Local Staging: Imaging Options and Core Biopsy Strategies

Published on 03/05/2015 by admin

Last modified 22/04/2025

Print this page

This article have been viewed 16810 times

CHAPTER 3 Local Staging: Imaging Options and Core Biopsy Strategies

Treatment planning for newly diagnosed breast cancer patients, including local and systemic therapies, is based on tumor type, extent of disease, and accurate staging. Imaging and image-directed needle biopsies play a critical role in establishing the local extent of disease and aid in staging. Surgical decision making, between breast-conserving therapy (also termed lumpectomy or segmental or partial mastectomy) and mastectomy, is primarily based on tumor size, extent and location within the breast, cosmetic implications, and patient preference. Imaging modalities, including mammography, ultrasound, MRI, and molecular imaging, are used to determine the extent of disease. However, disease extent cannot be reliably established solely by imaging. When preoperative imaging suggests more extensive disease than clinical impressions, histologic confirmation is necessary before performing a more extensive surgery.

Imaging is used to locally stage breast cancer. The TNM stage is based on the size of the tumor and whether the cancer has spread (Table 1). Identification of abnormal adenopathy (axillary, supraclavicular, or internal mammary nodes), as well as involvement of the skin, pectoralis muscle, or chest wall, affects staging and therapeutic decision making. Options for local therapies include surgery and radiation therapy, whereas systemic treatments may include chemotherapy, hormone therapy, and biologic therapy. Imaging also is used to identify distant metastases.

Table 1 AMERICAN JOINT COMMITTEE ON CANCER STAGING SYSTEM FOR PATIENTS WITH BREAST CANCER

Primary Tumor (T)
TX	Primary tumor cannot be assessed
T0	No evidence of primary tumor
Tis	Carcinoma in situ
Tis (DCIS)	Ductal carcinoma in situ
Tis (LCIS)	Lobular carcinoma in situ
Tis (Paget’s)	Paget’s disease of the nipple with no tumor
TI	Tumor 2 cm or less in greatest dimension
T1mic	Microinvasion 0 to 1 cm or less in greatest dimension
T1a	0.1 to 0.5 cm
T1b	>0.5 to 1 cm
T1c	>1 to 2 cm
T2	Tumor >2 to 5 cm in greatest dimension
T3	Tumor >5 cm in greatest dimension
T4	Tumor of any size with direct extension to chest wall or skin
T4a	Extension to chest wall, not including pectoral muscle
T4b	Edema (including peau-d’orange) or ulceration of the skin of the breast, or satellite skin nodules confined to the same breast
T4c	T4a and T4b
T4d	Inflammatory carcinoma
Regional Nodes (N)
NX	Regional lymph nodes cannot be assessed (e.g., previously removed)
N0	No regional lymph node metastasis
N1	Metastasis in movable ipsilateral axillary lymph nodes
N2	Metastasis in ipsilateral axillary lymph nodes fixed or matted, or in clinically apparent ipsilateral internal mammary nodes in the absence of clinically evident axillary lymph node metastasis.
N2a	Metastasis in ipsilateral axillary lymph nodes fixed to one another or to other structures
N2b	Metastasis only in clinically apparent ipsilateral internal mammary nodes and in the absence of clinically evident axillary lymph node metastasis
N3	Metastasis in ipsilateral infraclavicular lymph node(s) with or without axillary lymph node involvement, or in clinically apparent ipsilateral internal mammary node(s) in the presence of clinically evident axillary lymph node metastasis; or metastasis in ipsilateral supraclavicular lymph node(s) with or without axillary or internal mammary lymph node involvement
N3a	Metastasis in ipsilateral infraclavicular lymph node(s) and axillary lymph node(s)
N3b	Metastasis in ipsilateral internal mammary lymph node(s) nodes and axillary lymph node(s)
N3c	Metastasis in ipsilateral supraclavicular lymph node(s)
Distant Metastasis (M)
MX	Distant metastasis cannot be assessed
M0	No distant metastasis
M1	Distant metastasis

Selection of the appropriate imaging tests to determine tumor extent and stage are not standardized and are often based on the particulars of each case.

EXTENT OF DISEASE

Most breast cancers are evaluated by mammography, ultrasound, or both modalities. It is important to document the size, location, and distribution of the primary lesion, but also to evaluate for satellite lesions. Preoperative identification of additional lesions improves the likelihood of obtaining clear margins if breast-conserving therapy (BCT) is performed. For masses, it is important to look for associated microcalcifications, which may represent an associated noninvasive (in situ) component. The term extensive intraductal component (EIC) is used to refer to invasive tumors in which ductal carcinoma in situ (DCIS) makes up at least 25% of the neoplasm. Of all invasive ductal carcinomas, 15% to 30% have an EIC.¹ Tumors that are predominantly DCIS with focal invasion are also classified as EIC. The presence of EIC may have prognostic implications on the likelihood of obtaining clear margins, as well as the risk for subsequent local recurrence.^2,³

As many as 30% to 60% of breast cancers are pathologically multifocal (more than one tumor focus, separated by normal tissue) at the time of diagnosis.^4,⁵ The term multicentric has been variably defined. It generally has been used to describe cancers separated by more than 4 cm or tumors located in different quadrants of the breast. BCT generally is not suitable for multicentric carcinomas because of poor cosmetic results, limitations of radiation therapy, and inability to obtain clear margins.

Synchronous contralateral breast cancer may occur in about 3% to 5% of women with breast cancer.^6,⁷ Identification of these lesions at the time of the contralateral index cancer diagnosis can facilitate treatment in a single surgery, thereby avoiding both delays in diagnosis, as well as the emotional stress of a later diagnosis and second surgery.

MRI is useful to identify residual disease and direct re-excision in patients with positive margins at initial lumpectomy (Figure 1).

FIGURE 1 MRI evaluation can be useful in patients with close or positive surgical margins. A, Post-contrast, subtracted MRI of a right breast shows a focus of residual DCIS (arrow) along the medial aspect of a large lumpectomy cavity. B, Post-contrast, subtracted MRI shows residual invasive tumor (arrow) along the anterior lateral aspect of a small lumpectomy cavity.

EVALUATING ADENOPATHY

Identification of abnormal adenopathy, including axillary, supraclavicular, and internal mammary nodes, is important in staging. Metastatic adenopathy is suspected on imaging when there is cortical thickening (generally >3 mm), loss of the fatty hilum, and enlargement, particularly with increasingly round shapes ⁸ (Figure 2 and Figure 3). However, because many benign processes may cause reactive nodes with similar imaging findings, fine-needle aspiration (FNA) or core needle biopsy is necessary to confirm suspected metastatic nodal disease. Conversely, the absence of suspicious imaging findings, whether on mammography, ultrasound, MRI, or molecular imaging studies, does not exclude metastatic nodal involvement, particularly for micrometastasis. Therefore, in addition to imaging, sampling, either with axillary dissection or sentinel lymph node biopsy, is essential in the staging of breast cancer patients.

FIGURE 2 Sonographic appearance of normal axillary lymph nodes. A, Longitudinal image shows uniform thickness of the hypoechoic cortex, surrounding the echogenic fatty hilus. B, Transverse image of the same lymph node. The cortical thickness of 2 mm is within normal limits (<3 mm). C, Color Doppler image shows vascularity in the hilus.

FIGURE 3 The spectrum of abnormal axillary node sonographic findings (all FNA-proven to have breast cancer metastases). A, Cursors delineate borderline thickening of the hypoechoic cortex. The fatty hilus is preserved. A second, similar-appearing axillary lymph node is seen to the left. These findings are not clearly pathologic by sonographic criteria. Subtle findings on staging MRI, CT, and positron emission tomography (PET) studies (for inflammatory breast cancer) suggested axillary involvement. FNA confirmed metastatic carcinoma. B, Cortical thickening in this case is more eccentric and measures 5 mm. Ultrasound-guided FNA of the axillary lymph node confirmed metastatic carcinoma. C, Axillary node sonography of a 55-year-old woman with locally advanced breast cancer (LABC) shows highly suspicious morphology. The cortex is very hypoechoic, as well as abnormally thickened and nodular. There is a rat-bite, scalloped appearance and mass effect on the echogenic hilus. D, Same case as C, with color Doppler. Increased and abnormal vascularity is seen. Normal lymph node vascularity is seen only at the hilus. E, A 45-year-old woman with a new infiltrating ductal carcinoma (IDC) diagnosis. The cortical mantle of this axillary lymph node is markedly thickened, and the echogenic hilus is nearly completely effaced. F, Although small, this axillary node morphology is quite abnormal. The cortex is very hypoechoic. The node is “thick-waisted,” with nearly complete effacement of the fatty hilus, which is hinted at in profile. G, A 56-year-old woman with a neglected LABC (IDC with secondary inflammation) and multiple abnormal axillary nodes, which were FDG-avid on PET. This lymph node shows diffuse cortical thickening, without hilar effacement. H, Another lymph node of the same patient shows complete hilar loss and an abnormally rounded shape. I, A 41-year-old woman with a large postpartum pregnancy-associated breast cancer. This lymph node is massively enlarged and would be considered abnormal on any modality. Size, the primary criterion for judging normalcy on CT and MRI, is a weaker criterion by which to judge axillary nodes. Other morphologic criteria, including shape, cortical thickening and nodularity, and mass effect on or effacement of the fatty hilus, are more reliable in the assessment of the axillary lymph nodes of breast cancer patients. These more subtle findings are most easily assessed with ultrasound. This lymph node shows many of these features: marked cortical hypoechogenicity, nodularity, and thickening, with partial loss of the fatty hilus. The intracortical and peripheral vascularity is highly abnormal and also suggests malignancy. J, The same lymph node seen transversely shows that the hilus is incompletely effaced. Echogenic remnants are seen surrounded by the grossly thickened hypoechoic cortex. The abnormally increased vascularity, seen on the periphery of the cortex, is highly suggestive of malignancy. K, Even more normal-appearing adjacent lymph nodes, with visible echogenic fatty hila, have borderline thickened cortices and are notable for their ease of visualization and increased number. L, A 49-year-old woman with newly diagnosed, node-positive (five of eight nodes, largest 1.9 cm, with extracapsular extension), 1.8-cm left breast IDC. Her preoperative ultrasound predicted the extracapsular extension. This very hypoechoic axillary lymph node shows no echogenic hilus. The left side is rounded by a cortical nodule, and the right side has frankly angular margins. M, Oblique, noncontrast, T1-weighted MRI of the left axilla shows an abnormal lymph node (arrow), notable for the loss of the fatty hilum and the rounded and expanded shape.

The axillary nodes form a chain from the underarm to the collarbone (Figure 4). The axillary lymph nodes are named in relation to the pectoralis minor muscle, with level I the lowest, lateral to the pectoralis minor muscle. Level I receives the most lymphatic drainage from the breast. Level II axillary nodes are beneath the pectoralis minor muscle. Level III is above and medial to the pectoralis minor muscle. A traditional axillary lymph node dissection usually removes nodes in levels I and II. Sentinel lymph node sampling involves the mapping and removal of the first lymph node or nodes (usually 1 to 3) that drain the involved area of the breast (Figure 5 and Figure 6). Instead of removing 10 or more lymph nodes as performed in a standard dissection, the status of the axilla can be predicted by excision and close pathologic examination of the sentinel node. Sentinel lymph node biopsy has significantly reduced the number of women undergoing standard axillary dissection, avoiding dissection-associated side effects such as arm lymphedema. The identification of abnormal lymph nodes on physical exam or imaging studies favors proceeding directly to axillary dissection over sentinel node biopsy.

FIGURE 4 Schematic of axillary lymph node nomenclature. The axillary lymph nodes ascend from the axilla medially to the clavicle and are named in relation to the pectoralis minor muscle. Level I, the lowest, is lateral to the pectoralis minor; level II, deep to it; and level III, medial to it.

FIGURE 5 Schematic of the technique of sentinel lymph node injection. The breast lymphatics drain to the skin, and there is a rich periareolar network of lymphatics, with one or two major lymphatic channels draining to the axilla. Anatomic studies indicate 75% of the lymphatic drainage of the breast is to the axilla, with 25% to the internal mammary nodes. In this drawing, a peritumoral injection is depicted, but success has been reported for a variety of techniques (including subdermal or intradermal overlying the tumor, and periareolar), presumably owing to the rich network of the breast lymphatic plexus.

FIGURE 6 Lymphoscintigraphy (from left to right: anterior, left anterior oblique, and left lateral) can be performed in conjunction with injections for sentinel lymph node identification. In this example, intense activity is seen at the injection site of the tumor in the left breast. A lymphatic drainage channel extending to the axilla is clearly delineated, as are a faint, more superior secondary channel and the axillary sentinel lymph node. If a sentinel node is identified at imaging, skin marking can save surgical dissection time. Breast cancer sentinel node mapping with lymphoscintigraphy is not universally performed, unlike in melanoma.

The use of molecular imaging studies, CT, or MRI may identify adenopathy in areas other than the axilla (Figure 7). The presence of internal mammary node adenopathy (Figure 8) affects staging and radiation therapy planning. Identification of abnormal Rotter’s nodes (Figure 9), nodes between the pectoralis minor and major muscles, also has staging and therapeutic implications.

FIGURE 7 Schematic of the lymphatic drainage pathways of the breast. By anatomic studies, 75% of the lymphatic drainage is to the axilla, with 25% to the ipsilateral internal mammary nodes. Note lymphatic communications extending across the midline of the chest to contralateral internal mammary nodes and inferiorly extending channels, which communicate with the abdomen.

FIGURE 8 Internal mammary node metastasis. A, Axial postcontrast MRI in a patient with an implant with an abnormally enlarged left internal mammary node (arrow) adjacent to the internal mammary artery and vein, suspicious for involvement. B, Axial postcontrast MRI shows a small (2–3 mm) right internal mammary node (left). The same area, 5 years later, has a chest wall recurrence (arrow), presumably from metastasis to the internal mammary node (right).

FIGURE 9 Rotter’s nodes. A, Oblique T1-weighted (left) and axial fat-saturated, enhanced T1-weighted (right) MRIs of the left axilla show a small Rotter’s node (arrows) between the pectoralis major and pectoralis minor muscles. B, Enhanced CT scan demonstrates a metastasis to a large Rotter’s node (arrow), as well as within the right axilla.

SKIN, PECTORALIS, AND CHEST WALL INVOLVEMENT

Identification of breast edema and skin thickening in patients with invasive breast cancer may represent an inflammatory component (tumor involving the dermal lymphatics). This materially affects staging and therapeutic approach. Skin punch biopsy may be necessary to confirm the diagnosis if the clinical picture is not characteristic. Identification of pectoralis muscle or chest wall involvement also affects treatment planning. Pectoralis muscle involvement should be looked for in women with posterior lesions. The diagnosis on MRI requires not just effacement or obliteration of the pectoralis fascia but also enhancement of the muscle (Figure 10). Identification of either skin or chest wall involvement classifies a tumor as a locally advanced breast carcinoma (LABC). Large (>5 cm) tumors and those with clinically matted or fixed axillary node involvement or involved supraclavicular or internal mammary nodes by imaging are also considered locally advanced. LABC is generally treated with preoperative (neoadjuvant or primary systemic) chemotherapy, which converts some patients into operative candidates. Multiple examples are presented in Chapter 5.

FIGURE 10 Chest wall invasion. Postcontrast axial MRI in a patient with a breast sarcoma, demonstrating enhancement of the pectoralis major muscle and intercostal muscles (arrows). Chest wall involvement is confirmed by identification of intercostal muscle enhancement, whereas the pectoralis muscle is considered part of the breast.

PREOPERATIVE MAGNETIC RESONANCE IMAGING

MRI is becoming increasingly established as a useful modality in patients with known breast cancer. However, patterns of use of preoperative MRI in women with biopsy-proven breast cancer remain highly variable in practice because of the lack of randomized control trials. It has been firmly established that preoperative MRI can detect unsuspected disease and often changes management in women with known breast cancer.⁹ However, the use of preoperative MRI to alter surgical management has been criticized because of the lack of studies evaluating its effect on tumor recurrence and mortality.¹⁰^–¹² Fischer and colleagues,¹³ in a study involving more than 40 months of follow-up, reported a reduction in ipsilateral breast tumor recurrence, from 6.8% to 1.2%, in patients who underwent preoperative MRI. However, this study was a retrospective, singleinstitution review of only 346 patients. It has been argued that unsuspected disease detected on MRI and treated with BCT is of little clinical consequence and is controlled by radiation therapy. This argument is primarily based on the fact that 10-year local recurrence rates as low as 10% have been reported in patients with negative surgical margins.¹⁴^–¹⁶ Concerns about the use of preoperative MRI involve the potential for unnecessary biopsies, delays in treatment, and an increase in unnecessary mastectomies. Despite these arguments, there is ongoing evidence and experience accumulating that supports its benefit in preoperative staging, as follows:

1. Although no randomized controlled trials have been performed, preoperative MRI has the potential to reduce recurrences in many patients. Recurrence rates overall are very low; however, higher recurrence rates have been reported in select patients, such as those with high-grade DCIS.¹⁷ In addition, recurrence rates as high as 35% have been reported in younger women and in women who do not undergo radiation therapy.¹⁸ With the increasing use of partial-breast irradiation, it will be important to monitor the recurrence rates in patients thus treated, to compare with established rates for whole-breast irradiation. Preoperative MRI may prove essential to identify patients likely to fail partial-breast irradiation because of the presence of occult disease outside of the local radiation field.

2. Preoperative MRI may help to improve surgical outcomes. The previously reported low breast cancer recurrence rates following BCT are based on patients with negative surgical margins. In the United States, the number of women requiring multiple surgeries to obtain clear margins is significant. Positive margin rates at initial lumpectomy may be as high as 40%.¹⁹ By better defining tumor extent and providing three-dimensional information, MRI can potentially improve surgical outcomes and decrease positive margin rates.

3. In addition to better defining disease extent, preoperative MRI can improve treatment planning in other ways. By demonstrating abnormal axillary adenopathy, MRI can help to identify patients who may be better served by standard axillary dissection, rather than sentinel node biopsy. Surgical planning may also be altered in patients in whom MRI demonstrates pectoralis or chest wall involvement. MRI also has the ability to demonstrate abnormal internal mammary nodes, information that may be incorporated in radiation therapy planning.

4. Preoperative MRI has the ability to detect mammographically and clinically occult carcinoma in the contralateral breast in 3% to 5% of patients.^7,²⁰ Failure to detect these cancers at the time of initial diagnosis exposes the patient to potential risks associated with repeat general anesthesia and surgery, as well as possible delays in diagnosis. In addition, the psychological toll of dealing with a second cancer diagnosis in the opposite breast is not insignificant.

These benefits, although individually small, together have the potential to positively affect a significant number of patients. Preoperative MRI can improve surgical clear margin rates, guide surgical and radiation therapy planning, detect occult contralateral cancers, and potentially reduce recurrences. However, the number of false-positive results and delays in treatment needs to be minimized. Other breast imaging studies, such as mammography and ultrasound, must be correlated with the MRI findings to provide the most accurate interpretation and appropriate recommendations.

FUNCTIONAL (MOLECULAR) BREAST IMAGING: BREAST-SPECIFIC GAMMA IMAGING AND POSITRON EMISSION MAMMOGRAPHY

Functional breast imaging is a growing and evolving field that is assuming a larger role in breast cancer diagnosis, providing complementary information to anatomically based breast imaging modalities. Scintimammography has matured from initial versions using standard gamma cameras, which were limited in resolution and positioning flexibility, to breast-specific gamma imaging (BSGI), which obtains higher-resolution planar images in views emulating mammography.²¹^–²⁵ Similarly, positron emission mammography (PEM) is performed using a small-field-of-view, high-resolution PET scanner that resembles a mammogram unit and acquires tomographic data sets in planes analogous to mammography.²⁶^–²⁹ Scintimammography is performed after the intravenous administration of 25 mCi of either ^99mTc-sestamibi or ^99mTc-tetrofosmin, whereas PEM is performed after an hour’s uptake of 10 mCi of ¹⁸F-fluorodeoxyglucose (FDG) admin-istered intravenously. The radiation dose is about the same, about 0.4 rad. Imaging time is also similar. Scintimammography planar views are acquired for 5 to 10 minutes, or at least 150,000 counts, whereas PEM tomographic data sets are acquired for 4 to 10 minutes per projection. The lower limit of BSGI detector resolution is 3 mm (although smaller lesions may be identifiable), whereas the in plane resolution of PEM is on the order of 2 mm (Table 2).

Table 2 FUNCTIONAL BREAST IMAGING COMPARISON CHART

	BSGI	PEM
Radiopharmaceutical	^99mTc-sestamibi (Miraluma) (^99mTc-methoxyisobutylisonitrile)

¹⁸F-FDG (2-[fluorine-18] fluoro-2-deoxy-d-glucose Half-life 6 hr 110 min Emission energy 140 keV 511 keV Dose 25 mCi 10 mCi Sensitivity

85%–93% (scintimammography)

>90% (BSGI)

91% Specificity 87%–89% 93% Whole-body dosimetry About 0.4 rad About 0.4 rad Acquisition Planar Tomographic Binding target Intracellular mitochondria Intracellularly phosphorylated by hexokinase Theoretical basis Cancer cells have greater cytoplasmic mitochondrial density than normal breast Higher glycolytic rate of cancer cells results in increased cellular uptake and glucose utilization Negative predictive value (NPV)

82%–98% (scintimammography)

90%–100% (BSGI)

88% Target population Suspicion of breast abnormality (breast cancer diagnosis not required) Approved for patients with known or past history of breast cancer Mechanism of cellular transport Passive diffusion through potassium channels Active transport into cell Lower limit of resolution 3-mm detector spatial resolution 2 mm

BSGI, breast-specific gamma imaging; PEM, positron emission mammography.

The uptake of sestamibi is dependent on regional blood flow and cellular mitochondrial density. Enhanced blood flow due to tumorinduced neo-angiogenesis results in increased delivery of radiopharmaceutical. Cancer cells have higher cytoplasmic mitochondrial density than normal breast tissue and bind more of the radiopharmaceutical than the surrounding tissue.

PEM is performed with ¹⁸F-FDG, a radioactive glucose analogue in wide use as a cancer-imaging agent with whole-body PET. PEM and PET with FDG capitalize on the higher glycolytic rate of cancer cells, with FDG actively transported intracellularly through an up-regulated transmembrane GLUT-1 receptor mechanism. Once intracellular, FDG is phosphorylated by hexokinase like glucose, but because it is not metabolized further, it is trapped intracellularly in proportion to glucose utilization.

At this writing, BSGI is becoming more available, with sites with early experience finding it useful both for problem solving of ambiguous conventional breast imaging findings, and as an adjunct to local staging, looking for multifocality, multicentricity, and contralateral lesions (Figure 11). BSGI does not require a diagnosis of breast cancer for reimbursement, and increasingly is being performed in patients with suspicious conventional breast imaging findings as an aid to biopsy decision making (e.g., deciding how many areas need sampling). Currently, PEM is approved for patients with known or prior breast cancer diagnoses. Several cases incorporating the use of PEM have been included in this chapter. Its performance compared with MRI in preoperative local staging of apparently localized breast cancer is being assessed by a multicenter, prospective clinical trial, which at this writing is accruing patients. Early pilot studies show high sensitivity for depiction of primary breast cancers, on the order of 91%. Similarly high sensitivity for primary tumor depiction is reported for BSGI. Both modes of functional breast imaging appear to have improved specificity compared with MRI, on the order of 87% to 89% for BSGI and 93% for PEM, offering hope that increased use of functional breast imaging in the future may decrease the number of unnecessary benign biopsies now being performed.

FIGURE 11 An 89-year-old woman with a prior right upper outer quadrant benign biopsy (scar marked on mammographic views (arrows)—A, MLO; B, CC). Multiple lobular and circumscribed masses are seen within moderately dense breast tissue, especially in the upper outer quadrant near the scar. Ultrasound of the corresponding area suggested two cysts and one solid mass, for which biopsy was recommended. Breast-specific gamma imaging (BSGI) (C, MLO; D, CC) shows multifocal upper outer quadrant increased activity, with three discrete sites of increased uptake. At surgical pathology, a multilobed mucinous carcinoma, with ductal carcinoma in situ was confirmed.

(Case courtesy of Dilon Technologies)

Currently, no biopsy capability using functional imaging modalities for guidance is readily available, although feasibility studies have been done and this should be available in the near future.

DISTANT METASTASES

Routine imaging with bone scan, CT, and PET is not indicated for early-stage (stage I and II) breast cancers. Patients with limited-stage breast cancer who have symptoms, such as back or bone pain, or right upper quadrant pain; physical examination abnormalities (such as jaundice or hepatomegaly), or laboratory abnormalities can be considered for imaging to exclude distant metastatic disease. Patients with locally advanced breast cancer or evidence of axillary nodal involvement by physical examination or imaging should be considered for systemic imaging after a detailed history, physical exam, and laboratory evaluation.

REFERENCES

1 van Dongen JA, Fentiman IS, Harris JR, et al. In situ breast cancer: the EORTC consensus meeting. Lancet. 1989;2:25-27.

2 Schnitt SJ, Connolly JL, Recht A, et al. Breast relapse following primary radiation therapy for early breast cancer. II. Detection, pathologic features and prognostic significance. Int J Radiat Oncol Biol Phys. 1985;11:1277-1284.

3 Bartelink H, Borger JH, van Dongen JA, Peterse JL. The impact of tumor size and histology on local control after breast conserving treatment. Radiother Oncol. 1988;11:297-303.

4 Holland R, Veling SH, Mravunac M, Hendricks JH. Histologic multifocality of Tis, T1–2 breast carcinomas: implications for clinical trials of breast conserving surgery. Cancer. 1985;56:979-990.

5 Wilkinson LS, Given-Wilson R, Hall T, et al. Increasing the diagnosis of multifocal primary breast cancer by the use of bilateral whole-breast ultrasound. Clin Radiol. 2005;60:573-578.

6 Heron DE, Komarnicky LT, Hyslop T, et al. Bilateral breast carcinoma: risk factors and outcomes for patients with synchronous and metachronous disease. Cancer. 2000;88:2739-2750.

7 Liberman L, Morris EA, Kim CM, et al. MR imaging findings in the contralateral breast of women with recently diagnosed breast cancer. AJR Am J Roentgenol. 2003;180:333-341.

8 Yang WT, Chang J, Metreweli C. Patients with breast cancer: differences in color Doppler flow and gray-scale US features of benign and malignant axillary lymph nodes. Radiology. 2000;215:568-573.

9 Bedrosian I, Mick R, Orel SG, et al. Changes in the surgical management of patients with breast carcinoma based on preoperative magnetic resonance imaging. Cancer. 2003;98:468-473.

10 Morrow M. Magnetic resonance imaging in the preoperative evaluation of breast cancer: primum non nocere. J Am Coll Surg. 2004;198:240-241.

11 Morrow M. Magnetic resonance imaging in breast cancer: one step forward, two steps back? JAMA. 2004;292:2779-2780.

12 Morrow M. Magnetic resonance imaging in breast cancer: is seeing always believing? Eur J Cancer. 2005;41:1368-1369.

13 Fischer U, Zachariae O, Baum F, et al. The influence of preoperative MRI of the breasts on recurrence rate in patients with breast cancer. Eur Radiol. 2004;14:1725-1731.

14 Smitt MC, Nowels KW, Zdeblick MJ, et al. The importance of the lumpectomy surgical margin status in long-term results of breast conservation. Cancer. 1995;76:259-267.

15 Neuschatz AC, DiPetrillo T, Safaii H, et al. Long-term follow-up of a prospective policy of margin-directed radiation dose escalation in breast-conserving therapy. Cancer. 2003;97:30-39.

16 Obedian E, Haffty BG. Negative margin status improves local control in conservatively managed breast cancer patients. Cancer J Sci Am. 2000;6:28-33.

17 Provenzano E, Hopper JL, Giles GG, et al. Histological markers that predict clinical recurrence in ductal carcinoma in situ of the breast: an Australian population-based study. Pathology. 2004;36:221-229.

18 Borg MF. Breast-conserving therapy in young women with invasive carcinoma of the breast. Australas Radiol. 2004;48:376-382.

19 Smitt MCMD, Horst K. Association of clinical and pathologic variables with lumpectomy surgical margin status after preoperative diagnosis or excisional biopsy of invasive breast cancer. Ann Surg Oncol. 2007;14(3):1040-1044.

20 Lehman CD, Gatsonis C, Kuhl CK, et al. ACRIN Trial 6667 Investigators Group. MRI evaluation of the contralateral breast in women with recently diagnosed breast cancer. N Engl J Med. 2007;356(13):1295-1303.

21 Schillaci O, Buscombe JR. Breast scintigraphy today: indications and limitations. Eur J Nucl Med Mol Imaging. 2004;31(Suppl 1):S35-S45.

22 Brem RF, Schoonjans JM, Kieper DA, et al. High-resolution scintimammography: a pilot study. J Nucl Med. 2002;43:909-915.

23 Brem RF, Rapelyea JA, Zisman G, et al. Occult breast cancer: scintimammography with high-resolution breast-specific gamma camera in women at high risk for breast cancer. Radiology. 2005;237:274-280.

24 Brem RF, Fishman M, Rapelyea JA. Detection of ductal carcinoma in situ with mammography, breast specific gamma imaging, and magnetic resonance imaging: a comparative study. Acad Radiol. 2007;14(8):945-950.

25 Brem RF, Petrovitch I, Rapelyea JA, et al. Breast-specific gamma imaging with (99m) Tc-sestamibi and magnetic resonance imaging in the diagnosis of breast cancer: a comparative study. Breast J. 2007;13(5):465-469.

26 Levine EA, Freimanis RI, Perrier ND, et al. Positron emission mammography: initial clinical results. Ann Surg Oncol. 2003;10:86-91.

27 Rosen EL, Turkington TG, Soo MS, et al. Detection of primary breast carcinoma with a dedicated, large-field-of-view FDG pet mammography device: initial experience. Radiology. 2005;234:527-534.

28 Berg WA, Weinberg IN, Narayanan D, et al. High-resolution fluorodeoxyglucose positron emission tomography with compression (“positron emission mammography”) is highly accurate in depicting primary breast cancer. Breast J. 2006;12(4):309-323.

29 Tafra L, Cheng Z, Uddo J, et al. Pilot clinical trial of ¹⁸F-fluorodeoxyglucose positron-emission mammography in the surgical management of breast cancer. Am J Surg. 2005;190(4):628-632.

CASE 1 Mammography: Extent of disease

A 75-year-old woman presented with a palpable left breast mass. Mammography demonstrated a dense, spiculated breast cancer, corresponding to the palpable mass (Figure 1). Closer review of the mammogram showed a second abnormality distant from the palpable mass, with suspicious linear microcalcifications in the central left breast (Figure 2). Biopsy of the mass revealed invasive ductal carcinoma and stereotactic biopsy of the calcification found intermediate-grade ductal carcinoma in situ (DCIS). Because the two areas of disease involved a large portion of the left breast, mastectomy was performed.

FIGURE 1 CC mammographic view of the left breast: A dense spiculated mass is seen in the medial left breast, corresponding to the palpable finding.

FIGURE 2 Close-up view of the CC mammographic view: An area of suspicious linear microcalcifications is seen (arrows) in the central breast, distant from the palpable mass. These calcifications exhibit classic linear and casting features of DCIS.

TEACHING POINTS

Breast imagers need to constantly guard against “satisfaction of search.” By fully defining the extent of disease, the appropriate surgical therapy can be planned. The occurrence of synchronous ipsilateral or bilateral breast cancer is about 3% to 5%. For localized breast cancer, the 20-year follow-up reports of two pivotal prospective randomized trials demonstrated that survival after a lumpectomy with radiotherapy is equal to survival after mastectomy. However, these studies are based on patients in whom clear margins were obtained. By better defining the extent of disease preoperatively, the likelihood of obtaining clear margins at initial surgery can be improved. Pathologic margin status is the most important predictor of local recurrence after breast conservation with radiation. Even when clear margins are obtained, many patients have remaining foci of disease. The role of radiation therapy is to treat these areas of occult residual tumor. The risk for local recurrence may be diminished if the residual tumor burden after surgery is minimized. This is even more important in patients who undergo partial breast irradiation rather than whole breast irradiation.

SUGGESTED READINGS

Fisher B, Anderson S, Bryant J, et al. Twenty year follow-up of a randomized trial comparing total mastectomy, lumpectomy, and lumpectomy plus irradiation for the treatment of invasive breast cancer. N Engl J Med. 2002;347:1233-1241.

Holland R, Veling SH, Mravunac M, Hendriks JH. Histologic multifocality of Tis, T1–2 breast carcinomas: implications for clinical trials of breast-conserving surgery. Cancer. 1985;56(5):979-990.

Veronesi U, Cascinelli N, Mariani L, et al. Twenty year follow-up of a randomized study comparing breast conserving surgery with radical mastectomy for early breast cancer. N Engl J Med. 2002;347:1227-1232.

Wazer DE, Schmidt-Ullrich RK, Schmid CH, et al. The value of breast lumpectomy margin assessment as a predictor of residual tumor burden. Int J Radiat Oncol Biol Phys. 1997;38(2):291-299.

CASE 2 Use of ultrasound to find invasive disease within extensive microcalcifications; depiction of disease extent by breast MRI versus PEM versus whole-body PET

An asymptomatic 40-year-old woman had extensive new left lateral microcalcifications identified on screening mammography. These showed suspicious linear and branching pleomorphism on magnification views (Figure 1), and extensive ductal carcinoma in situ (DCIS) was suspected. Ultrasound was performed to determine whether an invasive component could be identified for biopsy. Five solid masses were identified in the left lateral breast, ranging up to 1.8 cm in size, as well as a suspicious abnormal axillary lymph node (Figure 2).

FIGURE 1 Lateral (A) and CC (B) magnification views show moderate parenchymal density. Diffuse pleomorphic microcalcifications are present throughout the left lateral breast (arrows).

FIGURE 2 Left lateral breast ultrasound, performed to search for invasive components within suspected extensive DCIS, identified five discrete masses, measuring up to 1.8 cm in size. (A) At 2 o’clock, an oval hypoechoic mass without a well-defined capsule shows tiny punctate foci of hyperechoic calcium. Similar foci within the surrounding parenchyma can also be appreciated (arrows). (B) At 2:30 to 3 o’clock, two additional, similar-appearing, solid, hypoechoic masses with ill-defined margins are seen.

Biopsies with ultrasound guidance of the two largest upper outer quadrant masses at 2 and 3 o’clock confirmed grade 3 infiltrating ductal carcinoma, high-grade DCIS, with comedonecrosis, estrogen receptor and progesterone receptor positive, and HER-2/neu negative, from both sites. Axillary lymph node fine-needle aspiration confirmed metastatic carcinoma.

The patient desired breast conservation, and so extensive staging studies were performed to assess the extent of disease. These included breast MRI (Figure 3), positron emission mammography (PEM) (Figure 4 and Figure 5), and whole-body positron emission tomography (PET) (Figure 6). These studies confirmed multicentric disease with axillary nodal involvement, but showed no distant metastases. Neoadjuvant chemotherapy was given.

FIGURE 3 Axial breast MRI enhanced, subtracted maximal intensity projection view shows diffuse, confluent enhancement of the left lateral breast, corresponding to the distribution of pleomorphic DCIS calcifications on mammography. Within this field can be seen a variety of intensely enhancing masses. The largest display rim enhancement and washout. Eight discrete masses could be delineated on MRI in the left upper and lower outer quadrants. An enlarged left axillary lymph node can also be seen here.

FIGURE 4 Digital mammographic views and comparable PEM views, for correlation: (A) MLO mammogram. (B) MLO “volume” PEM image. Seven discrete foci of intense increased FDG uptake are seen. (C) CC mammogram. (D) CC “volume” PEM image. The most posterior of the seven lesions seen in the MLO projection is not included in the field of view and is not seen. Lower-level uptake of FDG is seen in the upper outer breast, corresponding both to the distribution of parenchyma and the extensive microcalcifications. The mammographic parenchymal density conceals the masses which could be identified by ultrasound, MRI, and PEM.

FIGURE 5 PEM data are a volumetric, tomographic acquisition and are divided into 12 slices: (A) MLO projection. PEM progresses from medial (upper left) to lateral (lower right). With windowing, the largest lesion, projecting above the nipple, shows a rim pattern of uptake. On the edge of the field of view (middle row, upper left, arrow), hypermetabolic axillary lymph nodes can be seen. (B) CC projection PEM is displayed from cranial (upper left) to caudal (lower right).

FIGURE 6 Fused whole-body PET/CT images show three hypermetabolic foci in the left lateral breast as well as three hypermetabolic left axillary lymph nodes. No evidence of distant metastatic disease was found.

TEACHING POINTS

Extensive, new, suspicious microcalcifications suggested an extensive intraductal component in this case. Stereotactic biopsy of microcalcifications undoubtedly would have established a diagnosis of DCIS. However, ultrasound can be useful to search for a mass for ultrasound sampling within an area of suspicious microcalcifications, which may enable a diagnosis of invasive disease to be made.

This is a very extensive, multicentric malignancy with an extensive intraductal component. It is interesting to compare the performance of the various modalities in depicting the extent of this neoplasm. The mammogram best depicts the extensive calcified DCIS, diffusely involving the upper and lower outer quadrants, but the breast tissue density obscured the invasive components. Five masses (two proven invasive) were identified by ultrasound within the area encompassed by the microcalcifications. The microcalcifications can be recognized sonographically, but would likely not be recognized prospectively without mammographic correlation.

Both invasive and noninvasive disease components are well depicted by MRI. The DCIS manifested as confluent, segmental enhancement in the lateral breast. Eight intensely enhancing discrete masses with washout could be identified by MRI within this extensive intraductal component. PEM performed nearly as well, depicting seven discrete sites of intense increased fluorodeoxyglucose (FDG) uptake. The intraductal disease (as represented by the distribution of calcifications on mammography) would be difficult to recognize prospectively on PEM, without mammographic correlation. The lower-level segmental FDG uptake seen here in the distribution of the DCIS calcifications also corresponds to the distribution of parenchymal density and so would be difficult to differentiate from background parenchymal uptake.

PEM is essentially a small field of view, high-resolution dedicated breast PET scan. Tomographic volumetric data is acquired with gentle compression (for immobilization, not for tissue thinning) in planes analogous to mammography. Because the detectors (in compression plate-like arrays) on either side of the breast are extremely close to the radioactive source (FDG taken up by glycolytically active tumor), the resolution is considerably higher than in whole-body PET, wherein a patient’s body is surrounded by a ring of detectors. State-of-the-art whole-body PET scanners have a lower limit of resolution today of about 6 mm, whereas resolution on the order of 2 mm can be expected with PEM. This is illustrated in this case. Only the three largest foci of FDG uptake could be visualized within the breast on whole-body PET, as compared with seven discrete lesions on PEM. A limitation of PEM is also illustrated here. Two hypermetabolic axillary nodes are at least partially visualized on PEM. Axillary visualization on PEM is variable, depending on patient anatomy and positioning. Whole-body PET in this patient readily demonstrated three hypermetabolic axillary nodes.

CASE 3 MRI: Extent of disease

A 54-year-old woman presented with a palpable mass in the upper outer left breast. Diagnostic mammographic and ultrasound evaluations demonstrated multiple masses in the upper outer quadrant of the left breast at the site of the palpable abnormality (Figure 1 and Figure 2). Core needle biopsy confirmed invasive ductal carcinoma. Preoperative MRI suggested much more extensive involvement of the left breast, with multiple enhancing masses extending from area of the known cancer toward the nipple (Figure 3). Because of the MRI findings, second-look ultrasound was performed and identified several small masses in the subareolar region (Figure 4). Core needle biopsy confirmed the extensive nature of the patient’s disease, and she was treated surgically with mastectomy and axillary dissection.

FIGURE 1 MLO mammographic view of the left breast: Multiple masses (arrows) are seen in the upper outer quadrant at the site of the patient’s palpable complaint.

FIGURE 2 Ultrasound evaluation of the left breast axillary tail region: A dominant mass is seen corresponding to the patient’s palpable lump. The mass is highly suspicious for malignancy with its irregular shape and margin features.

FIGURE 3 Postcontrast maximum intensity projection MRI: Multiple enhancing masses are demonstrated involving a large portion of the left breast, extending from the left axillary tail to the subareolar region (arrows).

FIGURE 4 Ultrasound image of the subareolar left breast: A lobular, hypoechoic mass (arrow) corresponds to the MRI finding in this location.

TEACHING POINTS

This case illustrates the use of MRI to evaluate the extent of disease in a patient with a known carcinoma. MRI may detect areas of additional disease unsuspected on mammography and ultrasound. In this case, MRI allowed targeting of ultrasound to identify more subtle additional lesions that could then undergo biopsy to establish the extent and aid in planning the appropriate surgical treatment.

CASE 4 Multicentric IDC and DCIS: Local staging with MRI

A 47-year-old woman was evaluated with mammography and ultrasound for a palpable lump in the right breast upper outer quadrant.

Mammography showed very dense breast tissue, with no correlate for the palpable abnormality (Figure 1). Ultrasound of the palpable lump showed a 2-cm heterogeneous, solid mass, with irregular margins and vascularity and a highly suspicious appearance (Figure 2). An ultrasound-guided core needle biopsy confirmed infiltrating ductal carcinoma (IDC), with high-grade ductal carcinoma in situ (DCIS). At initial surgical consultation, breast conservation therapy with partial mastectomy and radiation was discussed. Because of the mammographic density of the patient’s breasts, the patient was referred for preoperative breast MRI to more fully evaluate her suitability for breast conservation therapy.

FIGURE 1 Mammography [MLO (A), CC (B), and spot compression (C)] shows a marker placed at the level of the palpable lump in the right upper outer quadrant. The breast tissue is very dense, and no correlate is seen.

FIGURE 2 Ultrasound of the right breast palpable lump shows a mildly hypoechoic, solid mass, with lobular and irregular borders, with peripheral vascular flow on color Doppler. Several tail-like areas of suspected ductal extension are depicted.

Bilateral enhanced subtracted breast MRI showed the known right breast carcinoma mass to be intensely enhancing, with irregular margins and spiculation. Multiple smaller, additional foci of enhancement were noted throughout the right breast, markedly asymmetric compared with the left side (Figure 3 and Figure 4). A few of these foci were larger and more morphologically concerning, including an irregular mass at 5 o’clock (Figure 5) and clumped contiguous foci of enhancement at 9 o’clock (Figure 6). A second-look ultrasound was performed of the right breast seeking correlates for biopsy, to prove the patient’s disease was multicentric and that she was not a conservation candidate.

FIGURE 3 Maximal intensity projection from the enhanced, subtracted dynamic breast MRI series shows marked asymmetry between sides. The left shows no concerning finding. The right shows an irregularly marginated, dominant, intensely enhancing mass in the upper outer quadrant. Multiple smaller nodules are noted throughout the breast as well, larger than the tiny (<5 mm) foci frequently seen with physiologic and fibrocystic enhancement.

FIGURE 4 A section from the dynamic enhanced breast MRI series through the upper outer quadrants shows the known right infiltrating ductal carcinoma, with intense peripheral rim enhancement. The margins are irregular, with tiny spiculations.

FIGURE 5 Another enhancing, irregularly bordered mass is seen just below the level of the nipple (partially visualized here). This was reported as being at 12 o’clock. However, the sonographic correlate (see Figure 7) was best seen at 5 o’clock. This discrepancy illustrates the difficulty in lesion localization between modalities, which is due both to the mobility of breast tissue and differences in positioning. This patient had very small breasts, and this is a centrally positioned lesion.

FIGURE 6 Another section from the dynamic enhanced MRI series shows additional concerning findings in the right lateral breast, at about 9 o’clock. Irregular clumped enhancement, with linear extension anteriorly toward the nipple, suggests additional disease.

Ultrasound identified two subtle correlates for biopsy, confirming DCIS at both sites (Figure 7 and Figure 8). With pathologic confirmation of multicentric disease, the surgery treatment was changed to mastectomy.

FIGURE 7 Ultrasound of the right breast at 5 o’clock: This lesion was identified using the MRI as a roadmap and was not recognized prospectively. The correlate is subtle, being minimally hypoechoic. Once located, there are some suspicious sonographic features, including being taller than wide and having marginal angularity. Core needle biopsy confirmed DCIS.

FIGURE 8 Targeted right breast second-look ultrasound at 9 o’clock showed a very subtle correlate. A, A lobular, minimally hypoechoic focus seemed to correlate in shape and location to the MRI findings. B, Because of the lack of conspicuity of this area, a larger-volume, vacuum-assisted sampling was performed, also confirming DCIS. The needle was maneuvered underneath the target. With pathology-proven multicentricity, the surgical treatment was changed to mastectomy.

Final pathology was a 2.1-cm IDC, with associated high-grade (comedo) DCIS extending into lobules. The tumor was noted to be multicentric, with 50% of the gross tumor at the index cancer site, an 8-mm residual focus at 5 o’clock, and microscopic residual at 9 o’clock. Margins were negative, and two sentinel lymph nodes were also negative. Final stage was stage II, T2N0, and the patient was additionally treated with chemotherapy.

TEACHING POINTS

This case illustrates well how a multimodality approach can yield critical additional information to most accurately stage a newly identified breast cancer. Mammography was not particularly useful in this patient, owing to breast density, but confirmed that there were no microcalcifications. Prospectively, ultrasound identified the index tumor and provided guidance to confirm the diagnosis of IDC. However, no additional disease sites were recognized prospectively on ultrasound to suggest multicentric disease. However, with the breast MRI as a roadmap for a directed, second-look ultrasound, more subtle correlates could be discerned and targeted for sampling, confirming two separate sites of DCIS remote from the IDC, indicating that the patient’s disease should be treated surgically with mastectomy.

This case also illustrates ductal extension on ultrasound, a sonographic sign suspect for malignancy, as delineated by Stavros and colleagues. This feature is less frequently seen or recognized than other features of malignancy, such as taller-than-wide shape and shadowing.

SUGGESTED READING

Stavros AT, Thickman D, Rapp CL, et al. Solid breast nodules: use of sonography to distinguish between benign and malignant lesions. Radiology. 1995;196:123-134.

CASE 5 Additional disease site identified by PEM

A 46-year-old woman was noted to have a suspicious 1-cm spiculated mass overlying the right pectoral muscle on baseline screening mammography (Figure 1). It was confirmed on spot compression views, and a suspicious sonographic correlate was found (Figure 2), as well as a second concerning ultrasound finding. Both were biopsied with ultrasound guidance, confirming infiltrating ductal carcinoma (IDC) at 12 o’clock and sclerosing adenosis at 10 o’clock.

FIGURE 1 Baseline screening mammogram [MLO (A) and CC (B)] shows moderate parenchymal density. An irregularly marginated mass is seen overlying the pectoral muscle (circled and arrow on MLO view) and on the edge of the field in the CC projection (arrow).

FIGURE 2 Ultrasound of the right breast identified two abnormalities. A, At 12 o’clock, corresponding to the mammographic abnormality, there is a taller-than-wide, hypoechoic, solid, irregularly marginated, suspicious mass. Ultrasound-guided biopsy proved IDC. B, At 10 o’clock, a hypoechoic mass with ill-defined margins and no capsule is indeterminate, and also a concerning finding. Ultrasound-guided biopsy identified sclerosing adenosis.

The patient desired breast conservation and appeared by conventional imaging to be a suitable candidate. She was enrolled in a clinical trial prospectively comparing breast MRI to positron emission mammography (PEM) in preoperative staging of newly diagnosed breast cancer.

PEM identified two fluorodeoxyglucose (FDG)-avid abnormalities in the right breast. One corresponded to the known IDC. The other showed a linear and ductal pattern of uptake and was suspicious for a second site of malignancy, but did not clearly correspond to the second ultrasound abnormality in location or morphology (Figure 3).

FIGURE 3 PEM images of the right breast [MLO (A) and CC (B)]. A, MLO views show two FDG-avid abnormalities. Regions of interest have been placed around these (blue squares). The scintigraphically active mass in the upper posterior breast corresponds to the known IDC. The second site, projecting above the nipple at mid-breast level, shows suspicious intensity and morphology, with a linear and ductal configuration. This did not seem to clearly correlate in either location or morphology to the sclerosing adenosis lesion known from ultrasound at 10 o’clock. B, Both of the abnormalities seen in the MLO projection are less clearly seen in the CC projection. The known IDC is so far posterior that it is not included in the field of view on the CC projection and can only be suspected (arrowheads). The arrow indicates the location suspected to correspond to the linear/ductal PEM abnormality seen in the MLO projection.

MRI was performed subsequently. The known IDC formed a bilobed mass, but no additional abnormality was recognized prospectively (Figure 4A). When this study was then correlated with PEM, a possible correlate was recognized as localized clumped enhancement centrally, with late appearance of a 1-cm mass (see Figure 4B).

FIGURE 4 Breast MRI: A, Axial maximal intensity projection view shows the right posterior known IDC as a bilobed mass (arrow). No additional abnormality was noted prospectively. After correlation with PEM, the central region of clumped foci of enhancement (circled) was suspected to correspond with the second PEM abnormality. B, Sagittal, delayed (5 minutes after contrast), enhanced, fat-saturated image shows the clumped area of enhancement, which seemed to correlate in location with the second PEM abnormality (measured). C, Sagittal, enhanced, subtracted localizer image at the time of breast MRI biopsy showed a more convincing correlate for the linear-ductal PEM abnormality. A wishbone-shaped region of branched linear-ductal enhancement (circled) was targeted for MRI-guided vacuum-assisted sampling, confirming high-grade DCIS. This site was 8 cm away from the known IDC, indicating multicentric disease.

MRI-guided biopsy was recommended. At the time of biopsy, sagittal localization scans showed a clearer, branched, linear pattern of abnormal enhancement, which correlated with the morphology and location of the second PEM abnormality (see Figure 4C). Biopsy was performed with MRI guidance, confirming high-grade ductal carcinoma in situ (DCIS) as well as a complex sclerosing lesion and atypical ductal hyperplasia. This site was 8-cm away from the known IDC. Mastectomy was recommended but refused by the patient.

Double lumpectomy was performed after triple-needle localization. The 12-o’clock IDC was localized with ultrasound guidance. Ultrasound guidance was also used to place two additional needles to bracket the second abnormality. One was placed at the lateral margin of the post-MRI biopsy cavity, and the other was placed at the medial margin of this cavity, where the 10-o’clock sclerosing lesion was seen.

The 12-o’clock specimen removed a 1.2-cm IDC with DCIS. Margins were positive for DCIS inferiorly and posteriorly. The 9-o’clock specimen contained a 1.2-cm high-grade DCIS lesion as well as a complex sclerosing lesion and biopsy site changes. The medial margin was positive.

Mastectomy was again recommended, and declined. Re-excision of the positive margins was again positive for DCIS. She underwent a total of four re-excision procedures from both sites for recurrently positive margins, before undergoing mastectomy.

TEACHING POINTS

PEM is an emerging modality that functionally images breast tissue. PEM is essentially a highresolution, small-field-of-view, dedicated breast PET scan, performed after intravenous administration of FDG and an hour’s uptake time. It is performed utilizing a device resembling a mammogram unit, with detector arrays in place of compression plates, which are positioned on either side of the breast. Views in projections analogous to mammographic views can be obtained, allowing for ready correlation with mammograms. The breast is compressed to immobilize it, but not to the degree required to thin the tissue for optimal mammography. A volume of data is acquired tomographically. At this writing, the only U.S. Food and Drug Administration–approved device is manufactured by Naviscan PET Systems, but analogous PET and gamma camera–based devices by other manufacturers are in various stages of development. The Naviscan PEM tomographic acquisition is divided into 12 slices.

In this case, the PEM scan was the first study to indicate that this patient had multicentric disease. The MRI correlate was not recognized prospectively and was subtle to identify even with PEM correlation. PEM-guided biopsy capability is expected in the future and would provide an alternative means of staging in a case like this, when correlation is uncertain with conventional imaging or MRI.

The multicenter prospective trial in which this patient was enrolled is accruing patients at this writing. It is hoped that the data will provide valuable information on the performance of breast MRI compared with PEM in preoperative staging of breast cancer. The precise role that PEM will play in the breast imaging armamentarium is still being defined. Pilot studies suggest sensitivity for breast cancer identification is on a par with MRI, perhaps with greater specificity, but this remains to be established. Also unclear at this point is whether PEM could function as a stand-alone modality, or whether it is best used in conjunction with conventional breast imaging correlation.

SUGGESTED READINGS

Berg WA, Weinberg IN, Narayanan D, et al. High-resolution fluorodeoxyglucose positron emission tomography with compression (“positron emission mammography”) is highly accurate in depicting primary breast cancer. Breast J. 2006;12(4):309-323.

Tafra L. Positron emission tomography (PET) and mammography (PEM) for breast cancer: importance to surgeons. Ann Surg Oncol. 2007;14:3-13.

Tafra L, Cheng Z, Uddo J, et al. Pilot clinical trial of ¹⁸F-fluorodeoxyglucose positron-emission mammography in the surgical management of breast cancer. Am J Surg. 2005;190(4):628-632.

CASE 6 Subtle axillary nodal involvement

A 52-year-old woman presented with several months of left nipple change, with retraction and infra-areolar fullness. She had had tenderness and nodularity in this breast for more than a year. Her gynecologist confirmed an abnormal clinical exam and referred her to surgery, where the left side was noted to be asymmetrically enlarged, with faint peau d’orange of the periareolar area. A full, irregular, and retracted appearance of the nipple-areolar complex was noted. A large, firm, nodular mass encompassed the central breast on palpation, including the nipple-areolar complex. An axillary lymph node could be palpated. Palpation-guided core needle biopsy of the mass was performed, returning a diagnosis of infiltrating ductal carcinoma (IDC) with focal high-grade ductal carcinoma in situ (DCIS). A punch biopsy of the skin identified carcinoma in dermal lymphatics, confirming the clinical diagnosis of inflammatory breast cancer. The imaging evaluation included mammography, ultrasound, and breast MRI. Mammography showed slightly increased trabecular thickening and breast density. Sonography showed a hypoechoic, solid, irregular, highly vascular mass involving the nipple and retroareolar region (Figure 1). Multiple, sub-centimeter adjacent hypoechoic nodules were noted, and there was diffuse skin thickening. The axilla showed several lymph nodes, without clearly pathologic cortical mantle thickening or other particularly suspicious alterations in morphology prospectively (Figure 2).

FIGURE 1 Ultrasound of the retroareolar region shows a solid, hypoechoic, highly vascular mass with irregular angular margins.

FIGURE 2 Sonography of the axilla shows two normal-sized lymph nodes. Fatty hila are readily identified, and the cortical mantles are not clearly pathologically thickened or nodular.

The pretreatment staging of this locally advanced inflammatory breast cancer, with clinical and pathologic proof of skin involvement, was completed with breast MRI, body positron emission tomography (PET)/CT, and contrast-enhanced diagnostic chest, abdomen, and pelvis CT scans. Breast MRI showed a diffusely infiltrating, intensely enhancing central left breast mass, extending to and retracting the nipple (Figure 3). Breast MRI, PET/CT (Figure 4), and contrast-enhanced chest CT (Figure 5) also showed subtle but suspicious findings suggesting left axillary nodal involvement. Axillary fine-needle aspiration (FNA) of an axillary lymph node with only mild cortical mantle thickening was performed and confirmed metastatic carcinoma.

FIGURE 3 A, Axial maximal intensity projection view of both breasts from enhanced dynamic breast MRI series shows marked asymmetry, with intense central left breast enhancement, as well as asymmetrical nipple enhancement. An occasional, sub-centimeter, small, benign-appearing focus of enhancement is noted on the right. B, Axial short tau inversion recovery (STIR) image of both breasts shows marked left skin thickening and edema as well as diffuse increased edema of the entire breast. C, Axial, subtracted images of the breasts, from above to below (1–3), showing the diffusely infiltrating enhancing mass, with intense nipple enhancement and skin thickening and enhancement. B, Sagittal, enhanced, subtracted view (obtained about 5 minutes after contrast injection) shows diffuse infiltrative parenchymal enhancement, extending to the nipple. Enhancement of the thickened skin can be seen. E, Coronal STIR images through the thorax (1 and 2), obtained with the body coil as part of the breast MRI, show several left axillary lymph nodes. Although not clearly pathologically enlarged in size (1 cm), they are notable for the asymmetry from the opposite side. Hyperintensities in the liver are cysts.

FIGURE 4 A, Sagittal images from PET/CT (from left to right: CT, attenuation-corrected PET, fused PET/CT, non-attenuation-corrected PET) show hypermetabolism of the central left breast mass and nipple. B, PET images through the axilla (from left to right: coronal, sagittal, axial, and coronal projection volume image) show a small focus of mildly increased metabolic activity at the left axillary level, suggesting nodal involvement.

FIGURE 5 Axial contrast-enhanced chest CT images (A and B) through the axilla show an asymmetrical increased number of identifiable left axillary lymph nodes. Individual lymph nodes are not remarkably enlarged (a fatty hilus can still be seen in the largest) but are larger than the right-sided findings. In addition, there is subtle smudging of fat in the left axilla.

TEACHING POINTS

The imaging findings of involved axillary or regional lymph nodes can be subtle on both anatomic and functional imaging modalities. Relying on the traditional anatomic criteria of size allows us only to suggest disease when lymph nodes are clearly pathologically enlarged. Dimensions defining normalcy have been suggested previously for the axilla of less than 1 cm, or 5 mm in short axis. However, use of dimensional criteria alone will underestimate nodal involvement, which could be suspected based on other findings, such as asymmetry from the opposite, asymptomatic side. Either an asymmetrical increase in number of lymph nodes compared with the normal side, or larger (but still “normal”) size of nodes than on the opposite side, could be an imaging manifestation of nodal involvement. Extranodal extension of disease is suggested by fuzziness of nodal margins or infiltration of axillary fat. This case illustrates these more subtle findings of axillary involvement well. The PET scan also was abnormal, but not floridly so, with modest metabolic activity discernible in a normal-sized lymph node.

Preservation of the fatty hilus of a lymph node should not be regarded as complete reassurance regarding the status of that node. It is important to realize that mass effect and scalloping of the hilus, and ultimately replacement and complete effacement of the fatty hilus, are late manifestations of extensive lymph node involvement. Before these signs develop, cortical thickening, with or without nodularity, may signify a potentially abnormal lymph node and serve as a target for sampling with fine-needle aspiration.

CASE 7 Breast MRI problem solving: Deciding among sites for additional sampling

A 63-year-old woman was noted to have new right upper outer quadrant (UOQ) nodularity and increasing microcalcifications on screening mammography. Previously, she had undergone two benign stereotactic left breast biopsies for microcalcifications. Magnification spot compression confirmed the edge of a partially visualized mass as well as architectural distortion. Multiple clusters of microcalcifications were noted. At least three of these seemed to be progressive compared with prior studies, two in the right UOQ and one in the upper inner quadrant. Ultrasound of the right UOQ showed a highly suspicious, vascular, solid, hypoechoic mass at 10 o’clock (Figure 1), which correlated with the partially visualized mammographic mass.

FIGURE 1 Ultrasound of the right UOQ shows a highly suspicious, vascular, solid, hypoechoic mass with ill-defined margins at 10 o’clock. Ultrasound-guided biopsy confirmed IDC.

Core needle biopsy of the right UOQ mass with ultrasound guidance confirmed invasive ductal carcinoma (IDC), with associated calcifications. Because of the multiplicity of microcalcification clusters and prior history of two benign contralateral biopsies for microcalcifications, breast MRI was recommended after biopsy confirmation of UOQ malignancy to see where additional sampling should be undertaken, if breast conservation was considered.

Breast MRI demonstrated the known right UOQ carcinoma at 10 o’clock, with persistent enhancement (Figures 2, 3, and 4). A separate site of clumped enhancement was seen in the lower outer quadrant, 6 cm inferior to the known IDC, which subsequently underwent biopsy with MRI guidance (Figure 5). This proved to be ductal carcinoma in situ (DCIS), intermediate to high grade. This site did not correspond in location to the indeterminate microcalcifications.

FIGURE 2 Axial maximal intensity projection view from dynamic, contrast-enhanced MRI shows a dominant, heterogeneously enhancing mass in the right lateral breast, corresponding to the known IDC in Figure 1. An oblong focus of irregular enhancement is seen in the posterior lateral right breast (arrow), against a background of bilateral diffuse scattered enhancing foci.

FIGURE 3 Enhanced views [axial (A) and sagittal (B)] of the right UOQ known IDC. Margins are irregular and spiculated, and there is rim enhancement, MRI features of malignancy.

FIGURE 4 A second oblong small mass is seen in the lateral posterior LOQ right breast [axial (A) and sagittal (B)]. Biopsy was performed with MRI guidance, confirming DCIS.

FIGURE 5 Images from the MRI-guided biopsy of the right LOQ mass seen in Figure 4. A, A lateral localizing grid has been incorporated into the dedicated breast coil. Two fiducials are seen (arrows), as well as the hypointense obturator (seen on end, arrowhead) placed to the prescribed depth of the LOQ abnormality. Note how posterior the lesion is; if it had been any further posterior, it might not have been possible to access with this technique. B, Fat-saturated sagittal enhanced imaging after placement of the obturator shows it (arrowhead) in the LOQ. The known UOQ spiculated IDC is also seen on this slice. C, Fat-saturated axial enhanced imaging after placement of the obturator shows it in position (arrowheads). D, Repeat sagittal imaging, after the biopsy, shows a small fluid collection (postbiopsy hematoma) at the site. E, Repeat axial imaging, after the biopsy and clip deployment, shows a new hypointensity within the postbiopsy hematoma, representing the clip. F, Post-MRI-guided biopsy digital mammogram [90-degree lateral (1) and CC (2)] shows dense breast parenchyma and the MRI-placed clip in the posterior LOQ, 6 cm away from the UOQ IDC (asterisks), which is best seen as density and spiculation in the upper breast on the lateral view.

With pathologic proof of multicentric disease established, right mastectomy was recommended. The patient ultimately opted to undergo prophylactic left mastectomy at the same time.

The pathology showed a right 2.1-cm UOQ IDC, estrogen receptor and progesterone receptor positive, with DCIS associated with the invasive cancer and in a separate 6-mm lower outer quadrant (LOQ) focus. The margins were negative, and three sentinel lymph nodes and one intramammary lymph node were negative. The left mastectomy specimen showed atypical ductal hyperplasia and one negative sentinel lymph node.

There was no indication for radiation therapy, and the patient elected not to have chemotherapy. She had an Oncotype DX score of 20, indicating intermediate to low risk (13% risk for recurrence in 10 years). She was started on anastrozole (Arimidex) but switched to tamoxifen 8 months later due to joint pain.

TEACHING POINTS

When this patient’s UOQ IDC was identified, she was also seen to have at least three progressive clusters of microcalcification. In considering breast conservation, it would have been necessary to sample one or more of these microcalcification clusters to assess their significance. Breast MRI was obtained to see if it might help in selecting a second site for biopsy, which might avoid the need for performance of a double or triple stereotactic biopsy of the similar-appearing microcalcifications. The breast MRI did identify a LOQ suspicious abnormality, unrelated to the microcalcifications, which were all in upper quadrants. This LOQ oblong mass proved to be DCIS, indicating that the patient was best treated with mastectomy, and avoiding the need to perform a battery of stereotactic biopsies for diagnosis of the microcalcifications.

CASE 8 Breast MRI problem solving: Assessing depth of involvement of posterior breast cancer

A 47-year-old premenopausal woman noted a palpable, growing left breast mass, which was confirmed mammographically as a deep central mass. Stereotactic biopsy at another facility made the diagnosis of infiltrating ductal carcinoma (IDC). By clinical exam, the mass was on the order of 4 cm in size, occupying much of the upper outer quadrant, and seemed to be affixed to the chest wall. By ultrasound, the mass appeared to involve pectoral muscle (Figure 1).

FIGURE 1 Ultrasound [radial (A) and antiradial (B)] shows a hypoechoic, solid, irregularly marginated mass, which extends into horizontally oriented fibers of the pectoralis muscle posteriorly.

Breast MRI was requested to assess the relationship of the mass to the chest wall. It confirmed extension of the mass into the pectoral muscle, without chest wall involvement (Figure 2 and Figure 3). The mass enhanced intensely, had lobular margins, and displayed washout on kinetic analysis.

FIGURE 2 Axial breast MRI shows an intensely enhancing, posterior lateral left breast mass, which clearly involves the pectoralis muscle. No chest wall involvement was seen. A, Axial T2. B, Axial STIR. C, Axial maximal intensity projection from enhanced dynamic series (minute 1). D, Enhanced subtracted single slice from dynamic series.

FIGURE 3 Sagittal views better show the depth of posterior muscular involvement. A, Fat-saturated, unenhanced, T1-weighted gradient echo. B, Subtracted, enhanced view shows the enhancing mass, extending into the pectoral muscle, but not into the chest wall. Susceptibility artifact associated with a clip placed at stereotactic biopsy (upper edge of the mass) is better seen on the unsubtracted view.

Neoadjuvant chemotherapy was administered with Adriamycin and Cytoxan (AC) and paclitaxel (Taxol), with shrinkage of the tumor.

Left mastectomy, with excision of some muscle, and sentinel node sampling were performed. The residual tumor was a 1-cm, well-differentiated IDC, estrogen receptor and progesterone receptor positive, HER-2/neu negative, with positive deep margin and perineural involvement, and two negative axillary lymph nodes.

Additional therapy given was chest wall radiation and tamoxifen.

TEACHING POINTS

It can be difficult to evaluate the extent of muscle involvement of posterior breast masses based on mammography, ultrasound, and clinical exam. This is an important determination, with implications for surgical decision making and prognostic significance. The pectoral muscle is considered part of the breast, and adequate surgical therapy of posterior breast masses superficially involving the pectoral muscle will need to include some muscle in the excision. A radical mastectomy can be considered for more extensive pectoral muscle involvement. If the chest wall (ribs, intercostal muscles, and serratus anterior muscle) is involved, the lesion is considered T4, and the patient’s stage IIIB.

Breast MRI is the most accurate imaging modality currently in use for assessment of the extent of muscle involvement by posterior breast cancers. Morris and associates have shown that abutment or effacement of the fat plane is not sufficient imaging evidence to suggest muscle invasion. Actual enhancement in the muscle is necessary to diagnose muscle invasion. In this case, pectoral muscle involvement can be suspected from ultrasound, but the delineation of the extent of muscle invasion is more precise with MRI.

SUGGESTED READING

Morris EA, Schwartz LH, Drotman MB, et al. Evaluation of pectoralis major muscle in patients with posterior breast tumors on breast MR images: early experience. Radiology. 2000;214:67-72.

CASE 9 Breast MRI problem solving: Chest wall invasion

A 45-year-old woman presented with a palpable mass in the upper inner right breast. Mammography showed a partially obscured mass with associated suspicious calcifications, corresponding to the palpable mass (Figure 1 and Figure 2). Ultrasound demonstrated the mass to be an irregularly marginated solid mass (Figure 3). Ultrasound-guided core needle biopsy identified high-grade invasive ductal carcinoma. Preoperative MRI demonstrated the known tumor to invade the pectoralis and chest wall (intercostal) muscles (Figure 4). The patient was treated with neoadjuvant chemotherapy, followed by surgery and radiation therapy.

FIGURE 1 ML (A) and CC (B) mammographic views of the right breast show an ill-defined mass with associated calcifications in the upper inner right breast (arrows).

FIGURE 2 Close-up view, right CC projection: Suspicious coarse and ductal calcifications are seen within the mass.

FIGURE 3 Ultrasound of the upper inner right breast demonstrates the mammographic and palpable mass to correspond to an irregular solid mass with posterior acoustic shadowing. Small echogenic foci with the mass (arrow) represent calcifications within the tumor.

FIGURE 4 Delayed post–contrast MRI axial image: MRI demonstrates the mass in the upper inner right breast. In addition, there is linear enhancement extending into the adjacent pectoralis muscle as well as the underlying intercostal muscles. This indicates pectoralis and chest wall invasion.

TEACHING POINTS

This case of locally advanced breast cancer illustrates the use of preoperative MRI to aid in tumor staging. Pectoralis and chest wall involvement should be considered in patients with posterior tumors. Effacement of the intervening fat or abutment of the pectoralis fascia does not correlate well with muscle involvement. However, enhancement of the muscle itself on MRI is highly suggestive of muscle invasion. Although pectoralis involvement alone does not change staging, chest wall (intercostal muscle) invasion would represent stage IIIB. In addition, involvement of the pectoralis or chest wall may affect surgical management as well as radiation therapy planning.

SUGGESTED READING

Morris EA, Schwartz LH, Drotman MB, et al. Evaluation of pectoralis major muscle in patients with posterior breast tumors on breast MR images: early experience. Radiology. 2000;214(1):67-72.

CASE 10 Breast MRI problem solving: MRI guidance for tailored lumpectomy

A 41-year-old woman underwent routine screening mammography. A nodular right upper outer quadrant (UOQ) density was questioned. Spot compression suggested persistent architectural distortion. Ultrasound of the right lateral breast at 9 o’clock showed a subtle, hypoechoic nodule (Figure 1). The axilla was more convincingly abnormal, with a lymph node with effaced hilus and abnormal vascularity (Figure 2 and Figure 3). Ultrasound-guided core needle biopsy of the small breast mass identified infiltrating ductal carcinoma (IDC), estrogen receptor negative, progesterone receptor positive, HER-2/neu negative. Malignancy was also confirmed in the axillary lymph node by fine-needle aspiration (FNA) (Figure 4).

FIGURE 1 Ultrasound of the right lateral breast shows an oval, taller-than-wide, deep, hypoechoic indeterminate breast mass (cursors). Ultrasound-guided core needle biopsy identified IDC.

FIGURE 2 Axillary ultrasound shows two adjacent lymph nodes. The one on the left is very vascular and has complete effacement of the fatty hilus.

FIGURE 3 Transverse ultrasound view of the vascular lymph node in Figure 2: Note abnormal vascularity on the periphery of the node. Identification of vascularity anywhere but in the hilus is abnormal.

FIGURE 4 Image from an ultrasound-guided axillary lymph node FNA. The skinny needle (25 gauge) is entering from the right and can be seen within the hypoechoic node.

Staging evaluations showed normal tumor markers and no evidence of distant metastatic disease by positron emission tomography, CT, and bone scans (Figure 5). Breast MRI was requested by the breast surgeon because of increased mammographic breast density. MRI showed the disease process to be larger than previously suspected. In the region identified by ultrasound, MRI showed a multinodular, clumped area of enhancement with washout (Figure 6 and Figure 7). Individual small mass components measured on the order of 1 cm each, with the overall process measuring about 3 cm in dimension. Given the discrepancy between the volume of disease displayed by ultrasound compared with MRI, there was concern that localization for breast conservation with ultrasound guidance would lead to positive margins.

FIGURE 5 Assessment of axillary lymph nodes by size criteria, used with CT and MRI, can underestimate axillary involvement. A, Enhanced chest CT shows an outsized right axillary lymph node, corresponding to ultrasound. Margins appear fuzzy here, but this could be post-FNA change. B, Axial STIR MRI of the corresponding axillary findings.

FIGURE 6 Axial enhanced subtracted T1-weighted breast MRI. A, Maximal intensity projection view shows a localized, multinodular abnormality in the posterior right lateral breast, seen against a background of scattered fibrocystic foci of enhancement. B, Single axial subtracted view of two of the larger, small, irregularly marginated masses constituting this abnormality.

FIGURE 7 Assessment of kinetic data is made easier with color mapping. These examples use DynaCAD software. A, Foci of washout are colored lighter (yellow and red). B, Representative kinetic curve from this lesion shows rapid upstroke of enhancement and washout.

Under MRI guidance, clips were placed to bracket the MRI abnormality (Figure 8). One was placed anterolaterally and the other posteromedially. Post-MRI procedure mammograms confirmed successful clip deployment. The clips were localized with ultrasound guidance on the day of surgery.

FIGURE 8 Images from MRI-guided biopsy and clip placement and subsequent ultrasound-guided needle localization. A, Sagittal subtracted view obtained in a coil with a lateral localizing grid in place to reidentify the target lesions. Targets anterolateral and posteromedial to the enhancement were selected for clip placement. B, Post-MRI-guided clip placement lateral digital mammogram confirms both clips to be successfully deployed, in the posterior upper outer quadrant, at 9 to 10 o’clock. C, Exaggerated craniocaudal mammogram shows both clips. D, On the day of surgery, ultrasound was used to identify the MRI-placed clips. The clips are hyperechoic, and annotated here. They can be difficult to visualize with ultrasound if not placed against a hypoechoic tissue background.

The partial mastectomy specimen showed a 2.4-cm IDC, with ductal carcinoma in situ and angiolymphatic invasion. The anterior and inferior margins were close (<2 mm), with others clear. Three of nine sampled lymph nodes were involved, and the largest was 1.1 cm. Re-excision was performed of the close margins, with no IDC and only biopsy site changes noted.

The final stage was stage II, T2N1. Chemotherapy was begun on a study protocol.

TEACHING POINTS

The intent of the MRI-guided clip placement was to maximize the likelihood of achieving clear margins at partial mastectomy. Similar in principle to bracketing of microcalcifications seen on mammography, the larger extent of disease seen by MRI suggested positive margins would be a likely outcome if localized any other way. In this case, clear but close margins were actually achieved on the first attempt, although that was not known with certainty until re-excision was performed subsequently.

CASE 11 Breast MRI problem solving: Assessment of completeness of breast cancer excision

A 48-year-old woman noted a palpable right breast lump, which on her physician’s physical exam was assessed to be 1 cm in the upper outer quadrant (UOQ) and possibly a cyst. Mammography showed dense breast tissue only. Ultrasound at the site of the palpable lump was negative. However, a sonographically suspicious 1.5-cm mass was found nearby, localizing to 10 o’clock, with a second nearby 7-mm nodule at 9 o’clock (Figure 1). Both were sampled with ultrasound-guided core biopsy, confirming infiltrating ductal carcinoma (IDC) at both sites.

FIGURE 1 Antiradial ultrasound of the right lateral breast shows adjacent hypoechoic, highly suspicious masses with irregular margins at 9 and 10 o’clock. Subsequent core biopsy with ultrasound guidance confirmed IDC from both masses.

MRI was ordered to assess the extent of tumor because the known IDC was mammographically and clinically occult. MRI suggested multifocal tumor, with three UOQ intensely enhancing tumor nodules, two of which were immediately adjacent to each other and could have been components of a bilobed mass (Figure 2).

FIGURE 2 Enhanced, subtracted, axial T1-weighted gradient echo MRI shows apparent multifocal tumors in the right UOQ with a background of diffuse, stippled, fibrocystic foci of enhancement. Three discrete (versus one bilobed and a separate smaller focus) tumor masses were suspected based on this study. A, Axial maximal intensity projection (MIP) view. B, Axial enhanced subtracted views, from above to below (1–3), show rim enhancement of the largest enhancing mass. The most inferior mass was immediately below this, and the most superior, smaller mass was 2.5 cm superolateral. C, Sagittal subtracted enhanced view, obtained about 5 to 6 minutes after contrast administration, shows the two larger masses. Although they appeared separate by earlier axial imaging, enhancing bridging joining the two masses is now seen.

The largest mass measured 1.8 cm, with the adjacent mass or component 1.2 cm in maximal dimension. Another smaller mass was identified 2.5 cm superolaterally. Second-look ultrasound was then undertaken and showed an additional 5 mm nodule at 10 o’clock, 2 cm away from the largest index lesion (Figure 3).

FIGURE 3 Second-look ultrasound. A, Ultrasound shows a subtle correlate for the separate, superolateral small mass seen on MRI as a vague, 5-mm hypoechoic focus within dense fibrous tissue, which by ultrasound was 2 cm from the index lesion (arrow). B, Additional ultrasound image shows the relationship between the new additional tumor nodule and the index lesion (cursors).

Because the patient desired breast conservation, triple-needle localization with ultrasound guidance was performed. The initial pathologic interpretation was confusing and reported as IDC and ductal carcinoma in situ (DCIS), larger than 2 cm and smaller than 5 cm, with no mention of multiple components. After discussion at breast conference, this was clarified, with pathology ultimately reporting a 2.8-cm IDC, consisting of a continuous region of admixed IDC and DCIS (only one of the localized regions was abnormal by gross). DCIS was present less than 1 mm from the final lateral margin. Angiolymphatic invasion was noted. One sentinel lymph node was negative.

Re-excision was performed 2 weeks later. The new superior and lateral margins were negative, with no residual IDC.

The patient was started on chemotherapy with Adriamycin and Cytoxan (AC) because of the angiolymphatic invasion and her intermediate-range Oncotype DX recurrence score of 21.

Because of the confusion resulting from the apparent multifocality of the tumor on imaging compared with the single tumor mass on pathology, breast MRI was obtained to assess the completeness of the resection before beginning radiation therapy. By this time, two cycles of AC had been given.

The breast MRI showed a 7-mm enhancing nodule medial to the postoperative seroma, suspicious for a residual tumor focus (Figure 4). Biopsy with MRI guidance confirmed IDC (Figure 5). A clip was placed, which was subsequently localized with two bracketing wires. The specimen showed a 0.6-cm residual IDC, and the margins were clear.

FIGURE 4 Repeat breast MRI, after two surgical excisions, the second with clear margins, after beginning chemotherapy and before radiation therapy, showed a small solitary intensely enhancing nodule in the postoperative bed, medial to a postoperative seroma (not shown). The nodule is larger and more intensely enhancing than the diffuse background lower level enhancement, and was subsequently proved to be a residual IDC by MRI-guided biopsy. A, Axial enhanced subtracted MIP image. B, Axial enhanced subtracted slice.

FIGURE 5 Images from MRI-guided vacuum-assisted core needle biopsy, which confirmed residual IDC, despite negative margins at prior excision. A, Enhanced, fat-saturated sagittal localizer shows the enhancing target. B, The target is initially easier to confirm and visualize with subtraction. C, The lesion is targeted for sampling with the aid of a software program (DynaCAD). D, The software calculates the depth from the lateral localizing grid, which contains two fiducials (arrows). E, The appropriate skin entry site and depth are determined. F, After placement of the localizing obturator, axial images are obtained to verify its position. The hypointense linear obturator is seen in the lateral breast, with the end at the level of the enhancement. This marks the center of the sampling notch. G, Repeat axial imaging after sampling. The enhancement is no longer seen. H, Repeat sagittal imaging after sampling and clip placement shows two adjacent hypointense foci at the level of the sampled target (compare with A). One represents the obturator and the other the deployed clip. I, Verification of successful clip deployment with a postbiopsy mammogram is advisable. Digital lateral (I1) and CC (I2) mammograms show three adjacent surgical clips in the posterior UOQ, on the edge of the field in both views. These were from the prior excisions. Anterior to the surgical clips is the MRI-placed clip, projecting at the inferomedial margin of the postoperative seroma.

A total of four cycles of AC were completed while the patient was undergoing these procedures. Radiation therapy was completed subsequently, and the patient was started on tamoxifen. Final stage was stage IIA, T2N0M0, estrogen receptor and progesterone receptor positive, and HER-2/neu negative.

TEACHING POINTS

Although it took three surgical excisions, this patient ultimately was successfully treated with breast conservation. Breast MRI was instrumental in enabling this patient to be thus treated. Breast MRI provided the most accurate map of the extent of disease, guiding the second-look ultrasound to a third focus of disease. Ultrasound-guided triple-needle localization removed the bulk of the disease. A close margin prompted re-excision, with the new margins negative.

The apparent discrepancy between the breast MRI, which suggested multifocal tumor nodules, and pathology, which reported only a single tumor mass, caused the pathology to be re-reviewed and a tighter correlation to be made. The tumor was admixed IDC and DCIS. A possible explanation could be that the enhancing components seen on MRI represented invasive components, with the DCIS poorly delineated. The ensuing discussion of this case led to the consensus that a repeat breast MRI study should be obtained to assess the completeness of the resection.

With subsequent confirmation of residual disease, and the patient still desirous of breast conservation, localization and excision of the confirmed residual were undertaken. Whether such small-volume residual disease as this, which undoubtedly would have gone undetected in the pre–breast MRI era, would have been adequately treated with radiation and chemotherapy or would have predisposed this patient to local recurrence in subsequent years, we can only speculate. The rationale of breast conservation is predicated on the surgical removal of all identifiable disease, a task that has become increasingly complex now that we are better able to identify disease. There is support in the literature that indicates that patients treated with breast conservation using preoperative breast MRI have fewer local recurrences. To date, no studies evaluating the effect of the use of preoperative breast MRI on overall survival have been reported.

SUGGESTED READING

Fischer U, Zachariae O, Baum F, et al. The influence of preoperative MRI of the breasts on recurrence rate in patients with breast cancer. Eur Radiol. 2004;14(10):1725-1731.

CASE 12 Contralateral breast carcinoma found on MRI; not seen on second-look ultrasound

A 48-year-old woman with mammographically detected ductal carcinoma in situ in the right breast underwent preoperative MRI. The MRI revealed a suspicious enhancing mass in the contralateral left breast (Figure 1). Subsequent targeted ultrasound was performed of the lateral left breast in the area of the MRI lesion (Figure 2 and Figure 3). Because no discrete lesion was seen on ultrasound, MRIguided biopsy was performed of the left breast mass (Figure 4). Core needle biopsy revealed an 8-mm intermediate-grade invasive ductal carcinoma.

FIGURE 1 Detail view of the subtracted postcontrast MRI. There is an irregularly shaped enhancing mass in the lateral left breast (arrow).

FIGURE 2 Ultrasound image of the lateral left breast. No discrete lesion is seen to account for the MRI finding.

FIGURE 3 Color Doppler ultrasound image of the lateral left breast. Increased vascularity is seen in the area of the MRI finding.

FIGURE 4 Axial image from an MRI-guided core biopsy of the lateral left breast. The biopsy needle is seen entering the lateral left breast with its tip at the level of the targeted mass.

TEACHING POINTS

This case illustrates the limitations of ultrasound in detecting correlates for suspicious MRI lesions. Although this patient had not had a prior ultrasound of the left breast, the term second-look ultrasound is used when targeted ultrasound is performed to find an ultrasound correlate to a finding on MRI. In this case, although some vascularity was detected on Doppler ultrasound in the corresponding area in the lateral left breast, this finding is not enough to allow confident correlation with the lesion seen on MRI. In general, ultrasound-guided needle biopsy is preferred over MRI-guided biopsy because of patient comfort, time, and cost. However, depending on the size of the lesion seen on MRI and the quality of the ultrasound evaluation, second-look ultrasound may fail to identify a sonographic correlate in as many as 77% of MRI-detected lesions referred for biopsy. Therefore, a suspicious lesion seen on MRI requires biopsy despite a negative ultrasound. For programs performing breast MRI, it is essential that there also be MRI-guided biopsy capability to assess sonographically occult lesions.

SUGGESTED READINGS

Dhamanaskar KP, Muradall D. MRI directed ultrasound: a cost effective method for diagnosis and intervention in breast imaging [abstract]. Radiology. 2002;225:653.

LaTrenta LR, Menell JH, Morris EA, et al. Breast lesions detected with MR imaging: utility and histopathologic importance of identification with US. Radiology. 2003;227:856-861.

Panizza P, De Gaspari A. Accuracy of post MR imaging second-look sonography in previously undetected breast lesions [abstract]. Radiology. 1997;205:489.

CASE 13 Evaluation of the other breast with MRI

A 43-year-old woman with 10-year-old breast implants noted left axillary lumpiness. Similar complaints had been evaluated 5 years before. When the resurgent axillary complaints failed to improve after antibiotic therapy, the patient was referred for breast imaging evaluation.

Mammography showed asymmetrical density and nodularity at the levels indicated by the patient. Ultrasound showed multiple abnormal axillary lymph nodes as well as round lymph nodes at the lateral margins of the implant (Figure 1). Two vascular, dominant, highly suspicious breast masses were identified in the upper outer quadrant (UOQ) at 1 and 2 o’clock (Figure 2). At least three additional discrete, 4- to 5-mm, hypoechoic nodules were noted in the same quadrant, interposed between and adjacent to the dominant masses. The two dominant breast masses were sampled with ultrasound-guided core needle biopsy, confirming infiltrating ductal carcinoma (IDC) from both sites. Two axillary lymph nodes were also sampled with fine-needle aspiration technique, confirming metastatic poorly differentiated carcinoma (consistent with breast primary) from both sampled sites (Figure 3).

FIGURE 1 A variety of abnormal ultrasound manifestations of axillary metastases were demonstrated in this case. A, Color Doppler of an enlarged axillary lymph node shows abnormal thickening of the cortical mantle (up to 8 mm in this case). The fatty hilus can still be visualized. Color Doppler features of this vascularity further confirm the suspicious nature of this lymph node. Instead of vascular flow confined to the hilus, flow can be readily identified at the periphery of the thickened cortex. B, Additional left axillary nodal findings suggest (as subsequently proved pathologically) that multiple lymph nodes are involved. This section shows two very hypoechoic, rounded lymph nodes, with the one on the right retaining only a hint of a reniform shape and hilar notch in profile. C, Two additional, abnormal, enlarged, adjacent rounded left axillary lymph nodes show peripheral vascularity and complete loss of the fatty hilus. D, Even small, completely round lymph nodes like this, adjacent to the implant margin, should be regarded as suspicious in the appropriate setting.

FIGURE 2 Sonographic images of the left UOQ show two dominant, vascular, highly suspicious masses at 1 (A and B) and 2 o’clock (C). Both are solid, hypoechoic, taller than wide, irregularly marginated, and highly vascular and both were confirmed to be IDC. The relationship of the masses to the underlying sonolucent implant can be appreciated in A and C.

FIGURE 3 Image obtained from ultrasound-guided fine-needle aspiration of one of the abnormal axillary lymph nodes shows the skinny needle within the hypoechoic, rounded lymph node.

Based on these evaluations, the patient’s best surgical treatment option appeared to be a mastectomy, with removal of the implant. A breast MRI was obtained to assess the proximity of the disease to the implant and to look for additional disease sites.

Breast MRI showed extensive left UOQ and axillary disease and correlated well with the sonogram (Figures 4, 5, 6, 7, 8, and 9). The two known foci of IDC manifested as intensely rim-enhancing, spiculated masses with washout kinetics. They were accompanied by an entire quadrant filled with enhancing smaller masses and clumped enhancement, extending down to the level of the nipple. Multiple, intensely enhancing lymph nodes were identified, also correlating with the ultrasound, both along the lateral margin of the implant and extending up into the axilla.

FIGURE 4 Enhanced subtracted axial maximal intensity projection view of both breasts shows large implants with extensive left lateral disease, with multiple enlarged axillary lymph nodes. A large draining vein extends to the internal mammary level. On the right, a localized, segmental region of enhancement is seen in the anterior lateral breast on the surface of the implant (arrow).

FIGURE 5 Axial subtracted image shows the two known rim-enhancing left UOQ IDC lesions, one near the midline and the other lateral to it. The entire region shows clumped enhancement in addition, suggesting an extensive intraductal component.

FIGURE 6 An additional slice from the same series shows two adjacent, rim-enhancing, round lymph nodes at the lateral margin of the implant, corresponding to the ultrasound findings depicted in Figure 1C. Anterior to these, in the lateral breast tissue adjacent to the implant, is a cluster of three tiny enhancing masses.

FIGURE 7 An additional section, obtained 3 minutes after contrast injection, shows tiny foci of clumped enhancement throughout the left breast, lateral to the nipple. A similar localized region of clumped enhancement is seen in the right breast, lateral to the nipple (arrow). Targeted ultrasound was recommended to evaluate the concerning right-sided findings.

FIGURE 8 Sagittal enhanced subtracted images of these findings. A, Slice through the left 1-o’clock IDC mass, which is seen here with rim enhancement. The adjacent tissue shows confluent, clumped enhancement as well. B, A more lateral section of the left breast, through the 2-o’clock IDC mass, also displays rim enhancement of the mass. Enhancing spicules project posteriorly from it, and there is lower-level confluent enhancement throughout the quadrant, both anterior and posterior to the IDC mass. Multiple adjacent enhancing nodules and lymph nodes are noted posteriorly. C, Sagittal view of the clumped localized enhancement in the right lateral breast.

FIGURE 9 Kinetic information is consistent with the known multifocal nature of the left-sided malignancy and heightens suspicion about the more indeterminate right breast MRI findings. Color map (A) and signal intensity–time curve (B) of the left UOQ mass at 1 o’clock show early, intense enhancement and washout, evidence of angiogenesis. C and D, The left UOQ mass at 2 o’clock shows similar kinetics. E and F, The lower-level, more confluent UOQ clumped enhancement shows a plateauing pattern of enhancement. G and H, Within the clumped localized right-sided enhancement is a small region of washout.

A suspicious, as yet unsuspected, contralateral abnormality was identified in the right breast on MRI, manifesting as a 2 × 2.6-cm region of clumped progressive enhancement, within which was a 6-mm nodule with washout (see Figures 7, 8C, and 9G and 9H). Targeted sonography identified tiny hypoechoic nodules in the expected region. The most suspicious measured 5 mm in maximal dimension and showed irregular, angular margins (Figure 10). Core needle biopsy under ultrasound guidance identified focal (1.5 mm) invasive ductal carcinoma (in one of six cores) and intermediate-grade ductal carcinoma in situ (DCIS) in all the cores.

FIGURE 10 Ultrasound of the right UOQ identified tiny hypoechoic nodules. This one had the most concerning appearance, with irregular margins and ductal extension suggested medially. This was cored with ultrasound guidance and a 1.5-mm IDC focus, and DCIS was confirmed.

Additional preoperative staging was obtained, including bone scan, enhanced chest, abdomen and pelvis CT scans, and positron emission tomography (PET)/CT (Figures 11, 12, 13, 14, 15, and 16). The PET scan showed the two known left UOQ IDCs to be hypermetabolic, as were multiple level I and II left axillary lymph nodes. No evidence of systemic breast cancer was seen. Hypermetabolism identified of the uterine lining was followed up with ultrasound, which was unremarkable, and the activity was presumed to be physiologic variation.

FIGURE 11 Coronal PET/CT images (left to right: CT, PET, fused PET/CT, and NAC-PET) show multiple hypermetabolic left axillary lymph nodes. A midline focus of increased activity in the pelvis above the bladder localized to the uterine lining, a physiologic normal variant.

FIGURE 12 Axial PET/CT images better depict the location of the axillary disease. A, Hypermetabolic level I (lateral to the pectoralis minor [PM] muscle) and level II (beneath the PM muscle) left axillary lymph nodes are depicted. B, More inferior axillary level I hypermetabolic nodes are lateral to the implant.

FIGURE 13 Axial PET/CT images through the two sites of known IDC in the left breast show corresponding hypermetabolism.

FIGURE 14 Corresponding contrast-enhanced chest CT images for correlation, from above to below. A, Enlarged left axillary level I and II lymph nodes are strikingly asymmetrical compared with the right side. Several show fuzzy margins. B, At the upper margin of the implant, adjacent lymph nodes with fuzzy margins correspond with findings depicted on ultrasound (see Figure 1C) and MRI (see Figure 6). C, The two IDC masses can be seen on CT owing to differential enhancement from the rest of the parenchyma, but not to the same advantage as on breast MRI (arrows).

FIGURE 15 Three-plane fused PET/CT images show increased metabolic activity in the pelvis, localizing to the uterine lining (cursors). The axial projection shows a right adnexal corpus luteum cyst, which is not hypermetabolic in this case.

FIGURE 16 Coronal PET/CT images show symmetric mildly increased activity of “brown fat” in the supraclavicular regions bilaterally, localizing to fat on the corresponding CT.

With the new contralateral breast information, the patient elected to undergo bilateral mastectomy and implant removal. The simple mastectomy pathology on the right showed a 2-cm high-grade DCIS lesion extending into lobules, and two sentinel lymph nodes were negative. The left mastectomy specimen showed two foci of IDC, one described as in the lower outer quadrant (LOQ) measuring 1.5 × 1.4 × 1.4 cm. A second IDC was 2 cm medial and superior to the first focus (in the mid-breast posterior to nipple) and measured 2.1 × 2.0 × 1.5 cm. There was extensive angiolymphatic invasion. DCIS (solid, with foci of comedo necrosis) was associated with the invasive tumors and extended into lobules. Nine of 18 lymph nodes showed metastatic tumor, with the largest 1.4 cm with extranodal extension. Margins were negative, by at least 5 mm. The tumors were estrogen receptor negative, progesterone receptor positive, and HER-2/neu negative.

After surgery, the patient was treated with chemotherapy (six cycles of Taxotere, Adriamycin, Cytoxan (TAC)), after which left chest wall and peripheral lymphatic radiation therapy to the supraclavicular and posterior axillary regions was performed. Additionally, the patient was placed on tamoxifen.

TEACHING POINTS

Implants in patients with newly diagnosed breast cancer introduce additional complexity into the therapeutic option decision making. In addition to the usual determinates of whether the patient is a lumpectomy candidate or not, the relationship of the disease to the implant and the impact of the implant on the ultimate cosmesis must be considered. If the patient is a candidate for breast conservation based on clinical and imaging assessments of the disease extent, the effect of radiation on a conserved breast with an implant must be taken into account. Breast radiation with an implant in place can be performed but may lead to contracture and an undesirable cosmetic result. In this case, the patient’s reconstruction options were further limited by a prior “tummy tuck” procedure.

In this case, ultrasound accurately suggested the extensive involvement of the left breast. Prospective ultrasound of the right breast (not performed) might have successfully identified the right-sided disease, but clearly, it is easier to find a target with a map in hand (namely, breast MRI). The localized nature of the clumped enhancement in the right breast by itself would have been concerning, given how “quiet” the rest of the breast was. Coupled with the small focus of angiogenesis, our suspicion level is appropriately heightened, and any reasonably concordant sonographic correlate should be sampled. If no ultrasound correlate had been identified, it might have been technically possible to undertake MRI-guided biopsy in this region, given its anterior location. MRI-guided clip placement for subsequent localization would have been another consideration.

The correlation of the ultrasound and MRI findings with the final pathologic results is quite good in this case. The right-sided MRI findings of clumped enhancement suggested DCIS as the primary diagnosis of concern, and the small focus of invasion found at pathology correlates with the small nodule with angiogenesis on MRI kinetics and the small but suspicious ultrasound mass. Similarly, the two known IDC lesions on the left were accompanied by an entire quadrant of additional nodules and lower-level, clumped, plateauing enhancement, suggested a significant component of DCIS as well as invasion, as was subsequently confirmed.

One apparent discrepancy, of the locations of the left-sided IDC masses compared with the final pathology, is probably artifactual because of the presence in vivo of the implant. The lesion identified by pathology as in the LOQ apparently corresponds with the lesion identified as the UOQ 2 o’clock IDC, whereas the other lesion, described by pathology as 2 cm medial and superior to the LOQ IDC and in the mid-breast posterior to the nipple, correlates with the 1-o’clock IDC. In this case, the pathologist contacted radiology and reviewed the MRI to resolve the apparent discrepancy.

Systemic staging studies obtained in this patient included PET/CT and enhanced body CT scans. This case shows one source of normal variant PET increased activity, which can be seen in the pelvis of premenopausal patients. Physiologic endometrial activity can be seen, as well as activity in uterine fibroids (see Chapter 4, Case 7) and in corpus luteum cysts (see Case 17 in this chapter). In this case, PET/CT allows us to localize the uterine activity to the endometrium. Subsequent pelvic sonogram showed a normal appearance of the endometrium, and the patient was asymptomatic, so this is presumed to be normal variant activity. This has been documented to occur most commonly during menstruation and ovulation (weeks 1 and 3 of the menstrual cycle). Right adnexal findings compatible with a corpus luteum cyst were seen in this patient, but no associated hypermetabolism was apparent. Another commonly seen source of normal variant activity, brown fat, is also demonstrated in this patient.

SUGGESTED READING

Lerman H, Metser U, Grisaru D, et al. Normal and abnormal ¹⁸F-FDG endometrial and ovarian uptake in pre- and postmenopausal patients: assessment by PET/CT. J Nucl Med. 2004;45:266-271.

CASE 14 Use of body coil STIR imaging for local staging of new diagnoses of breast cancer

A 51-year-old woman noted a palpable lump in her right breast for 2 to 3 months, which was confirmed by her physician as a palpable 1-cm mass above the nipple at 11 to 12 o’clock. Mammography of the area showed increased density and possible architectural distortion in a region of mammographically stable microcalcifications (Figures 1, 2, and 3). Sonography confirmed a suspicious, dominant, solid, vascular, irregularly marginated, palpable mass, measuring up to 2.3 cm (Figure 4 and Figure 5). Because it was palpable, the patient was referred to surgery for a palpation-guided biopsy, which diagnosed infiltrating ductal carcinoma (IDC).

FIGURE 1 Mammogram in the CC position with a marker at a palpable lump identified by the patient shows marked increased breast tissue density. Increased density and microcalcifications are noted at the marked level. The microcalcifications had been mammographically stable for 10 years.

FIGURE 2 Lateral spot compression view of the palpable lump shows a marker. Underlying it, there is locally increased breast density and a suggestion of architectural distortion.

FIGURE 3 Lateral magnification view of the area shows increased density, possible architectural distortion, and microcalcifications. Some represent milk of calcium. The calcifications were mammographically stable for 10 years.

FIGURE 4 Ultrasound of the palpable mass shows it to be highly suspicious—solid, very hypoechoic, and irregularly marginated.

FIGURE 5 Color Doppler evaluation shows the suspicious mass to be highly vascular.

Because of the marked breast density, completion of the evaluation was undertaken with breast MRI. This showed the known IDC to be a dominant, spiculated, intensely enhancing mass. The disease appeared to be unifocal (Figure 6 and Figure 7). However, 1-cm bilateral supraclavicular lymph nodes (one on each side) were suggested on a coronal STIR sequence of the thorax (Figure 8). Positron emission tomography (PET)/CT was obtained for further evaluation and showed hypermetabolism only at the site of the known IDC (Figure 9).

FIGURE 6 Axial maximal intensity projection from dynamic enhanced breast MRI shows a dominant, intensely enhancing mass with irregular margins.

FIGURE 7 Fat-saturated, enhanced, high-resolution MRI of the mass better shows the marginal spiculation.

FIGURE 8 Coronal STIR images of the thorax (A anterior to B), obtained with the body coil with breast MRI, show oval hyperintense foci bilaterally, suggesting supraclavicular nodes.

FIGURE 9 The known right IDC is hypermetabolic and well seen on sagittal PET/CT (left to right: CT, AC-PET, fused PET/CT, NAC-PET). No other abnormality was seen.

The patient was treated surgically with lumpectomy and axillary node dissection, finding an estrogen receptor– and progesterone receptor–positive, HER-2/neu-negative, 2.1-cm IDC, with 1 of 12 lymph nodes positive. Final stage was IIB, T2N1M0. Two cycles of Cytoxan, methotrexate, 5-fluorouracil (CMF) chemotherapy were administered. The patient’s therapy was changed to hormonal ablation with sulindac (Zoladex) and tamoxifen because of her desire to discontinue chemotherapy.

External-beam radiation therapy, with a boost to the lumpectomy bed, was also performed in this case.

TEACHING POINTS

A variety of protocols are used for performance of breast MRI, with little standardization across institutions. That said, most protocols have certain critical elements in common. In general, breast MRI should be performed on a high-field magnet (at least 1.0 Tesla) in a dedicated breast coil. At a minimum, the protocol must include at least a fluid-sensitive sequence (T2 or STIR), and fatsaturated or subtracted enhanced T1-weighted sequences. Acquisition can be sagittal or axial, unilateral or bilateral. We have a strong preference for axial imaging performed bilaterally to allow for ease of comparison with the opposite side. Choices for fluid-sensitive sequences include T2-weighted, fat-saturated T2 weighted, and STIR sequences. These sequences permit identification of fluid, such as in cysts or in postoperative collections. Bright signal intensity on STIR of common encountered enhancing benign entities, such as fibroadenomas and lymph nodes, aids in characterization.

All protocols evaluating the breast parenchyma rely on early differential enhancement between suspicious lesions and the normal breast parenchyma, which enhances more slowly. For this reason, imaging must commence soon after the intravenous administration of Gd-chelate contrast agents, with the peak enhancement of most cancers occurring 1 to 2 minutes after injection. Dynamic, repetitive imaging before and at multiple time points after injection of the contrast enables the before and after series to be subtracted from each other, increasing the visual conspicuity of any enhancement. Three-dimensional, T1-weighted gradient echo sequences can image both breasts in the axial plane in 60 to 90 seconds, depending on acquisition parameters. This type of sequence is commonly used for performance of dynamic imaging, with the sequence built to run once before the contrast injection is triggered, and then again 3 to 5 times after the contrast is given. The sequence can be performed with or without fat saturation. However, if the patient moves during the dynamic series, subtractions will be unsuccessful, and the interpretation will have to be made from the source images, for which reason fat saturation is preferred.

As an example of how different breast MRI protocols can be, here are the current protocols in use at our institutions:

University of California, San Diego

Angled T1-weighted parasagittal oblique through each axilla (using the body coil)

Unilateral T1-weighted axial series through each breast

Unilateral fat-saturated T2-weighted series through each breast

Dynamic, fat-saturated, enhanced, bilateral axial T1-weighted gradient recalled series, run 6 times, once before and 5 times after contrast given (each series 60 seconds)

High-resolution, fat-saturated, three-dimensional axial series

25 minutes total of imaging

Scripps Clinic

Bilateral axial STIR series (5 minutes, 34 seconds)

Bilateral sagittal, T1-weighted series (precontrast) (4 minutes, 58 seconds)

Dynamic, fat-saturated, enhanced bilateral axial T1-weighted gradient recalled series, run 5 times, once before and 4 times after contrast given (each series 1 minute, 12 seconds)

Repeat bilateral sagittal, T1-weighted series (postcontrast) (4 minutes, 58 seconds)

20 minutes of total imaging

OPTIONAL (bail-out sequence for the patient suspected of having moved during the dynamic series, or if fat saturation fails): high-resolution, fatsaturated, three-dimensional axial series (3 minutes, 6 seconds)

Another choice that can be considered in creation of breast MRI protocols is concurrent performance of a limited chest MRI, in the form of a coronal STIR series through the thorax, encompassing the sternum and internal mammary vessels anteriorly and the thoracic spine posteriorly. This approach is advocated by Dr. Bruce A. Porter of First Hill Diagnostic Imaging in Seattle, who has had considerable success in identifying unsuspected stage IV disease (e.g., bone metastases in the sternum or thoracic spine) and in assessment of regional lymph nodes. At the time of this case, we included this sequence in our protocol, but overall did not find it to be as helpful in our patient population as in Dr. Porter’s experience. As in this case, which suggested supraclavicular nodal involvement, we found that it not infrequently raised questions that required other imaging studies (in this case, a PET and CT) to answer them, without increasing the yield in terms of identifying more extensive disease than previously suspected. Another approach, which could have been considered in this case, would be to search sonographically for supraclavicular nodes, which could then undergo fine-needle aspiration.

CASE 15 MRI depiction of axillary and internal mammary node involvement

Annual screening mammography of a 72-year-old woman demonstrated suspicious masses in the medial left breast (Figure 1). Ultrasound confirmed these to be solid and demonstrated an abnormal left axillary node (Figure 2 and Figure 3). An MRI confirmed the left breast masses and the abnormal axillary node (Figure 4). The constellation of findings was suspicious for multifocal carcinoma with axillary metastasis. In addition, MRI showed an abnormal-appearing left internal mammary lymph node (Figure 5 and Figure 6). Ultrasound-guided core needle biopsy of the left breast confirmed multifocal invasive ductal carcinoma. The patient was treated with left mastectomy and radiation therapy.

FIGURE 1 MLO (A) and CC (B) mammographic views of the left breast show multiple irregular masses in the medial left breast (arrows).

FIGURE 2 Ultrasound images (A and B) of the medial left breast: multiple hypoechoic solid masses correspond to the mammographic masses (arrows). The findings are highly suspicious for multifocal carcinoma.

FIGURE 3 Ultrasound of the left axilla demonstrates an enlarged, abnormal lymph node with suspicious morphology for metastatic involvement. The lymph node is rounded, with effacement of the hilus.

FIGURE 4 Postcontrast subtracted, maximal intensity projection MRI. Multiple intensely enhancing masses are seen in the medial left breast. In addition, an abnormal-appearing lymph node (arrow) is seen in the left axilla.

FIGURE 5 Postcontrast subtracted axial MRI. An enlarged left internal mammary node is demonstrated (arrow).

FIGURE 6 Fat-saturated, T2-weighted axial MRI of the left breast. The abnormal internal mammary node shows slight increased T2 signal intensity (arrow).

TEACHING POINTS

This case illustrates the ability of MRI to demonstrate both axillary and internal mammary node (IMN) metastasis. Because normal-appearing nodes may harbor metastasis, MRI cannot be used to rule out nodal involvement. However, abnormal-appearing nodes may suggest metastasis. Internal mammary chain nodes may be seen incidentally in normal patients. Internal mammary nodes larger than 5 mm should be considered suspicious. In general, IMN metastases are located in the first or second intercostal spaces. More than 90% of women with metastases to the internal mammary nodes also have axillary node metastasis. Of the 5% to 10% of women who have isolated IMN involvement, most have tumors involving the medial aspect of the breast. Involvement of IMN nodes may indicate a worse prognosis and may benefit from radiation therapy.

SUGGESTED READINGS

Bundred NJ, Morgan DAL, Dixon JM. ABC of breast diseases: management of regional nodes in breast cancer. BMJ. 1994;309:1222-1225.

Jatoi I. Internal mammary sentinel nodes in primary breast cancer. Curr Med Res Opin. 2003;19(6):567-569.

CASE 16 Cautionary notes on the use of breast MRI

A 53-year-old woman was diagnosed with left breast infiltrating lobular carcinoma (ILC) and lobular carcinoma in situ (LCIS) by stereotactic biopsy of left upper outer quadrant (UOQ) microcalcifications. The pathology specimen identified atypical ductal hyperplasia (ADH) associated with microcalcifications, with the ILC and LCIS noncalcified. The patient underwent a lumpectomy, with no residual tumor in the specimen. One sentinel lymph node was negative. She was treated with radiation therapy in addition.

While undergoing radiation therapy, breast MRI was performed to assess the opposite side. This showed a segmental region of clumped enhancement in the right medial breast, where the breast was largely fatty (Figures 1, 2, 3, and 4). An MRI-guided biopsy was performed when no ultrasound correlate could be found, and pathology showed ductal carcinoma in situ (DCIS).

FIGURE 1 Axial enhanced subtracted maximum intensity view of both breasts shows low-level enhancement on the left posterolaterally, corresponding to the patient’s recent surgery site. In the right medial breast, there is a segmental region of enhancement.

FIGURE 2 Axial T2-weighted image shows the right medial breast to be largely fatty.

FIGURE 3 Representative enhanced subtracted image from the dynamic series shows an elongated region of clumped enhancement in the posteromedial right breast.

FIGURE 4 Sagittal subtracted view, obtained about 5 minutes after contrast injection, shows the right medial segmental enhancement.

Because of the extent of the abnormal segmental enhancement, over a 6-cm expanse, the patient was not thought to be a breast conservation candidate by multiple surgeons. Concurrently, microcalcifications in the right UOQ were noted to have progressed from prior studies. Two separate sites were sampled stereotactically, about 3 cm apart. One site returned a diagnosis of ADH and the other showed ADH and atypical lobular hyperplasia (ALH).

Despite multiple second opinions advising her to undergo mastectomy, the patient was quite reluctant to do so. Eventually, 16 months after her MRI-guided breast biopsy establishing the diagnosis of DCIS, she did undergo a simple mastectomy and placement of a tissue expander.

The mastectomy pathology showed ADH and ALH, but no evidence of carcinoma in situ or invasive carcinoma. Two sentinel lymph nodes were negative.

TEACHING POINTS

This is an unusual outcome. Generally, when a malignant diagnosis is established by a large-volume core needle biopsy, we expect the same or worse to be obtained when the area is excised. However, the opposite is quite possible, as this case and Case 6 in Chapter 6 demonstrate. It is certainly possible in image-guided biopsies to sample a small malignant component, only to find no residual tumor in the excision. In fact, that had happened in this patient in her original left breast diagnosis of ILC and LCIS. The targeted microcalcifications corresponded to ADH at pathology, whereas the ILC and LCIS were noncalcified but present in the stereotactic specimen. No residual tumor was found at pathologic examination of her lumpectomy specimen.

Presumably, the most intensely enhancing component visualized on breast MRI on the day of biopsy was targeted and proved to be DCIS. Because it would be difficult to sample multiple areas within such a large target area, we frequently must make the assumption that the entire process is probably all the same. Could mastectomy have been avoided in this case? A good cosmetic result would probably have been difficult to achieve, given that the patient would have required two large-volume excisions just to remove all of her known sites of disease (the right UOQ ADH and the long expanse in the right medial breast).

CASE 17 Whole-body PET as an adjunct to initial staging of node-positive breast cancer: Benign PET pelvic uptake in a corpus luteum cyst

A 41-year-old premenopausal woman presented with a few weeks’ history of a palpable left upper outer quadrant breast lump. Mammography identified a dominant 2 cm mass, with irregular and spiculated margins (Figure 1 and Figure 2). Ultrasound confirmed a 1.7-cm highly suspicious corresponding mass, as well as a small but abnormal looking axillary lymph node (Figures 3, 4, 5, and 6). Ultrasound-guided biopsy of the breast mass confirmed invasive lobular carcinoma (ILC), and fine-needle aspiration of the axillary lymph node showed metastatic adenocarcinoma.

FIGURE 1 MLO (A) and CC (B) mammograms show moderate breast parenchymal density. A dominant mass with spiculated margins is seen posteriorly overlying the left pectoral muscle on the MLO view but is largely excluded from the field of view on the CC view. This corresponds in location to the palpable lump, marked with a radiopaque BB.

FIGURE 2 Additional mammographic views include a 90-degree lateral view (A) and an exaggerated CC spot view (B). The margins can be better seen. Although some are obscured by adjacent parenchyma, where they can be seen, they range from indistinct to spiculated.

FIGURE 3 Ultrasound of the left upper outer quadrant palpable lump shows a highly suspicious appearance. The mass is solid, vascular, and very hypoechoic and has irregular, angular margins.

FIGURE 4 Axillary ultrasound shows a hypoechoic lymph node (cursors). Although not very large by size criteria (7 × 4 mm), its morphologic features are concerning. There is partial effacement of the fatty hilus.

FIGURE 5 Another ultrasound view shows a cortical thickness of 2 mm, also not clearly pathologic. In other regions, it measured up to 3 mm.

FIGURE 6 Color Doppler shows abnormal vascularity of this lymph node. In addition to expected hilar flow, there is peripheral, cortical flow. Ultrasound-guided fine-needle aspiration confirmed metastatic involvement.

Breast MRI was obtained because of the ILC histology. The known tumor mass showed intense rim enhancement and washout (Figures 7, 8, and 9). Tissue anterior to the known tumor showed an asymmetrical, less intense, plateauing pattern of more indeterminate enhancement (see Figure 7; Figure 10). MRI-guided biopsy was recommended.

FIGURE 7 Enhanced subtracted axial breast MRI shows central necrosis (rim enhancement) of the left posterior lateral dominant mass. A concerning swath of lower-level enhancement is seen extending anteriorly toward the nipple from the known tumor.

FIGURE 8 Color map of the enhancement (DynaCAD software) shows intense rim enhancement with washout (color-coded red) of the known cancer on the left posterolaterally. Less intense asymmetrical enhancement is seen anterior to the dominant mass (color-coded blue).

FIGURE 9 Corresponding signal intensity change versus time graph from the dominant mass shows a washout pattern of angiogenesis, with rapid upstroke (here to 200% of baseline) and subsequent decline (washout).

FIGURE 10 Signal intensity change versus time graph from the less intensely enhancing tissue anterior to the known cancer shows a plateauing pattern of enhancement. MRI-guided biopsy was recommended.

Positron emission tomography (PET)/CT was ordered because of the known axillary involvement. The known breast cancer mass was quite hypermetabolic and readily identified. No additional breast abnormality was seen. A small punctate left axillary focus of activity was also seen, corresponding to the known involved lymph node (Figure 11 and Figure 12). An additional focus of hypermetabolism in the pelvis appeared to correspond to a right ovarian cyst (Figure 13 and Figure 14). This was presumed to be a corpus luteum cyst, a known cause of normal variant benign PET activity.

FIGURE 11 Coronal (A) and sagittal (B) PET/CT images show the intense hypermetabolism of the known left breast cancer as well as the small, involved axillary lymph node.

FIGURE 12 Corresponding enhanced chest CT images (A through axilla, and B through left breast mass) show the counterpart CT findings.

FIGURE 13 Coronal PET/CT images show an intense focus of hypermetabolism in the right hemipelvis, separate from the bladder (cursors).

FIGURE 14 Corresponding enhanced pelvic CT image shows a right ovarian cyst at the same level. Corpus luteum cysts can be hypermetabolic on PET and are a known source of benign PET scan activity in the pelvis.

A second-look left breast ultrasound showed no additional anterior abnormality. It was elected to sample more anterior tissue at the time of lumpectomy and axillary dissection. The pathology showed a 2-cm infiltrating ductal carcinoma (IDC) with angiolymphatic invasion and high histologic grade (8 of 9) with close (<2 mm) deep margin and other margins clear by more than 1.5 cm. The invasive tumor was noted to extend close to the deep margin at two levels, both of which contained skeletal muscle at the deep surface. Fibrocystic changes and fibroadenomatoid changes were also noted in the specimen. Metastatic carcinoma was found in 1 of 17 lymph nodes.

The patient was additionally treated with chemotherapy (four cycles of dose-dense Adriamycin and Cytoxan (AC)) and radiation to the left breast and supraclavicular fossa. The final cycle was poorly tolerated by the patient. Planned additional chemotherapy with a taxane was declined, and the patient instead began tamoxifen.

TEACHING POINTS

This case provides many takeoff points for discussion. The characteristics of the index lesion in this case are entirely typical for malignancy with all modalities. The MRI did raise a question of more extensive disease extending anteriorly from the index lesion. The negative second-look ultrasound did not identify a correlate. This does not entirely resolve the question because correlates for MRI abnormalities which prove to be cancer are found in as few as 23% of cases. Rather than delay this patient’s planned surgery with an MRI-guided biopsy, the surgeon used the MRI information to extend the excised area, which showed fibrocystic and fibroadenomatoid changes on histology.

Preoperative identification of axillary node involvement is desirable because patients who are known to have axillary metastases can proceed directly to axillary dissection and not undergo a sentinel node sampling. As this case illustrates, malignancy in lymph nodes can be suspected and confirmed histologically even in relatively small lymph nodes. Lymph nodes should be assessed for morphologic characteristics, which may suggest malignancy. Size itself can be misleading, but thickening or formation of nodules of the cortex, mass effect, effacement or loss of the fatty hilus, and abnormal vascularity (anywhere but at the hilus) are all features that may indicate involvement. Identification of such features should lead one to consider histologic sampling preoperatively.

The pathology of this malignancy was revised from ILC at ultrasound-guided core biopsy to IDC at partial mastectomy. The particular histology should not influence one unduly in interpretation of breast MRI. In the setting of a known IDC, the segmental regional enhancement noted here anterior to the index lesion could be due to DCIS, whereas the propensity of ILC to infiltrate regionally is well known.

This case also illustrates a fairly commonly encountered source of benign PET scan activity, with the right ovarian corpus luteum cyst showing hypermetabolism. Other sources of benign PET scan activity that may be seen in the pelvis of women being imaged systemically for breast cancer include physiologic activity of the endometrium, which will be most intense during the first (menstrual) and third (ovulatory) weeks of a premenopausal patient’s cycle, and activity in uterine fibroids. Endometrial activity is a more concerning finding in a postmenopausal patient, in whom it may reflect endometrial cancer, and should prompt further evaluation. Similarly, PET hypermetabolism in an ovary is a concerning finding in a postmenopausal patient and should prompt further evaluation for possible neoplasia.

SUGGESTED READING

LaTrenta LR, Menell JH, Morris EA, et al. Breast lesions detected with MR imaging: utility and histopathologic importance of identification with US. Radiology. 2003;227:856-861.

CASE 18 Whole-body PET as an adjunct to initial staging of node-positive breast cancer: Rotter’s node involvement

A 37-year-old woman found a right UOQ palpable breast lump. Mammography showed a dominant lobular breast mass, with some margins obscured by dense adjacent parenchyma (Figure 1). Sonography demonstrated highly suspicious features, including taller-than-wide shape, and angular and irregular margins (Figure 2). In addition, an axillary lymph node with a thickened (7-mm) cortical mantle was found (Figure 3). Ultrasound-guided core needle biopsy of the dominant mass proved infiltrating ductal carcinoma (IDC), and ultrasound guided fine-needle aspiration of the axillary lymph node confirmed malignant cells, consistent with metastatic breast carcinoma.

FIGURE 1 MLO (A) and CC (B) mammograms show dense anterior breast parenchyma. A dominant mass is seen in the right upper outer quadrant (asterisk), at the level of a marker identifying a palpable lump. Some margins appear lobular, but many are obscured. MLO (C) and CC (D) spot views of the palpable lump better show the margins, which are variably circumscribed, indistinct, and obscured by adjacent parenchyma.

FIGURE 2 Ultrasound shows the mass to have highly suspicious features. It is taller than wide and hypoechoic, and the margins are angular and ill-defined, with no surrounding capsule. Ultrasound-guided biopsy confirmed IDC.

FIGURE 3 Sonography of the ipsilateral axilla shows an abnormal lymph node. Although the fatty hilus is not effaced, the cortical mantle is abnormally thick (7 mm; normal <3 mm). Well-seen, abnormal lymph nodes like this are amenable to sampling with fine-needle (20–25 gauge) aspiration (FNA). FNA of this lymph node revealed malignant cells, consistent with metastatic breast carcinoma.

Because of the size of the patient’s cancer relative to her breast size, mastectomy was favored for surgical therapy over lumpectomy. Breast MRI confirmed unifocal disease (Figures 4, 5, 6, 7, and 8). Staging workup showed elevated tumor markers (CEA and CA27.29). A positron emission tomography (PET)/CT scan and enhanced diagnostic chest, abdomen, and pelvic CT scans showed intense hypermetabolism in the known breast cancer (Figures 9, 10, 15) and abnormal right axillary lymph node (Figures 8, 11 and 12). Increased metabolic activity was also seen in a 5-mm right interpectoral (Rotter’s) lymph node (Figure 13 and Figure 14). A commonly seen normal variant source of PET activity was also seen, with symmetrical supraclavicular brown fat uptake (Figure 16).

FIGURE 4 Maximal intensity projection from the enhanced, subtracted dynamic portion of a staging breast MRI shows a large, dominant, rim-enhancing right breast cancer mass. Note the marked asymmetry in vascularity of the right breast. Small scattered foci of enhancement are noted bilaterally. The pattern and low level of enhancement of these tiny foci are typical of fibrocystic enhancement.

FIGURE 5 A slice from the dynamic series (T1-weighted, fat-saturated, three-dimensional gradient echo sequence) through the right IDC shows the margin irregularity and rim enhancement of this lesion well.

FIGURE 6 A signal intensity change versus time curve graphically confirms washout of the index tumor, another MRI feature of malignancy.

FIGURE 7 Axial T2-weighted turbo spin echo (TSE) (A) and STIR (B) images show the dominant right breast mass to be predominantly isointense to breast parenchyma. Central hyperintensity suggests necrosis.

FIGURE 8 A coronal STIR image (limited chest MRI performed in conjunction with the breast MRI, using the body coil) shows an enlarged right axillary lymph node.

FIGURE 9 Sagittal images from PET/CT (from left to right: noncontrast CT, PET, fused PET/CT; non-attenuation-corrected PET) show intense hypermetabolism of the right breast infiltrating ductal carcinoma. Activity at the inferior tip of the liver is physiologic, in bowel.

FIGURE 10 Axial images from PET/CT (upper left: noncontrast CT; upper right: non-attenuation-corrected (NAC) PET; lower left: fused PET/CT; lower right: PET) show the dominant right IDC to be intensely metabolically active. Physiologic myocardial activity is noted on these PET views.

FIGURE 15 Another slice from the enhanced chest CT shows the right IDC, with central hypodensity. Although breast cancers can often be seen on CT, the soft tissue contrast is generally insufficient to rely on CT for primary identification and local staging of breast cancers.

FIGURE 11 Axial PET/CT slices through the axilla (upper left: noncontrast CT; upper right: NAC PET; lower left: fused PET/CT; lower right: attenuation-corrected PET) show marked hypermetabolism of a right level I axillary lymph node.

FIGURE 12 Diagnostic chest CT with intravenous contrast enhancement better shows the right axillary lymph node, corresponding to PET. It is enlarged, with no discernible fatty hilus, and corresponds to the abnormal lymph node seen on ultrasound, which was malignant on FNA.

FIGURE 13 Another axial slice from PET/CT (upper left: noncontrast CT; upper right: NAC PET; lower left: fused PET/CT; lower right: attenuation-corrected PET) shows a separate small focus of Rotter’s node hypermetabolism, interposed between the right pectoralis major and minor muscles.

FIGURE 14 Corresponding section from enhanced chest CT shows the right interpectoral (Rotter’s) lymph node (arrow) to be on the order of 5 mm in size.

FIGURE 16 Coronal PET/CT images show symmetrical supraclavicular uptake of FDG, localizing on CT to fat. The appearance and distribution are typical of brown fat, which is a commonly observed normal variant source of benign activity on PET. Although it can be seen in any PET patient, younger female patients who are cold appear to be more prone to such uptake. Note also the left mediastinal brown fat uptake.

The mastectomy specimen demonstrated a 3-cm infiltrating ductal adenocarcinoma, with clear margins. Because the patient was known preoperatively to have axillary involvement, axillary dissection was performed at the same time. Three of 22 lymph nodes showed metastatic adenocarcinoma. The largest involved lymph node measured 1.5 cm.

The patient was treated postoperatively with aggressive chemotherapy (four cycles of Adriamycin and Cytoxan [AC] chemotherapy and four cycles of paclitaxel [Taxol]), followed by right chest wall, supraclavicular fossa, and posterior axillary boost radiation.

Subsequently, the patient underwent genetic testing and proved to have a BRCA1 gene mutation. Her family history consisted of premenopausal breast cancer in a maternal aunt. Ten months after finishing radiation therapy, she underwent prophylactic left mastectomy, bilateral transverse rectus abdominis myocutaneous (TRAM) flap reconstruction, hysterectomy, and oophorectomy. Her postoperative course was complicated by development of abdominal wall infection, twice requiring operative débridement of infected, necrotic tissue and intravenous antibiotic therapy (see Case 3 in Chapter 6 for imaging features of the TRAM flap donor site complications). She also developed pulmonary embolism.

Nearly concurrent with these events, the patient developed right upper arm and shoulder pain, and she noted development of an infraclavicular lump. Recurrence was subsequently confirmed in right infraclavicular and mediastinal nodal regions. See Case 7 in Chapter 10 for the recurrent disease imaging findings.

TEACHING POINTS

At initial diagnosis, the patient was considered to have stage IIB disease, T2N1M0. The increased tumor markers and evidence of axillary involvement at initial diagnosis of this breast cancer led to a comprehensive initial staging evaluation with enhanced chest, abdomen and pelvic CT scans, and PET/CT.

Historically, comprehensive imaging staging of a newly diagnosed locally advanced breast cancer (LABC) would have included a chest x-ray, bone scan, and either ultrasound or CT of the liver. Today, arguments can be made for comprehensive initial staging with PET, enhanced body CT, and bone scan. Body CT with enhancement will allow simultaneous optimal evaluation of the lung parenchyma and solid viscera (especially the liver), and the bone windows provide tomographic radiographic evaluation of axial skeletal bone. PET imaging allows for the assessment of the metabolic (glycolytic) activity of any lesions encountered on CT.

Why not do PET alone? One argument is that PET scans have a lower limit of resolution, on the order of 1 cm, depending on location in the body and a tumor’s intrinsic metabolic activity. Current state-of-the-art body PET scan units have a lower limit of resolution of 6 mm, but other limitations may prevent malignant lesions of this size from being reliably visualized. One example is in the liver, which has a moderate level of intrinsic metabolic activity. Another is in the lung bases, where there is considerable respiratory excursion.

What about PET/CT? Can it substitute for PET and enhanced body CT? One limitation of PET/CT alone is in lung parenchymal evaluation. PET/CT is acquired over multiple respiratory cycles, at tidal lung volumes, rather than at full lung expansion and suspended respiration, as in diagnostic chest CT scanning.

Because PET scans are tomographic and can image the entire body, why do bone scanning? PET scans are sensitive for identification of bone marrow involvement and lytic bone metastases, areas where bone scans frequently underestimate disease. However, blastic bone metastases that are readily demonstrated on bone scans may not show increased activity on PET.

Without the hypermetabolism displayed by the PET scan, it is doubtful that this interpectoral (Rotter’s) node would have been seen. An interpectoral (Rotter’s) lymph node is considered a level II axillary node. The PET scan activity is powerful presumptive evidence of involvement of a lymph node in a difficult position to sample.

CASE 19 Bracketing needle localization of microcalcifications

A 62-year-old woman was referred to an orthopedic surgeon for evaluation of right shoulder pain. The surgeon noted asymmetry of the left breast and nipple inversion and referred the patient for a mammogram. Mammography demonstrated extensive microcalcifications, with an associated mass, in the left upper outer quadrant (UOQ) (Figure 1 and Figure 2). Palpation-guided biopsy by a breast surgeon of UOQ firmness did not confirm a malignancy, and the patient had an excisional surgical biopsy after needle localization (Figure 3 and Figure 4). The pathology showed an estrogen receptor–positive, HER-2/neu-positive, 1.5-cm infiltrating ductal carcinoma (IDC), associated with solid and comedo ductal carcinoma in situ (DCIS), which extended to the margin of resection. An intramammary lymph node removed in the specimen was involved with tumor. Mastectomy was performed subsequently, showing extensive residual comedo DCIS in the specimen and negative margins. Three of 15 lymph nodes proved to be positive. Tumor stage was IIA, T1N1. The patient was further treated with chemotherapy and 5 years of tamoxifen therapy. Subsequent aromatase inhibitor therapy with letrozole was discontinued because of side effects.

FIGURE 1 Mammograms [MLO (A) and CC (B)] show dense breast tissue, with extensive microcalcifications and a suggestion of an underlying mass in the left upper outer quadrant (UOQ), better seen in detail (Figure 2.

FIGURE 2 Detail of the left upper breast, MLO projection, shows an ill-defined mass with pleomorphic microcalcifications.

FIGURE 3 Images [lateral (A) and CC (B)] from the needle localization. An attempt was made to encompass the entire area with bracketing localization wires, placed at opposite ends of the abnormality.

FIGURE 4 Specimen x-ray shows a large mass within the specimen, with many pleomorphic calcifications, with the localizing wires at opposite ends of the specimen.

TEACHING POINTS

What a difference a few years makes! This patient was treated in 1998. Palpable firmness in the same quadrant as the highly suspicious mammographic findings led appropriately to palpation-guided biopsy, but the suspected malignancy was not confirmed. Today, this scenario would be cause for referral for an image-guided biopsy. Ultrasound would be an important component of the workup, to identify any associated mass. When no mass is seen on mammography within a large area of suspicious microcalcifications, there is utility in evaluating the area with ultrasound to determine whether a mass can be identified to target for biopsy. If there is a mass seen on ultrasound, sampling it with ultrasound guidance may increase the likelihood of obtaining a diagnosis of invasive cancer than sampling microcalcifications stereotactically.

In this case, the patient had a diagnostic surgical excision biopsy. The abnormality is extensive. An attempt to encompass its full extent was made by bracketing the abnormality with more than one localization needle. From the specimen x-ray, we can predict the margins will be involved.

With a smaller abnormality, the bracketing needle localization illustrated here may be more successful, helping guide the surgeon’s excision and, in some cases, achieving clear margins in fewer steps.

CASE 20 Medial breast cancer with internal mammary drainage on lymphoscintigraphy ^*

A new, suspicious mass was identified in the lower inner quadrant of the right breast in an 83-yearold woman on screening mammogram (Figure 1 and Figure 2). Biopsy was performed under ultrasound guidance, confirming mucinous carcinoma. A clip placed at the time of biopsy was subsequently used to guide needle localization for excision. Lymphoscintigraphy was performed at the time, through peritumoral and subdermal injections. Imaging was carried out to 4 hours and showed evidence of drainage to a right lower internal mammary lymph node (Figure 3).

FIGURE 1 MLO (A) and CC (B) mammograms show a dense mass in the posterior lower inner quadrant of the right breast, which is readily visualized in this fatty portion of the breast.

FIGURE 2 The margins are seen to be indistinct on spot compression.

FIGURE 3 Lymphoscintigraphy shows intense activity at the peritumoral injection site. A small focus of separate activity is inferomedial to the tumor site (arrow), suggesting drainage of this medial tumor to an internal mammary lymph node.

TEACHING POINTS

There is little consensus in the literature on the optimal technique for breast cancer sentinel lymph node identification, other than that utilization of both radiopharmaceutical and intraoperative blue dye injection identifies more sentinel lymph nodes than protocols using either approach solely. A variety of radiopharmaceutical injection techniques, including peritumoral, subareolar, and in-tradermal injections, solely or in combination, have been advocated. The good news is that it seems that all these methods can work. Povoski and colleagues report a single-institution, prospective trial of 400 patients randomized to undergo either intradermal, intraparenchymal, or subareolar injection of ^99mTc-sulfur colloid administration for sentinel lymph node mapping and biopsy in breast cancer. In this series, intradermal injection demonstrated significantly greater frequency of localization, decreased time to first localization on preoperative lymphoscintigraphy, and decreased time to harvest the first sentinel node.

Performance of lymphoscintigraphy is not universal. In this example, peritumoral and subdermal injections were performed, and imaging was carried out to 4 hours. At Scripps Clinic, our practice is to perform a single intradermal injection of 500 μCi of filtered ^99mTc-sulfur colloid in the smallest volume possible, administered at the areolar edge of the breast cancer–involved quadrant.

In this case, lymphoscintigraphy suggests internal mammary node drainage, not a surprising result given the lower inner quadrant location of this breast cancer. Because no clear advantage has been demonstrated from surgical series of extended internal mammary dissections, what to do with this information remains controversial. Even with-out lymphoscintigraphic evidence of internal mammary sentinel node drainage, many radiation oncologists would plan for inclusion of internal mammary basins in designing treatment ports for such far medial lesions. This case also illustrates one of the possible limitations of lymphoscintigraphic imaging: if the draining lymph node is close to the primary and peritumoral injection has been performed, the primary site injection activity may obscure uptake in a draining lymph node. In such a case, imaging with the primary site shielded may aid in recognition of sentinel nodes.

SUGGESTED READING

Povoski SP, Olsen JO, Young DC, et al. Prospective randomized clinical trial comparing intradermal, intraparenchymal, and subareolar injection routes for sentinel lymph node mapping and biopsy in breast cancer. Ann Surg Oncol. 2006;13(11):1412-1421.

CASE 21 Biopsy quality control: Mammographic lesion, wrong ultrasound correlate biopsied; rationale for post–ultrasound biopsy clip placement and mammogram

An asymptomatic 71-year-old woman underwent yearly screening mammography, which showed a small mass in the posterior lateral right breast (Figure 1). Ultrasound showed a small irregular mass in the 9-o’clock position, thought to correspond to the mammographic finding (Figure 2). An ultrasound-guided 14-gauge core needle biopsy was performed. Because of the lesion’s small size, a marker clip was placed. A postbiopsy mammogram showed that the marker clip location did not correspond to the lesion seen on mammography (Figure 3). Subsequently, additional ultrasound imaging demonstrated a second lesion at 8 o’clock (Figure 4). A second core biopsy and clip placement was performed. On the postbiopsy mammogram, this clip conformed to the site of the original mammographic lesion (Figure 5). Pathology reported atypical ductal hyperplasia for the 9-o’clock lesion and invasive ductal carcinoma for the 8-o’clock lesion. The patient went on to have a preoperative MRI (Figure 6 and Figure 7). This showed a small contralateral left breast lesion. A correlate was found on second-look ultrasound (Figure 8). On core biopsy, invasive ductal carcinoma was confirmed. The patient decided to undergo bilateral mastectomies.

FIGURE 1 Right MLO (A) and exaggerated CC (B) mammographic views show a small mass (arrow) in the posterior lateral breast.

FIGURE 2 Ultrasound image of the right breast at 9 o’clock demonstrates a small 5- to 6-mm mass with irregular margins and posterior acoustic shadowing (arrow). This was thought to correspond to the mammographic finding.

FIGURE 3 Post-biopsy right breast 90-degree lateral mammographic view demonstrates that the biopsy marker clip location does not correlate with the mammographic lesion (arrows).

FIGURE 4 Ultrasound image of the right breast. Upon further evaluation, a second suspicious lesion was seen on ultrasound, in the 8-o’clock position.

FIGURE 5 Postbiopsy mammogram following the second core biopsy demonstrates that the lesion in the 8-o’clock position on ultrasound corresponds to the original mammographic lesion (arrow).

FIGURE 6 Subtracted axial postcontrast MRI demonstrates a lobular enhancing mass at 8 o’clock in the right breast (arrow), corresponding to the known carcinoma.

FIGURE 7 Subtracted postcontrast MRI image also showed a small enhancing round 5-mm mass in the medial left breast (arrow).

FIGURE 8 Second-look ultrasound image of the medial left breast. A small hypoechoic solid mass in seen at 9 o’clock, corresponding to the MRI finding. Subsequent core biopsy showed invasive ductal carcinoma.

TEACHING POINT

This case illustrates the importance of careful correlation of findings seen on ultrasound with suspicious mammographic findings. In this example, the use of postbiopsy marker clips allowed the location of ultrasound findings to be correlated with the original mammographic abnormality. This is particularly important with small or subtle mammographic findings. It is also important to remember that a significant number of breast cancers are multifocal or multicentric at presentation. In this case, the patient also had a mammographically occult contralateral carcinoma.

CASE 22 Biopsy quality control: DCIS presenting as disappearing microcalcifications and subsequent development of a mass

An 82-year-old woman presented with clustered suspicious microcalcifications in the lateral left breast on routine screening mammography. After magnification, stereotactic biopsy was recommended (Figure 1 and Figure 2). The biopsy was technically difficult because the patient had trouble tolerating the biopsy position. Multiple sets of core samples were obtained; however, no microcalcifications were identified in the samples (Figure 3 and Figure 4). Pathology results indicated benign breast tissue with no calcifications. Given the patient’s age, 6-month follow-up was recommended over surgical excisional biopsy. At follow-up, the calcifications were seen to decrease in number, and there was interval development of an associated irregular solid mass (Figure 5). Subsequent ultrasound-guided core needle biopsy revealed an intermediate-grade invasive ductal carcinoma and DCIS (Figure 6). The patient was treated surgically with sentinel node biopsy and lumpectomy.