastatic to the breast. Cancer 1972; 29: 1691–1696. 4. Yang WT, Muttarak M, Ho LW. Nonmammary malignancies of the breast: ultrasound, CT, and MRI. Semin Ultrasound CT MRI 2000; 21:375–394. 5. Vergier B, Trojani M, de Mascarel I, et al. Metastases to the breast: differential diagnosis from primary breast carcinoma. J Surg Oncol 1991; 48: 112–116. 6. Harvey JA. Unusual breast cancers: useful clues to expanding the differential diagnosis. Radiology 2007; 242:683–694. 7. Alva S, Shetty-Alva N. An update of tumor metastasis to the breast. Arch Surg 1999; 134:450. 8. Wood B, Sterrett G, Frost F, et al. Diagnosis of extramammary malignancy metastatic to the breast by fine needle biopsy. Pathology 2008; 40: 345–351. 9. Abrams HL, Spiro R, Goldstein N. Metastases in carcinoma. Analysis of 1000 autopsied cases. Cancer 1950; 3:74–85. 10. Hejmadi RK, Day LJ, Young JA. Extramammary metastatic neoplasms in the breast: a cytomorphological study of 11 cases. Cytopathology 2003; 14:191–194.
METASTASES TO THE BREAST
11. Silverman JF, Feldman PS, Covell JL, et al. Fine needle aspiration cytology of neoplasms metastatic to the breast. Acta Cytol 1987; 31:291–300. 12. Georgiannos SN, Chin J, Goode AW, et al. Secondary neoplasms of the breast: a survey of the 20th Century. Cancer 2001; 92:2259–2266. 13. Amichetti M, Perani B, Boi S. Metastases to the breast from extramammary malignancies. Oncology 1990; 47:257–260. 14. Vaughan A, Dietz JR, Moley JF, et al. Metastatic disease to the breast: the Washington University experience. World J Surg Oncol 2007; 5:74. 15. Smyrniotis V, Theodosopoulos T, Marinis A, et al. Metastatic disease in the breast from nonmammary neoplasms. Eur J Gynaecol Oncol 2005; 26: 547–550. 16. Williams SA, Ehlers RA II, Hunt KK, et al. Metastases to the breast from nonbreast solid neoplasms: presentation and determinants of survival. Cancer 2007; 110:731–737. 17. David O, Gattuso P, Razan W, et al. Unusual cases of metastases to the breast. A report of 17 cases diagnosed by fine needle aspiration. Acta Cytol 2002; 46:377–385. 18. Bohman LG, Bassett LW, Gold RH, et al. Breast metastases from extramammary malignancies. Radiology 1982; 144:309–312. 19. McCrea ES, Johnston C, Haney PJ. Metastases to the breast. AJR Am J Roentgenol 1983; 141:685–690. 20. Derchi LE, Rizzatto G, Giuseppetti GM, et al. Metastatic tumors in the breast: sonographic findings. J Ultrasound Med 1985; 4:69–74. 21. Hebert G, Ouimet-Oliva D, Paquin F, et al. Diffuse metastatic involvement of the breast. Can Assoc Radiol J 1991; 42:353–356. 22. Toombs BD, Kalisher L. Metastatic disease to the breast: clinical, pathologic, and radiographic features. AJR Am J Roentgenol 1977; 129: 673–676. €tz K, Fuchs D, et al. Two cases of metastases to the 23. Wurdinger S, Schu breast on MR mammography. Eur Radiol 2001; 11:802–806. 24. Lehman C, Holt S, Peacock S, et al. Use of the American College of Radiology BI-RADS guidelines by community radiologists: concordance of assessments and recommendations assigned to screening mammograms. AJR Am J Roentgenol 2002; 179:15–20. 25. Kim EK, Ko KH, Oh KK, et al. Clinical application of the BI-RADS final assessment to breast sonography in conjunction with mammography. AJR Am J Roentgenol 2008; 190:1209–1215. 26. Kuhl CK, Mielcareck P, Leutner C, et al. Dynamic breast MR imaging: are signal intensity time course data useful for differential diagnosis of enhancing lesions. Radiology 1999; 211:101–110. 27. Szabo BK, Aspelin P, Wiberg MK, et al. Dynamic MR imaging of the breast: analysis of kinetic and morphologic diagnostic criteria. Acta Radiol 2003; 44:379–386. 28. Baum F, Fischer U, Vosshenrich R, et al. Classification of hypervascularized lesions in CE MR imaging of the breast. Eur Radiol 2002; 12:1087–1092. 29. Macura KJ, Ouwerkerk R, Jacobs MA, et al. Patterns of enhancement on breast MR images: interpretation and imaging pitfalls. Radiographics 2006; 26:1719–1734. 30. Erguvan-Dogan B, Whitman GJ, Kushwaha AC, et al. BI-RADS-MRI: a primer. AJR Am J Roentgenol 2006; 187:W152–W160. 31. Morris EA. Breast MR imaging lexicon updated. Magn Res Imaging Clin N Am 2006; 14:293–303. 32. Sneige N, Zachariah S, Fanning T, et al. Fine-needle aspiration cytology of metastatic neoplasms in the breast. Am J Clin Pathol 1989; 92:27–35. 33. Van Hoeven KH, Hibbard CA, Flax H, et al. Metastatic malignant neoplasms and secondary lymphomatous involvement of the breast: a study of 43 cases. Pathol Annu 1993; 28:221–241. 34. Lee SK, Kim WW, Kim SH, et al. Characteristics of metastasis in the breast from extramammary malignancies. J Surg Oncol 2010; 101:137–140. 35. Noguera JJ, Martinez-Miravete P, Idoate F, et al. Metastases to the breast: a review of 33 cases. Australas Radiol 2007; 51:133–138. 36. Muttarak M, Nimmonrat A, Chaiwun B. Metastatic carcinoma to the male and female breast. Australas Radiol 1998; 42:16–19. 37. Vizcaino I, Torregrosa A, Higueras V, et al. Metastasis to the breast from extramammary malignancies: a report of four cases and a review of literature. Eur Radiol 2001; 11:1659–1665. 38. Soo MS, Williford ME, Elenberger CD. Medullary thyroid carcinoma metastatic to the breast: mammographic appearance. AJR Am J Roentgenol 1995; 165:65–66. rel Y, et al. Breast metastases in adolescent 39. Chateil JF, Arboucalot F, Pe girls: US findings. Pediatr Radiol 1998; 28:832–835. 40. Lee WK, Cawson JN, Hill PA, et al. Renal cell carcinoma metastasis to the breast: mammographic, sonographic, CT, and pathologic correlation. Breast J 2007; 13:316–317.
573
SUROV ET AL
41. Kim SM, Ko EY, Han BK, et al. Metastatic thymoma of the breast. Korean J Radiol 2008; 9:80–83. 42. Francois C, Rangachari B, Bova D. Mammography and sonography of pathologically proven adrenal cortical carcinoma metastatic to the breast. AJR Am J Roentgenol 2005; 184:1279–1281. 43. Ashkar L, Mesurolle I, Trembley F. Sonographic target sign and rim enhancement on magnetic resonance imaging in metastatic melanoma to the breast. Australas Radiol 2006; 50:237–240. 44. Yang WT, Kwan WH, Chow LT, et al. Unusual sonographic appearance with color Doppler imaging of bilateral breast metastases in a patient with alveolar rhabdomyosarcoma of an extremity. J Ultrasound Med 1996; 15:531–533. 45. Tukel S, Dogan BE, Ozcan H. Alveolar rhabdomyosarcoma metastatic to the breast. Curr Probl Diagn Radiol 2003; 32:102–104. 46. Kalra N, Ojili V, Gulati M, et al. Metastatic choriocarcinoma to the breast: appearance on mammography and Doppler sonography. AJR Am J Roentgenol 2005; 184(suppl):S53–S55. 47. Upalakalin JN, Collins LC, Tawa N, et al. Carcinoid tumors in the breast. Am J Surg 2006; 191:799–805. 48. Jakovljevic B, Stevanovic O, Bacic G. Metastases to the breast from smallcell lung cancer: MR findings. A case report. Acta Radiol 2003; 44:485–488.
574
Academic Radiology, Vol 18, No 5, May 2011
49. Perlet C, Sittek H, Forstpointner R, et al. Metastases to the breast from rhabdomyosarcoma: appearances on MRI. Eur Radiol 1999; 9:1113–1116. 50. Kelkar PS, Helbich TH, Becherer A, et al. Solitary breast metastasis as the first sign of a squamous cell carcinoma of the cervix: imaging findings. Eur J Radiol 1997; 24:159–162. 51. Briest S, Horn LC, Haupt R, et al. Metastasizing signet ring cell carcinoma of the stomach-mimicking bilateral inflammatory breast cancer. Gynecol Oncol 1999; 74:491–494. 52. Birjawi GA, Haddad MC, Tawil AN, et al. Metastatic rhabdomyosarcoma to the breast. Eur Radiol 2001; 11:555–558. 53. Ho LW, Wong KP, Chan JH, et al. MR appearance of metastatic melanotic melanoma in the breast. Clin Radiol 2000; 55:572–573. € u €roglu M, Ersavasti G, Ilvan S, et al. Bilateral inflammatory breast 54. Ozg metastases of epithelial ovarian cancer. Am J Clin Oncol 1999; 22: 408–410. 55. Gomori JM, Grossman RI, Shields JA, et al. Choroidal melanomas: correlation of NMR spectroscopy and MR imaging. Radiology 1986; 158: 443–445. 56. Woodruff WW Jr, Djang WT, McLendon RE, et al. Intracerebral malignant melanoma: high-field-strength MR imaging. Radiology 1987; 165: 209–213.
M
ost secondary intramammary tumors occur as metastatic involvement from the contralateral breast (1–3). Breast metastases (BM) from nonmammary malignancies are very rare (4–6). This may related to the fact that the breast contains large areas of fibrous tissue with a relatively poor blood supply (1,2). Alva and Shetty-Alva (7) found in their review of the literature that most BM arise from malignant melanomas, sarcomas, lung cancers, ovarian tumors, renal carcinomas, and thyroid tumors in decreased order of frequency (7). Some single-institution studies have suggested, however, that BM from ovarian and renal carcinomas are very rare (8). The reported data regarding the prevalence of BM in nonmammary malignancies are also controversial. Abrams et al (9) detected BM in 5% of 1,000 autopsied cases of carcinoma. Hejmadi et al (10), however, reported a prevalence of BM of 0.11%.
Most reports regarding BM are clinical observations or pathologic series and provide either no or inconsistent radiologic information (11–17). Radiologic publications consist predominantly of isolated case reports or small series (18–23). Furthermore, most radiologic studies were conducted from 1970 to 1990, before the Breast Imaging Reporting and Data System (BI-RADS) classification had been introduced (18–22). In addition, ultrasound and magnetic resonance imaging (MRI) of BM have been reported only sporadically in the literature (2,23). Therefore, the purpose of this study was to estimate the prevalence of BM and to describe their radiologic appearance.
MATERIALS AND METHODS This retrospective study was approved by the institutional ethics committee.
Acad Radiol 2011; 18:565–574 From the Departments of Radiology (A.S., K.R., R.-P.S.), Dermatology (E.F.), Pathology (H.-J.H.), and Oncology (H.-J.S.), Martin-Luther-University HalleWittenberg, Ernst-Grube-Straße 40, 06097 Halle, Germany. Received October 17, 2010; accepted December 14, 2010. Address correspondence to: A.S. e-mail: [email protected] ªAUR, 2011 doi:10.1016/j.acra.2010.12.009
Patients
A retrospective search in the statistical database of our institution from January 2000 to December 2009 revealed 6,668 patients with metastasizing solid malignancies who were examined and treated at our institution. In addition, 2,896 565
SUROV ET AL
patients with breast carcinoma were treated at our breast center in the investigated time period. Furthermore, we retrospectively reviewed the radiologic database for metastatic breast involvement in oncologic diseases (through a computerized search for key words in the reports). BM were identified in 51 patients, including 43 women and eight men with a median age of 61 years (range, 24–84 years). In all of the identified patients, the diagnosis of metastatic disease was proven histopathologically. In 36 patients (71%), the diagnosis was confirmed on breast biopsy specimens. In the other 15 patients (29%) with multiple metastases, soft tissue or cutaneous lesions of another localization had been investigated pathologically. All radiologic images of the identified patients were reevaluated. Patients with direct invasion of tumors into the breast from the thoracic wall were excluded from the study. Furthermore, patients with primary breast sarcoma, primary or secondary breast lymphoma, and breast involvement in hematologic oncologic diseases were also not included in the study. Additionally, we validated the radiologically identified BM cases by comparison with the database of the department of pathology of our hospital. Imaging and Review
Computed Tomography. Computed tomography (Somatom Plus 4 VZ and Somatom Sensation 64; Siemens Medical Systems, Erlangen, Germany) of the thoracic region was performed as a staging investigation in 38 patients. In all cases, 1.5 mL of iodinated intravenous contrast medium (Solutrast; Bracco Altana GmbH, Konstanz, Germany) per body mass was given at a rate of 1.5 to 3.5 mL/s by a power injector (Medtron GmbH, Germany), with a scan delay of 60 seconds after the onset of injection. Typical imaging parameters were 120 kVp, 150 to 300 mAs, and a 0.6-mm to 6-mm slice thickness, with a pitch of 0.8 to 1.2. Mammography. Mammography was performed in 37 patients using dedicated mammographic equipment (Mammomat 3000 Nova and Mammomat Novation DR; Siemens Medical Systems). Standard mediolateral oblique and craniocaudal mammograms were obtained, with additional views if necessary. Breast findings were classified according to the American College of Radiology’s BI-RADS lexicon (24).
Academic Radiology, Vol 18, No 5, May 2011
TABLE 1. Prevalence of BM in Metastasizing Oncologic Diseases Malignant Disease
Prevalence of BM (%)
Malignant melanoma of skin Carcinoma of ovary Carcinoma of stomach Renal cell carcinoma Sarcoma Carcinoma of bronchus and lung Carcinoma of larynx Carcinoma of cervix uteri Seminoma of testis Carcinoma of prostate Carcinoma of thyroid gland Carcinoma of colon
4.92 4.43 1.47 1.17 1.13 0.89 0.77 0.76 0.52 0.24 0.22 0.12
BM, breast metastases.
thickness using a T1-weighted (T1W) three-dimensional fast low-angle shot sequence (repetition time, 14 ms; echo time, 7 ms; flip angle, 25 ). Measurement time was 87 seconds, once before and five times after intravenous bolus injection of 0.2 mmol gadolinium diethylenetriamine penta-acetic acid per kilogram body weight (Schering AG, Berlin, Germany). The flow rate was 2 mL/s, and the bolus was followed by a 20-mL saline flush. Central k-space data were acquired at 43, 130, 217, 304, and 391 seconds after injection. Automatic image subtraction was performed routinely. T2-weighted images included a turbo spin-echo sequence (repetition time, 4,200 ms; echo time, 90 ms; flip angle, 180 ) with a 252 256 matrix, a 350-mm field of view, and a slice thickness of 5 mm. Kinetic analysis of contrast enhancement was performed additionally. Time–signal intensity curves were drawn using operator-defined regions of interest. The region of interest was smaller than the lesion size. The time to peak enhancement (ie, the time between administration of contrast agent and the maximum signal intensity value in the postcontrast phase) was estimated. Furthermore, the initial signal increase (SI) from the precontrast value (SIp) to the maximum peak within the first 3 minutes after the administration of contrast medium (SI1–3 min) was calculated (26–29): initial SI ð%Þ ¼ ðSI13min SIP Þ=SIP 100%:
Ultrasound. Ultrasound evaluation was performed in 43 patients using a 10.5-MHz linear-array transducer (Sonoline Elegra and Acuson Antares; Siemens Medical Systems). Additionally, power and color Doppler ultrasound scanning was performed in 20 patients. Sonographic findings were classified according to the BI-RADS ultrasound lexicon (25).
The postinitial behavior of the signal curve from the maximum peak (SIpeak) to the end of the examination (SIend) was also analyzed (26–29):
MRI. In 24 patients, MRI of the breast was performed additionally using a dedicated receive-only breast coil on a 1.0-T scanner (Impact Expert; Siemens Medical Systems). Both breasts were imaged in the coronal plane with 2.5-mm slice
MRI findings were classified according to the American College of Radiology’s BI-RADS lexicon (30,31). All available images were interpreted by two radiologists (A.S. and K.R., with 7 and 11 years of experience,
566
postinitial SI ð%Þ ¼ SIpeak SIend
SIpeak 100%:
Academic Radiology, Vol 18, No 5, May 2011
METASTASES TO THE BREAST
Figure 1. Imaging findings in a 48-year old woman with a known history of bronchial carcinoma. (a) Staging computed tomographic examination detecting an oval mass with moderate homogenous enhancement within the left breast (arrow). (b) Bilateral mediolateral oblique mammograms show multiple round and oval dense lesions with circumscribed and microlobulated margins (arrows). (c) Ultrasound of the left breast documents an oval mixed isoechoic to hyperechoic mass with small boundary posterior shadowing. (d) Transverse power Doppler ultrasound scan shows no hypervascularity of the mass. (e) On T1W images before contrast administration, the mass is hypointense to isointense compared to breast tissue (arrow). (f) Subtracted images after intravenous administration of contrast medium demonstrated marked homogenous enhancement of the mass (arrow). Time-intensity curve of the lesion documenting rapid enhancement in the first minute associated with wash out (type 3 curve). Time to peak enhancement was 90 seconds, initial increase of signal intensity was 132%, and washout ratio was 22%. (g) Histologic analysis after ultrasound-guided biopsy revealed a metastasis of bronchial carcinoma (hematoxylin and eosin stain, 100).
respectively). Consensus of the investigators was obtained on the following features of the identified breast lesions: number, shape, localization, size, margin, attenuation, homogeneity, and contrast enhancement patterns. Lesion size was determined by measuring the maximum diameter. Furthermore, case histories were reviewed retrospectively to determine clinical signs at presentation. Histologic Analysis
Biopsy specimens were evaluated by one pathologist (H.-J.H., with 37 years of experience). All available histologic materials (sections stained by hematoxylin and eosin and by immunohistochemistry) were analyzed.
Statistical Analysis
For statistical analysis, SPSS (SPSS, Inc, Chicago, IL) was used. Collected data were evaluated by means of descriptive statistics (absolute and relative frequencies). Continuous variables are expressed as mean standard deviation and categorical variables as percentages. RESULTS Primary Tumors and Number and Prevalence of BM
One hundred eighty-four metastases were diagnosed in the breast in 51 of 6,668 oncology patients (0.76%) who met 567
SUROV ET AL
Academic Radiology, Vol 18, No 5, May 2011
Figure 2. Radiologic findings in a 67-year-old woman with a known history of small cell lung cancer. (a) Staging computed tomographic examination detecting a round mass with moderate inhomogenous enhancement within the right breast (arrow). Additionally, right pulmonary masses are seen. (b) On mammography, the mass was round with microlobulated margins (arrow). (c) Transverse ultrasound documents irregular lesions with microlobulated margins and lightly posterior acoustic enhancement. (d) Transverse power Doppler shows marked internal hypervascularity. Histologic examination of the mass revealed a metastasis of the known bronchial carcinoma (not shown).
the inclusion criteria. The number of BM varied from one to 26. In 24 patients, BM were localized in the right breast (47 lesions [25.5%]) and in 13 patients (28 lesions [15.2%]) in the left breast, and in 14 patients (109 lesions [59.3%]), BM were bilateral. The size of metastases varied from 2 to 60 mm (median size, 8 mm). Systemic metastases outside the breast were diagnosed in 40 these 51 patients (78.4%). In 11 patients (21.6%), BM were the only manifestation of metastatic disease. Furthermore, in three of the 11 patients, BM were the first sign of malignancy. The intramammary metastases were derived from the following malignancies: malignant melanoma (31%), genital tumors (18%), pulmonary tumors (16%), urologic tumors (12%), gastrointestinal tumors (7%), and sarcomas (4%). Other neoplasia were rare. In 4% of the cases, no primary tumor could be identified. The intramammary lesions showed here no typical immunohistochemical patterns for primary breast cancer. The prevalence of detected BM associated with the different tumors is given in Table 1. Clinical Findings
Clinically, 16 patients (31%) presented with solitary or multiple painless breast lumps. In five patients (10%), unilateral axillary lymphadenopathy occurred additionally. Skin edema was seen in two patients (3.9%). In 30 patients (59%), however, breast involvement was diagnosed incidentally during staging examination on computed tomography. Furthermore, in five patients with primary unknown malignancies (10%), BM were found during a thorough tumor search on mammography or breast ultrasound. Radiologic Features
Computed Tomography. On computed tomography, which was available for 38 patients (108 lesions), most metastases 568
were round or oval in shape with circumscribed margins (Figs 1–3). They showed a marked enhancement in 52% and moderate and light enhancement in 37% and 11%, respectively (Table 2). Mammography. Mammography was performed in 37 patients. Mammographic findings are summarized in Table 3. Solitary or multiple breast masses (Figs 1–3) were the most common pattern of BM (27 patients [73%]). In two patients (5%), architectural distortion (diffuse unilateral or bilateral breast opacity) was seen. In eight patients (22%), however, no abnormalities could be identified on mammography. Twenty percent of the mammographic findings were interpreted as BI-RADS category 0 (American College of Radiology 4), 14% as BI-RADS category 3, 37% as BI-RADS category 4, and 29% as BI-RADS category 5. Ultrasound. Ultrasound images were available for 43 patients (71 lesions). The ultrasound features are shown in Table 4. Most of the intramammary masses were hypoechoic, oval or round in shape, with microlobulated or circumscribed margins and posterior acoustic enhancement (Figs 1–3). Color and power Doppler imaging showed hypervascularity in 16 of the identified lesions (39%; Table 5). US findings were categorized as BI RADS 2 in 1 (2%), BI RADS 3 in 8 (19%), BI RADS 4 in 29 (67%) and BI RADS 5 in 5 (12%) cases. MRI. MRI was performed in 24 patients (93 lesions). On T2weighted images, most lesions were isointense in comparison to the normal breast parenchyma (Table 6). On T1W images, most BM were isointense to the normal breast tissue and demonstrated marked homogenous contrast enhancement after intravenous administration of contrast medium (Figs 1, 3, and 4). Kinetic analysis of contrast enhancement was performed for 62 lesions. It showed in most cases (89%) a rapid initial SI >100% in comparison with precontrast signal intensity. The median initial
Academic Radiology, Vol 18, No 5, May 2011
METASTASES TO THE BREAST
Figure 3. Imaging findings in a 46-year old woman with a known history of medullary thyroid carcinoma. (a) Staging CT examination detecting an oval mass with circumscribed margins and marked enhancement within the left breast (arrow). On mammography (not shown), there was no evidence of malignancy (American College of Radiology 4). (b) Ultrasound documents a round hypoechoic to hyperechoic mass with indistinct margins without posterior acoustic phenomena. (c) Transverse power Doppler shows no hypervascularity of the lesion. (d) On T1W images before contrast administration, the mass is hypointense to isointense compared to breast tissue (arrow). (e) Subtracted images after intravenous administration of contrast medium demonstrated marked homogenous enhancement of the lesion (arrow). Time-intensity curve of the lesion showing rapid enhancement in the first minute associated with washout (type 3 curve). Time to peak enhancement was 90 seconds, initial increase of signal intensity was 506%, and washout ratio was 14.2%. (f) Histologic examination after ultrasound-guided biopsy revealed metastases of medullary thyroid carcinoma (hematoxylin and eosin stain, 50).
SI was 167% (range, 20%–672%). The median time to peak enhancement was 180 seconds (range, 90–270 seconds).
MRI findings were categorized as BI-RADS category 4 in 11 patients (46%) and BI-RADS category 5 in 13 patients (54%). 569
SUROV ET AL
Academic Radiology, Vol 18, No 5, May 2011
TABLE 2. Computed Tomographic Findings of Breast Metastases
TABLE 4. Ultrasound Findings of Breast Metastases Characteristic
Characteristics Lesions Shape Oval Round Lobulated Margin Circumscribed Lobulated Enhancement High Moderate Low
n (%) 108 (100) 18 (17) 82 (76) 8 (7) 101 (94) 7 (6) 56 (52) 40 (37) 12 (11)
TABLE 3. Mammographic Findings in 37 Patients with Breast Metastases Characteristic Pattern Mass Architectural distortion No abnormalities Mass features (n = 52) Shape Round Oval Lobular Margin Circumscribed Microlobulated Density High Isodense Calcifications Yes No ACR 1 2 3 4 BI-RADS 0 3 4 5
Patients 27 (73%) 2 (5%) 8 (22%) Lesions
41 (95%) 2 (5%) Lesions 30 (44%) 28 (41%) 10 (15%) 34 (50%) 18 (26%) 16 (24%) 13 (19%) 35 (52%) 20 (29%) 52 (76%) 11 (16%) 4 (6%) 1 (2%)
24 (46%) 28 (54%)
needle was performed in 26 patients. In two patients, stereotactic biopsy (11-gauge needle) was performed. In eight patients, incisional breast biopsy was performed. In the other 15 patients (29%) with multiple metastases, soft tissue or cutaneous lesions of another localization had been investigated pathologically. In 11 patients (22%), in whom BM were the only manifestation of metastatic disease, metastases were resected surgically. Patients with systemic metastases were treated with chemotherapy. Thirty patients (59%) died during therapy within 1 year after the diagnosis of BM. The remaining 10 patients are alive to date under chemotherapy.
5 (10%) 47 (90%)
DISCUSSION
37 (71%) 9 (17%) 6 (12%) 46 (88%) 6 (12%)
31% 31% 12% 26% 20% 14% 37% 29%
ACR, American College of Radiology; BI-RADS, Breast Imaging Reporting and Data System.
Diagnosis and Follow-Up
In all of the identified cases, the diagnosis of metastatic disease was proven histopathologically. In 36 patients (71%), the diagnosis was confirmed on breast biopsy specimens. Ultrasound-guided core-needle biopsy using a 14-gauge 570
Pattern Mass Architectural distortion (diffuse infiltration) Mass features (n = 68) Shape Round Oval Lobular Margin Microlobulated Circumscribed Indistinct Echo pattern Anechoic Hypoechoic Mixed hypoechoic to hyperechoic Posterior phenomena Enhancement None Shadowing Mixed
Patients
The prevalence of BM varies significantly in several studies, depending on how the prevalence is calculated. In relation to all breast biopsies performed in several single-institution studies, it ranged from 0.11% to 1.93% (10,32). Furthermore, it was 0.33% to 6.3% of cases with identified breast malignancies (8,33). The frequency of metastatic involvement of the breast in several tumors has been reported in only one series previously (34). Here, it ranged from 0.019% in hepatocellular carcinoma to 0.621% in thymus carcinoma (34). The investigators estimated the prevalence for each tumor with BM. However, it was calculated related to the total number of malignancies, regardless whether they had metastasized or not. In our study, the prevalence of BM was generated in relation to metastasizing tumors only and ranged from 0.12% to 4.92% (Table 1). Furthermore, the frequency of BM from all breast malignancies was 1.76% in our study. According to previous reports, BM are more common solitary than bilateral or multiple lesions (1–5,12–16). In the
Academic Radiology, Vol 18, No 5, May 2011
METASTASES TO THE BREAST
TABLE 5. Doppler Findings of 42 Breast Metastases in 20 Patients Hypervascularity None Light Moderate High
n (%) 25 (61) 10 (24) 2 (5) 4 (10)
TABLE 6. Magnetic Resonance Findings of Breast Metastases Characteristics
Figure 4. Magnetic resonance imaging of the breasts in a 28-yearold woman with a known history of ovarian cancer and right breast lumps. Mammography (not shown) demonstrated an architectural distortion (diffuse dense opacity) of the right breast. (a) Subtracted images after intravenous administration of contrast medium showing diffuse infiltration of the right breast with marked inhomogenous enhancement (arrow). (b) Transverse reconstruction of the breasts (subtracted images) reveals also diffuse infiltration of the right breast (arrow). Time-intensity curves document continuous increase in signal intensity on each successive contrast-enhanced image >10% in comparison with initial enhancement (type 1). (c) Histologic analysis after magnetic resonance–guided biopsy showing diffuse metastatic infiltration of the breast (hematoxylin and eosin stain, 100).
present study, however, 59.3% of BM were bilateral. Furthermore, most BM were multiple. Clinically, most intramammary metastases have been reported to be rapidly growing masses without pain (5,12– 16). Retraction of the skin or nipple was typically not seen (1,13–15). In our study, in 59% of patients, breast involvement was asymptomatic and diagnosed incidentally during staging examination. Only 31% of the patients presented clinically with solitary or multiple painless breast lumps with or without unilateral axillary lymphadenopathy. Radiologic series have described relatively small numbers of patients, ranging from six to 33 (18–22,35). Furthermore, patients were often investigated only by a single
Pattern Mass Diffuse infiltration Mass features (n = 91) Shape Round Oval Lobular Margin Smooth Irregular Signal intensity (T1W images) Hypointense Isointense Mixed isointense to hyperintense Signal intensity (T2W images) Isointense Hyperintense Internal mass enhancement Homogenous Inhomogenous Rim enhancement Time–signal intensity curve pattern Type 1 Type 2 Type 3
Patients 22 (92%) 2 (8%) Lesions 54 (59%) 32 (35%) 5 (6%) 84 (92%) 7 (8%) 10 (11%) 79 (87%) 2 (2%) 76 (84%) 15 (16%) 73 (80%) 7 (8%) 11 (12%) 11 (18%) 32 (52%) 19 (30%)
examination. Therefore, our study, with 184 findings in 51 patients, is the largest to date (Table 7). Computed tomographic appearances of BM have been described previously in rare cases (35). According to the present study, most metastases were round or oval in shape, with circumscribed margins and high or moderate enhancement. This finding is unspecific and could equally represent a cyst, fibroadenoma, or breast carcinoma. Therefore, further radiologic investigations are needed to determine these lesions. Mammographic appearances of BM have been reported as well known (1,5,15,16,18–22,35–37). According to previous reports, the most common mammographic appearances are circumscribed masses, located in the upper outer quadrants without spiculation (5,18–22,35–37). However, other findings are also documented. For example, in the largest mammographic study (n = 18), by Noguera et al (35), the 571
SUROV ET AL
Academic Radiology, Vol 18, No 5, May 2011
TABLE 7. Comparison of Reported Studies Regarding Radiologic Findings of Breast Metastases Radiologic Investigations (Patients) Study Bohman et al (18)* McCrea et al (19)* Derchi et al (20) Hebert et al (21) Vergier et al (5) Muttarak et al (36) Yeh et al (2) Noguera et al (35)* Present study
Patients
CT
Mammography
Ultrasound
Doppler
MRI
11 8 6 21 8 7 15 22 51
— — — — — — — 4 38
11 8 — 5 8 7 — 18 37
— — 6 — 8 6 11 11 43
— — — — — — — — 20
— — — — — — — 1 24
CT, computed tomography; MRI, magnetic resonance imaging. *Cases of lymphoma, leukemia, and intramammary metastases from contralateral breast cancer were excluded from the analysis.
findings of three BM cases (17%) resembled those of inflammatory breast carcinoma. In addition, asymmetric density occurred in one case (6%). Furthermore, the identified masses showed circumscribed margins in 58% only, and in 42%, indistinct margins were seen (35). According to the literature, some metastases present microcalcifications (1,5,18–22). Typically, they occur in mucinous ovarian carcinoma (18–22). However, metastases of hepatocellular carcinoma, gastric cancer, and medullary thyroid carcinoma can also develop calcifications (34,35,38). In our series we analyzed 37 mammograms (56 findings). Therefore, this is the largest study to date. In 5% of the cases, BM manifested as diffuse unilateral or bilateral breast opacity. Most BM, which presented as breast masses, were round or oval in shape, with circumscribed margins. Lobular lesions occurred in 12%. Calcifications were seen in 10%. There were cases with metastatic ovarian cancer. Ultrasound findings of BM have been described in small series ranging from six to 11 cases (2,5,20,35,36,39). According to the literature, intramammary metastases were hypoechoic or isoechoic in comparison to the breast parenchyma, with circumscribed margins (2,20,35). In addition, most showed no posterior acoustic phenomena. In a study by Yeh et al (2), however, five of 11 lesions had irregular margins. In our study, ultrasound investigation was performed in 43 patients (71 lesions). In contrast to the previous reports, several patterns could be identified here. Fifteen percent of intramammary masses were lobular, 41% oval, and 44% round. Twenty-six percent of lesions had circumscribed margins, 50% were microlobulated, and 24% had indistinct margins. Most lesions were homogenously hypoechoic with posterior acoustic enhancement. Interestingly, many BM were anechoic and demonstrated posterior acoustic enhancement. This pattern is typical of breast cysts, and these lesions can be misdiagnosed. Therefore, further investigations, such as Doppler imaging, are needed here. Doppler imaging findings in BM have been reported in only eight lesions in the literature (40–47). In most cases, metastatic lesions were hypervascular (40,42,44–47). The published 572
metastases arise from sarcomas, carcinoid tumors, melanoma, and renal cell carcinoma (ie, from hypervascularized tumors). Our study, however, showed that in 61% of BM, no hypervascularity was seen. Light vascularity was documented in 24%, moderate in 5%, and high in 10%. MRI findings of BM have been reported in only 12 patients previously (23,35,43,48–54). There were isolated case reports describing MRI findings in several malignancies. Furthermore, in most reports, no information of enhancement kinetics was provided. In our series, 24 patients with 93 lesions were analyzed. On T2-weighted images, most lesions were isointense in comparison to the normal breast parenchyma and hypointense compared to fat. On T1W images, most BM (87%) were isointense or lightly hypointense compared to the normal breast tissue. However, 2% of the metastases were mixed isointense to hyperintense. High signal intensity of metastases on T1W images was observed in our study in metastases from melanoma. Metastases of malignant melanoma contain melanin, which shows a hyperintense signal on T1W images (55). On the other hand, melanoma metastases often bleed (56). However, SI on T1W images was seen in only two of 10 patients with melanoma, who were investigated using MRI. After intravenous contrast administration, several patterns of BM were identified in our series. The diagnosed metastases presented as solitary or multiple lesions, round or oval in shape with smooth margins and marked homogenous contrast enhancement. Inhomogenous and rim enhancement were rarely present. Furthermore, in 8%, diffuse infiltration of breast parenchyma was identified. As seen, MRI appearances of BM may mimic benign breast masses, such as fibroadenoma. Therefore, it is important to use further analysis of breast MRI findings, particularly kinetic analysis. In our series, kinetic analysis of contrast enhancement showed in most cases a rapid initial SI >100% in comparison to precontrast signal intensity. This pattern is typical for breast malignancies (26,29). However, in 11% of the lesions, initial SI was <100%, ranging from 20% to 92%.
Academic Radiology, Vol 18, No 5, May 2011
Type 2 and 3 kinetics were present in most cases (82%). According to the literature, both types of curves suggest malignancy (26,29). In 18% of BM, type 1 kinetic curve was seen. This pattern is usually associated with benign findings (26,29). As seen, BM presented with different radiologic appearances and may be misdiagnosed as primary breast carcinoma or as benign lesions. Therefore, the diagnosis of BM should be confirmed histopathologically. Some reports have suggested that fine-needle biopsy can definitively confirm the diagnosis of BM (10,11,17). However, several metastatic malignancies, such as endometrial or prostatic cancers, and some variants of malignant melanomas have cytologic features resembling mammary carcinoma (33). In our study, core needle and incision biopsy was used in all cases. There were several limitations to our study. First, this was a retrospective analysis. Second, not each patient underwent all investigations (mammography, ultrasound, Doppler, and MRI). Third, in only 71% of the patients was the diagnosis established by breast biopsy. In the other 29% of patients, soft tissue or cutaneous lesions of another localization had been investigated pathologically. CONCLUSIONS BM manifest most frequently as round or oval masses with circumscribed margins on mammography and as hypoechoic masses with microlobulated or circumscribed margins and posterior acoustic enhancement on ultrasound. On MRI, most BM present as circumscribed masses with marked or moderate homogenous enhancement. These radiologic features should be taken into consideration in the differential diagnosis of breast lesions. REFERENCES 1. Bartella L, Kaye J, Perry NM, et al. Metastases to the breast revisited: radiological-histopathological correlation. Clin Radiol 2003; 58:524–531. 2. Yeh CN, Lin CH, Chen MF. Clinical and ultrasonographic characteristics of breast metastases from extramammary malignancies. Am Surg 2004; 70: 287–290. 3. Hajdu SI, Urban JA. Cancers metastatic to the breast. Cancer 1972; 29: 1691–1696. 4. Yang WT, Muttarak M, Ho LW. Nonmammary malignancies of the breast: ultrasound, CT, and MRI. Semin Ultrasound CT MRI 2000; 21:375–394. 5. Vergier B, Trojani M, de Mascarel I, et al. Metastases to the breast: differential diagnosis from primary breast carcinoma. J Surg Oncol 1991; 48: 112–116. 6. Harvey JA. Unusual breast cancers: useful clues to expanding the differential diagnosis. Radiology 2007; 242:683–694. 7. Alva S, Shetty-Alva N. An update of tumor metastasis to the breast. Arch Surg 1999; 134:450. 8. Wood B, Sterrett G, Frost F, et al. Diagnosis of extramammary malignancy metastatic to the breast by fine needle biopsy. Pathology 2008; 40: 345–351. 9. Abrams HL, Spiro R, Goldstein N. Metastases in carcinoma. Analysis of 1000 autopsied cases. Cancer 1950; 3:74–85. 10. Hejmadi RK, Day LJ, Young JA. Extramammary metastatic neoplasms in the breast: a cytomorphological study of 11 cases. Cytopathology 2003; 14:191–194.
METASTASES TO THE BREAST
11. Silverman JF, Feldman PS, Covell JL, et al. Fine needle aspiration cytology of neoplasms metastatic to the breast. Acta Cytol 1987; 31:291–300. 12. Georgiannos SN, Chin J, Goode AW, et al. Secondary neoplasms of the breast: a survey of the 20th Century. Cancer 2001; 92:2259–2266. 13. Amichetti M, Perani B, Boi S. Metastases to the breast from extramammary malignancies. Oncology 1990; 47:257–260. 14. Vaughan A, Dietz JR, Moley JF, et al. Metastatic disease to the breast: the Washington University experience. World J Surg Oncol 2007; 5:74. 15. Smyrniotis V, Theodosopoulos T, Marinis A, et al. Metastatic disease in the breast from nonmammary neoplasms. Eur J Gynaecol Oncol 2005; 26: 547–550. 16. Williams SA, Ehlers RA II, Hunt KK, et al. Metastases to the breast from nonbreast solid neoplasms: presentation and determinants of survival. Cancer 2007; 110:731–737. 17. David O, Gattuso P, Razan W, et al. Unusual cases of metastases to the breast. A report of 17 cases diagnosed by fine needle aspiration. Acta Cytol 2002; 46:377–385. 18. Bohman LG, Bassett LW, Gold RH, et al. Breast metastases from extramammary malignancies. Radiology 1982; 144:309–312. 19. McCrea ES, Johnston C, Haney PJ. Metastases to the breast. AJR Am J Roentgenol 1983; 141:685–690. 20. Derchi LE, Rizzatto G, Giuseppetti GM, et al. Metastatic tumors in the breast: sonographic findings. J Ultrasound Med 1985; 4:69–74. 21. Hebert G, Ouimet-Oliva D, Paquin F, et al. Diffuse metastatic involvement of the breast. Can Assoc Radiol J 1991; 42:353–356. 22. Toombs BD, Kalisher L. Metastatic disease to the breast: clinical, pathologic, and radiographic features. AJR Am J Roentgenol 1977; 129: 673–676. €tz K, Fuchs D, et al. Two cases of metastases to the 23. Wurdinger S, Schu breast on MR mammography. Eur Radiol 2001; 11:802–806. 24. Lehman C, Holt S, Peacock S, et al. Use of the American College of Radiology BI-RADS guidelines by community radiologists: concordance of assessments and recommendations assigned to screening mammograms. AJR Am J Roentgenol 2002; 179:15–20. 25. Kim EK, Ko KH, Oh KK, et al. Clinical application of the BI-RADS final assessment to breast sonography in conjunction with mammography. AJR Am J Roentgenol 2008; 190:1209–1215. 26. Kuhl CK, Mielcareck P, Leutner C, et al. Dynamic breast MR imaging: are signal intensity time course data useful for differential diagnosis of enhancing lesions. Radiology 1999; 211:101–110. 27. Szabo BK, Aspelin P, Wiberg MK, et al. Dynamic MR imaging of the breast: analysis of kinetic and morphologic diagnostic criteria. Acta Radiol 2003; 44:379–386. 28. Baum F, Fischer U, Vosshenrich R, et al. Classification of hypervascularized lesions in CE MR imaging of the breast. Eur Radiol 2002; 12:1087–1092. 29. Macura KJ, Ouwerkerk R, Jacobs MA, et al. Patterns of enhancement on breast MR images: interpretation and imaging pitfalls. Radiographics 2006; 26:1719–1734. 30. Erguvan-Dogan B, Whitman GJ, Kushwaha AC, et al. BI-RADS-MRI: a primer. AJR Am J Roentgenol 2006; 187:W152–W160. 31. Morris EA. Breast MR imaging lexicon updated. Magn Res Imaging Clin N Am 2006; 14:293–303. 32. Sneige N, Zachariah S, Fanning T, et al. Fine-needle aspiration cytology of metastatic neoplasms in the breast. Am J Clin Pathol 1989; 92:27–35. 33. Van Hoeven KH, Hibbard CA, Flax H, et al. Metastatic malignant neoplasms and secondary lymphomatous involvement of the breast: a study of 43 cases. Pathol Annu 1993; 28:221–241. 34. Lee SK, Kim WW, Kim SH, et al. Characteristics of metastasis in the breast from extramammary malignancies. J Surg Oncol 2010; 101:137–140. 35. Noguera JJ, Martinez-Miravete P, Idoate F, et al. Metastases to the breast: a review of 33 cases. Australas Radiol 2007; 51:133–138. 36. Muttarak M, Nimmonrat A, Chaiwun B. Metastatic carcinoma to the male and female breast. Australas Radiol 1998; 42:16–19. 37. Vizcaino I, Torregrosa A, Higueras V, et al. Metastasis to the breast from extramammary malignancies: a report of four cases and a review of literature. Eur Radiol 2001; 11:1659–1665. 38. Soo MS, Williford ME, Elenberger CD. Medullary thyroid carcinoma metastatic to the breast: mammographic appearance. AJR Am J Roentgenol 1995; 165:65–66. rel Y, et al. Breast metastases in adolescent 39. Chateil JF, Arboucalot F, Pe girls: US findings. Pediatr Radiol 1998; 28:832–835. 40. Lee WK, Cawson JN, Hill PA, et al. Renal cell carcinoma metastasis to the breast: mammographic, sonographic, CT, and pathologic correlation. Breast J 2007; 13:316–317.
573
SUROV ET AL
41. Kim SM, Ko EY, Han BK, et al. Metastatic thymoma of the breast. Korean J Radiol 2008; 9:80–83. 42. Francois C, Rangachari B, Bova D. Mammography and sonography of pathologically proven adrenal cortical carcinoma metastatic to the breast. AJR Am J Roentgenol 2005; 184:1279–1281. 43. Ashkar L, Mesurolle I, Trembley F. Sonographic target sign and rim enhancement on magnetic resonance imaging in metastatic melanoma to the breast. Australas Radiol 2006; 50:237–240. 44. Yang WT, Kwan WH, Chow LT, et al. Unusual sonographic appearance with color Doppler imaging of bilateral breast metastases in a patient with alveolar rhabdomyosarcoma of an extremity. J Ultrasound Med 1996; 15:531–533. 45. Tukel S, Dogan BE, Ozcan H. Alveolar rhabdomyosarcoma metastatic to the breast. Curr Probl Diagn Radiol 2003; 32:102–104. 46. Kalra N, Ojili V, Gulati M, et al. Metastatic choriocarcinoma to the breast: appearance on mammography and Doppler sonography. AJR Am J Roentgenol 2005; 184(suppl):S53–S55. 47. Upalakalin JN, Collins LC, Tawa N, et al. Carcinoid tumors in the breast. Am J Surg 2006; 191:799–805. 48. Jakovljevic B, Stevanovic O, Bacic G. Metastases to the breast from smallcell lung cancer: MR findings. A case report. Acta Radiol 2003; 44:485–488.
574
Academic Radiology, Vol 18, No 5, May 2011
49. Perlet C, Sittek H, Forstpointner R, et al. Metastases to the breast from rhabdomyosarcoma: appearances on MRI. Eur Radiol 1999; 9:1113–1116. 50. Kelkar PS, Helbich TH, Becherer A, et al. Solitary breast metastasis as the first sign of a squamous cell carcinoma of the cervix: imaging findings. Eur J Radiol 1997; 24:159–162. 51. Briest S, Horn LC, Haupt R, et al. Metastasizing signet ring cell carcinoma of the stomach-mimicking bilateral inflammatory breast cancer. Gynecol Oncol 1999; 74:491–494. 52. Birjawi GA, Haddad MC, Tawil AN, et al. Metastatic rhabdomyosarcoma to the breast. Eur Radiol 2001; 11:555–558. 53. Ho LW, Wong KP, Chan JH, et al. MR appearance of metastatic melanotic melanoma in the breast. Clin Radiol 2000; 55:572–573. € u €roglu M, Ersavasti G, Ilvan S, et al. Bilateral inflammatory breast 54. Ozg metastases of epithelial ovarian cancer. Am J Clin Oncol 1999; 22: 408–410. 55. Gomori JM, Grossman RI, Shields JA, et al. Choroidal melanomas: correlation of NMR spectroscopy and MR imaging. Radiology 1986; 158: 443–445. 56. Woodruff WW Jr, Djang WT, McLendon RE, et al. Intracerebral malignant melanoma: high-field-strength MR imaging. Radiology 1987; 165: 209–213.