03 - radiol clin n am 2007 - imaging breast cancer

R A D I O L O G I CC L I N I C S

O F N O R T H A M E R I C A

Radiol Clin N Am 45 (2007) 45–67

45

Imaging Breast CancerLia Bartella, MD, FRCRa,b,*, Clare S. Smith, MB, BCh, BAO, FRCRa,D. David Dershaw, MDa,b, Laura Liberman, MDa,b

- Breast cancer screening- Percutaneous image-guided biopsy- Guidance

StereotaxisUltrasoundMR imaging

- Percutaneous biopsy: future directions- Staging breast cancer

Preoperative stagingSentinel lymph node biopsy

- Breast MR imagingIndications for the use of breast MR

imagingNeoadjuvant chemotherapy responseAssessment of residual diseaseTumor recurrence at the lumpectomy siteOccult primary breast cancer

- Proton MR spectroscopy of the breast

The use of proton MR spectroscopy in thebreast

Differentiating benign from malignantbreast lesions

Characterization of histopathologicsubtypes

Evaluation of normal and lactating breastparenchyma

Predicting response to neoadjuvantchemotherapy

MR spectroscopy of axillary lymph nodesin breast cancer patients

High-resolution magic angle spinning MRspectroscopy of breast tissue

- Positron emission tomography and breastcancer

- Other new techniques- Summary- References

Breast cancer is now the most common nonskincancer in women in the United States. Women havean average risk of one in eight of being diagnosedwith breast cancer at some time in their lives. Al-though the breast cancer diagnosis rate has in-creased, there has been a steady drop in theoverall breast cancer death rate since the early1990s [1], most likely due to a combination ofscreening, improved treatments, and betterawareness.

Invasive ductal carcinoma is the most commonbreast cancer histologic type, accounting for 70%

0033-8389/07/$ – see front matter ª 2006 Elsevier Inc. All righradiologic.theclinics.com

to 80% of all cases. Invasive lobular carcinoma isthe second most common histologic type (5% to10% of all breast cancers). It is associated witha high rate of multifocality and bilaterality and canbe difficult to diagnose clinically and mammo-graphically because of its tendency to spread dif-fusely through breast tissue instead of forminga mass and causing architectural distortion. Otherless common cancers include tubular, medullary,mucinous, and papillary cancers. Cystosarcoma,phyllodes, angiosarcoma, and lymphoma also occurin the breast but are not considered typical breast

a Department of Radiology, Breast Imaging Section H-118, Memorial Sloan-Kettering Cancer Center, 1275York Avenue, New York, NY 10021, USAb Weill Medical College of Comell University, New York, NY, USA* Corresponding author. Memorial Sloan-Kettering Cancer Center, 1275 York Avenue, New York, NY 10021.E-mail address: [email protected] (L. Bartella).

ts reserved. doi:10.1016/j.rcl.2006.10.007

mailto:[email protected]

Bartella et al46

cancers. Inflammatory cancer is diagnosed clinicallybased on the association with edema, erythema, andskin dimpling. Paget’s disease is a relatively rare dis-ease, affecting the nipple–areolar complex. It ac-counts for 1% of all breast cancer cases.

In situ carcinoma is contained within the duct,and the basement membrane surrounding theduct is not breached. Ductal carcinoma in situ(DCIS) originates from the major lactiferous ducts.Approximately 30% to 50% of patients who haveDCIS will develop invasive ductal carcinoma overa 10-year period [2]. Lobular carcinoma in situ(LCIS) arises from the terminal duct lobule andcan be distributed diffusely throughout the breast.In contrast to DCIS, women who have LCIS haveup to 30% risk of developing invasive carcinoma,mostly of the ductal type and with equal frequencyin both breasts [3]. Therefore, LCIS is considereda marker of increased risk rather than a precursorof breast cancer.

Controversy exists around the diagnosis andtreatment of DCIS, particularly in relation toscreening and the phenomenon of overdiagnosis(finding early neoplasms, of which many wouldnot become clinically evident if screening had notoccurred). It has been estimated that 1 in 3 in situtumors are overdiagnosed at the first screen and 1in 25 are overdiagnosed at subsequent screens [4].It is not yet possible to say which patients whohave DCIS will go on to develop invasive cancersand whether survival rates would be the same if sur-gery were undertaken only after early invasive can-cer had been diagnosed (Tables 1–3). Thediagnosis and management of breast cancer has un-dergone tremendous changes over the years. Themammogram has taken over from clinical examina-tion in the diagnosis of breast cancer. Ultrasoundand stereotactic biopsy have replaced many surgicalbiopsies, and early detection of breast cancer has re-sulted in breast conservation and sentinel lymphnode biopsy, replacing the radical mastectomyand axillary lymph node dissection. Mammographyremains the traditional first-line radiologic test ofchoice in the detection and diagnosis of breast can-cer; however, mammography is not perfect. About10% of cancers are mammographically occulteven after they are palpable and, in women whohave dense breasts, the sensitivity of mammogra-phy can be as low as 68% [5]. This low sensitivityhas led to the expansion of breast imaging to in-clude sonography and MR imaging and the devel-opment of newer imaging techniques such aspositron emission tomography (PET), lymphoscin-tigraphy, scintimammography, breast tomosynthe-sis, and contrast-enhanced mammography to aidin the detection and staging of breast cancer andto monitor response to therapy.

Mammography is used for diagnostic and screen-ing purposes. Diagnostic mammography is com-monly used to identify possible breast cancers inwomen who present with signs or symptoms andit has higher sensitivities (85%–93%) comparedwith screening mammography [6,7]. Tumors de-tected by diagnostic mammography are larger andmore likely to be node positive than those detectedby screening mammography [8]. The last decadehas seen the development of full-field digital mam-mography. Digital mammography devices are simi-lar to film-screen units except that the film-screencassette used to record the image is replaced bya digital detector. Digital mammography has a num-ber of advantages over traditional film-screen mam-mography. It has a higher contrast resolution yetmaintains a good dynamic range. It allows for dig-ital transmission and storage of images, eliminatingthe need for the film library. The images can be ma-nipulated to enhance visualization of subtle struc-tures and calcifications, and the procedure isquicker for the patient because there are no waittimes for the films to be processed. It also elimi-nates film artifacts such as dust and uses a lowerdose of radiation [9]. The major disadvantage ofdigital mammography is cost, with digital systemscurrently costing approximately one to four timesas much as film-screen systems. The results of thelargest trial to date comparing digital versus filmmammography for breast cancer screening, the Dig-ital Mammographic Imaging Screening Trial [10],were recently published. In this multicenter trial,the investigators found that digital mammographywas better than conventional film mammographyat detecting breast cancer in young, premenopausal,and perimenopausal women and in women whohave dense breasts; however, there was no signifi-cant difference in diagnostic accuracy between dig-ital and film mammography in the population asa whole or in the other predefined subgroups.

Breast sonography is well established as a valu-able imaging technique. The current indicationsfor performing breast ultrasound, as listed in the‘‘ACR Practice Guideline for the Performance ofBreast Ultrasound Examination,’’ include identifi-cation and characterization of palpable and non-palpable abnormalities, evaluation of clinicaland mammographic findings, guidance of inter-ventional procedures, evaluation of breast im-plants, and treatment planning for radiationtherapy [11]. It is also the imaging technique ofchoice to evaluate palpable masses in womenyounger than age 30 years and in lactating andpregnant women. Its advantage lies in the factthat it is easily accessible, relatively low in cost,and does not involve the use of ionizing radia-tion. Its main disadvantage is that its performance

Breast Cancer Imaging 47

Table 1: TNM staging system

Primary tumor – TTX Primary tumor cannot be assessedT0 No evidence of primary tumorTis Carcinoma in situTis (DCIS) Ductal carcinoma in situTis (LCIS) Lobular carcinoma in situTis (Paget’s) Paget’s disease of the nipple with no tumor (note: Paget’s disease associated

with a tumor is classified according to the size of the tumor)T1 Tumor %2.0 cm in greatest dimensionT1mic Microinvasion %0.1 cm in greatest dimensionT1a Tumor >0.1 cm but %0.5 cm in greatest dimensionT1b Tumor >0.5 cm but %1.0 cm in greatest dimensionT1c Tumor >1.0 cm but %2.0 cm in greatest dimensionT2 Tumor >2.0 cm but %5.0 cm in greatest dimensionT3 Tumor >5.0 cm in greatest dimensionT4 Tumor of any size with direct extension to (a) chest wall or (b) skin, only as

described belowT4a Extension to chest wall, not including pectoralis muscleT4b Edema (including peau d’orange) or ulceration of the skin of the breast, or

satellite skin nodules confined to the same breastT4c Both T4a and T4bT4d Inflammatory carcinoma

Regional lymph nodes – NNX Regional lymph nodes cannot be assessed (eg, previously removed)N0 No regional lymph node metastasisN1 Metastasis to movable ipsilateral axillary lymph node(s)N2 Metastasis to ipsilateral axillary lymph node(s) fixed or matted, or in clinically

apparent ipsilateral internal mammary nodes in the absence of clinicallyevident axillary lymph node metastasis

N2a Metastasis in ipsilateral axillary lymph nodes fixed to one another (matted)or to other structures

N2b Metastasis only in clinically apparent* ipsilateral internal mammary nodesand in the absence of clinically evident axillary lymph node metastasis

N3 Metastasis in ipsilateral infraclavicular lymph node(s) with or without axillarylymph node involvement, or in clinically apparent* ipsilateral internalmammary lymph node(s) and in the presence of clinically evident axillarylymph node metastasis; or, metastasis in ipsilateral supraclavicular lymphnode(s) with or without axillary or internal mammary lymph nodeinvolvement

N3a Metastasis in ipsilateral infraclavicular lymph node(s)N3b Metastasis in ipsilateral internal mammary lymph node(s) and axillary lymph

node(s)N3c Metastasis in ipsilateral supraclavicular lymph node(s)

Distant metastasis – MMX Distant metastasis cannot be assessedM0 No distant metastasisM1 Distant metastasis

Histologic grade – GGX Grade cannot be assessedG1 Low combined histologic grade (favorable)G2 Intermediate combined histologic grade (moderately favorable)G3 High combined histologic grade (unfavorable)

is operator dependent and it can be time-consuming.

Several studies have shown that breast sonogra-phy can help distinguish benign from malignantsolid nodules [12] and that the use of ultrasoundas an adjunct to mammography has led to an

overall increase in diagnostic accuracy [13]. Studieson the impact of ultrasound have also shown thatits use can affect management in 64% of patientsand prevent unnecessary biopsies in 22% [14]. Ul-trasound is also useful in the assessment of the ax-illa in a patient who has newly diagnosed breast

Bartella et al48

cancer. If lymph nodes are seen to have cortical con-tour bulges or masses, then ultrasound-guided per-cutaneous needle biopsy can confirm metastaticinvolvement, obviating the need for sentinel lymphnode biopsy [15].

An American College of Radiology Imaging Net-work trial (Protocol 6666) is now underway to as-sess the efficacy of screening breast sonography.The primary aim of this multicenter protocol is todetermine whether screening whole-breast sonog-raphy can identify mammographically occult can-cers and whether such results can be generalizedacross multiple centers.

Breast cancer screening

Abundant evidence has accumulated over the past 4decades to support the ability of mammographicscreening to decrease breast cancer mortality. Sevenprospective randomized trials have been con-ducted. Study designs of these trials have differed,with variable intervals between mammographicscreenings, variable ages at invitation to screeningand cessation of screening, and even with varyingmammographic techniques. (Table 4) [16].

Table 2: Stage grouping

Stage Tumor Node Metastasis

0 Tis N0 M0I T1 N0 M0IIA T0 N1 M0

T1 N1 M0T2 N0 M0

IIB T2 N1 M0T3 N0 M0

IIIA T0 N2 M0T1 N2 M0T2 N2 M0T3 N1 M0T3 N2 M0

IIIB T4 N0 M0T4 N1 M0T4 N2 M0

IIIC Any T N3 M0IV Any T Any N M1

Table 3: Stage and 5-year survival rate

Stage Rate (%)

0 100I 100IIA 92IIB 81IIIA 67IIIB 54IV 20

Because of these differences, the conclusions ofthese studies have varied from trial to trial. Basedon these trials, however, there is little doubt thatmammographic screening has efficacy. A recentmeta-analysis of data from these seven trials showsa 24% mortality reduction in women invited toscreening.

The estimation of mortality reduction with mam-mographic screening is generally considered to beunderestimated by these trials because of noncom-pliance of those invited to screening and contami-nation of the control groups (the womenincluded in these studies who were not invited tobe screened). The percentage of women invited toscreening who actually underwent screening wasas low as 67%. Although they underwent somescreening, many women did not undergo all thescreening mammography to which they were in-vited. This lack of compliance degrades the impactof screening on breast cancer mortality in all stud-ies. In addition, of the women not invited to screen-ing, some underwent mammographic screeningoutside of the study situation, further decreasingthe study’s estimate of the impact ofmammography.

Results of prospective randomized trials havenow been augmented by experience with popula-tion-based screening. In the Uppsala region of Swe-den, a comparison of breast cancer mortality beforeand after the introduction of mammographicscreening has estimated a 39% mortality reductiondue to screening [17]. In Italy, the rate of fatal breastcancer cases has been reduced by 50% with the in-troduction of mammographic screening [18]. Thesedata suggest that the impact of screening may begreater than that estimated by prospective random-ized trials.

Although the benefit of mammographic screen-ing is widely accepted, limitations and adverseeffects from screening are also generally acknowl-edged and include biopsies to diagnose benignlesions, anxiety about mammography and biopsyresults, scarring from biopsies, and time lost fromwork to undergo screening and follow-up. Of biop-sies done on the basis of mammographic abnor-malities, only 25% to 45% result in a diagnosis ofcarcinoma.

The failure of mammography to detect all breastcancers is also widely acknowledged, with the false-negative rate of screening mammography usually inthe 20% to 30% range. Tumors without associatedcalcifications and subtle masses are particularly dif-ficult to diagnose. Invasive lobular carcinoma anduncalcified DCIS are especially difficult to detectwith mammography.

Despite these limitations, mammography hasbeen incorporated in the routine medical care of


Table 4: Results of prospective randomized trials of mortality reduction by mammographic screening

StudyYearbegun

Age ofwomen (y)

Mammographyinterval (mo)

% Participationof invited women

% Mortalityreduction (95% CI)

HIP 1963 40–64 12 67 24 (7–38)Two county,Sweden

1977 40–74 24 89 32 (20–41)

Malmo 1976 45–69 18–24 74 19 (�8–39)Stockholm 1981 40–64 24 81 26 (�10–50)Gothenburg 1982 39–59 18 84 16 (�39–49)Canada NBSS1 1980 40–49 12 100 �3 (�26–27)Canada NBSS2 1980 50–59 12 100 �2 (�33–22)All trialscombined

24 (18–30)

Abbreviations: CI, confidence interval; HIP, health insurance plan of greater NY; NBSS, National Breast Cancer Screeningstudy.Data from Smith RA, Saslow D, Sawyer KA, et al. American Cancer Society guidelines for breast cancer screening: update2003. CA Cancer J Clin 2003;53:141–69; and Heywang-Koebrunner SH, Dershaw DD, Schreer I. Diagnostic breast imaging.2nd edition. New York: Thieme; 2001.

women. The usual recommendation for mammo-graphic screening in the United States is currentlyannual mammography starting at age 40 years[19]. No upper age limit has been applied to thescreening recommendation in the United States.

In women at higher risk than the general popula-tion, screening for the development of breast cancermay be more aggressive. Women at highest risk arethose who are gene positive on testing for BRCAgenes or who have a very strong family history.These histories include multiple first- and second-degree relatives who have had breast or ovarian car-cinoma, a first-degree relative who has had breastcancer before age 50 years, male relatives whohave had breast cancer, and Ashkenazi Jewishwomen who have a family history of breast or ovar-ian cancer. In some families, gene-positive womenhave been calculated to have up to an 85% lifetimerisk of developing breast cancer. In families inwhich premenopausal breast cancer develops, ithas been recommended that women should startscreening 10 years earlier than the youngest age atwhich breast cancer was diagnosed, starting as earlyas age 25 years. Because breast cancers in youngerwomen may grow more quickly and because famil-ial breast cancers may grow more quickly than spo-radic cancers, it has been suggested that screeningmay be useful more frequently than every 12months in this population. Although 6-monthmammographic screening has been suggested forthese women, there are no data to indicate whetherit is of any advantage over annual examinations.

Other women at significantly higher risk includethose who have a personal history of breast canceror prior biopsy diagnosis of atypical ductal hyper-plasia (ADH) or LCIS. Screening for these women

should commence at the time of diagnosis. Womentreated for Hodgkin’s disease with mantle radiationare at risk for developing radiation-induced breastcancer and should commence screening as earlyas 8 years after their cure [20].

The addition of other imaging modalities tomammography in the screening algorithm forhigh-risk women has undergone some study, butscreening with nonmammographic imaging re-mains controversial. It should be rememberedthat mammography is the cornerstone of breastcancer screening, and there are no recommenda-tions that it be abandoned for other screening mo-dalities. The ability of mammography to detectsubcentimeter carcinomas based on easily identi-fied microcalcifications has not been replaced byany other screening tool [21].

When used in a high-risk population, there aredata to suggest that sonography and MR imagingcan detect early, curable cancers not found by mam-mography. In a study of Dutch women who hada genetic predisposition to develop breast cancer,MR imaging was able to find more cancers thanmammography at initial and follow-up screenings.Cancers found by both modalities had a similarprognosis, and the positive predictive values ofmammography and MR imaging were comparable[22]. Other studies from the United States and Eu-rope support these results. Data for sonographicscreening are less compelling, suggesting a sensitiv-ity that is inferior to MR imaging and an inability todetect most in situ disease. There may also bea lower positive predictive value for sonographi-cally recommended biopsies than those based onmammographic or MR imaging findings. The wideravailability and lesser cost of sonography compared

Bartella et al50

with MR imaging, however, may make it the studyof choice in some situations when a second modal-ity is desired.

Percutaneous image-guided biopsy

Percutaneous image-guided biopsy is increasinglyused as an alternative to surgical biopsy for the his-tologic assessment of breast lesions [23]. Guidancefor percutaneous biopsy may be provided by stereo-taxis, ultrasound, or MR imaging. Most often,stereotactic guidance is used for biopsy of calcifica-tions; ultrasound guidance for biopsy of masses;and MR imaging guidance for lesions identifiedonly with breast MR imaging. Tissue acquisitionfor percutaneous biopsy is usually accomplishedwith automated core needles or vacuum-assisted bi-opsy probes. Vacuum-assisted probes are preferablefor stereotactic-guided or MR imaging–guided bi-opsies; for ultrasound-guided biopsies, automatedcore needles and vacuum-assisted probes are useful.For small imaging lesions in which the imaging tar-get is removed at percutaneous biopsy, placementof a localizing marker is helpful.

Percutaneous image-guided biopsy, comparedwith surgical biopsy, is faster, less invasive, hasfewer complications, and causes minimal to noscarring. Percutaneous biopsy spares the need forsurgical biopsy in approximately 80% of patients,often obviating surgery in women who have benigndisease and decreasing the number of operationsnecessary in women who have breast cancer [24].Finally, percutaneous biopsy, compared with surgi-cal biopsy, has a lower cost of diagnosis. At the au-thors’ facility, the cost of diagnosis was decreased by20% by using stereotactic 11-gauge vacuum-assistedbiopsy and by more than 50% with ultrasound-guided 14-gauge automated core biopsy [25,26].

Guidance

Stereotaxis

Stereotactic biopsy is based on the principle that thethree-dimensional location of a lesion can be as-sessed based on its apparent positional change ontwo angled (stereotactic) images. Validation studiesof stereotactic 14-gauge automated core biopsydemonstrated 87% to 96% concordance betweenresults of stereotactic and surgical biopsies; obtain-ing multiple specimens with a long excursion gunand a dedicated table yielded the best results [27].Although prone tables provide more workingroom, decrease the likelihood of patient motionor vasovagal reaction, and provide a physical andpsychologic barrier between the patient and theprocedure than upright units, they are more expen-sive and require more space. Digital imaging is

valuable in stereotactic biopsy because it enablesimage processing that improves lesion conspicuityand shortens procedure times.

Stereotactic biopsy is most often used for lesionsevident as calcifications. Vacuum-assisted biopsyprobes are preferable to automated core biopsy nee-dles for tissue acquisition of calcifications at stereo-tactic biopsy because they provide better retrieval ofcalcifications [28] and better characterization ofcomplex lesions such as ADH and DCIS, lesionsthat are often evident as calcifications at mammog-raphy [29,30]. ‘‘Underestimation’’ is defined as thediagnosis of cancer at surgery in a lesion thatyielded ADH at percutaneous biopsy or the diagno-sis of invasive cancer at surgery in a lesion thatyielded DCIS at percutaneous biopsy. The likeli-hood of underestimation is significantly loweramong lesions that undergo stereotactic 11-gaugevacuum-assisted biopsy compared with 14-gaugeautomated core biopsy [31,32]. There is a learningcurve for stereotactic biopsy, with better results ob-tained after the first 5 to 20 cases for 14-gauge auto-mated core biopsy and after the first 5 to 15 casesfor 11-gauge vacuum-assisted biopsy [33]. In a vali-dation study of stereotactic 11-gauge vacuum-assis-ted biopsy, false-negative cases were encountered in3% of all cancers; the false-negative rate was signif-icantly higher among radiologists who had previ-ously performed fewer than 15 cases rather than15 or more cases (10% versus 0.6%, P < .01) [32].

Ultrasound

Ultrasound-guided biopsy, first performed with 14-gauge automated core needles, is a fast, safe, and ac-curate procedure. Ultrasound guidance has numer-ous advantages over stereotactic guidance,including speed, multipurpose equipment, lack ofionizing radiation, accessibility of all areas of thebreast and axilla, real-time needle visualization,multidirectional sampling, and lower cost. Themain disadvantage of ultrasound guidance is thatsonographically inapparent lesions (eg, specific le-sions evident as calcifications or masses that arenot visualized on ultrasound) may not be amena-ble to ultrasound-guided biopsy. Ultrasound-guided biopsy can also be performed withvacuum-assisted devices [33]. The vacuum-assistedbiopsy devices are faster and more often achievecomplete excision of the imaging target but haveno other significant benefit compared with14-gauge automated core needles for ultrasound-guided biopsy [34].

Ultrasound-guided biopsy is generally performedfor sonographically evident masses initially identi-fied by imaging or palpation. A mass identified atmammography and ultrasound could potentiallyundergo biopsy under stereotactic or ultrasound


guidance; however, ultrasound guidance is oftenpreferable due to shorter procedure time, lack ofionizing radiation, and lower cost. Soo and col-leagues [35] found that a subset (23%) of calcificlesions had a sonographic correlate and were there-fore amenable to percutaneous biopsy under ultra-sound guidance [35]. A sonographic correlate wasmore likely in calcific lesions categorized as BreastImaging Reporting and Data System (BI-RADS) 5,highly suggestive of malignancy, compared withcalcific lesions classified as BI-RADS 4, suspiciousof malignancy (89% versus 17%, P < .001). Target-ing the sonographic mass (if present) associatedwith the calcifications may facilitate diagnosis ofthe invasive component of a lesion containing inva-sive cancer and DCIS, thereby decreasing the fre-quency of underestimation. If screening breastultrasound proves to be efficacious, then ultra-sound-guided biopsy will be the method of choicefor diagnosis of lesions identified at screeningsonography [36].

MR imaging

MR imaging can demonstrate breast cancers that arenot identified by mammography, sonography, orphysical examination. The specificity of breast MRimaging, however, is limited, ranging from 37%to 97%. Furthermore, among lesions identified atMR imaging that warrant biopsy, ultrasound failsto reveal a sonographic correlate in up to 77%[37]. To benefit from breast MR imaging, it is neces-sary to have the capability to perform biopsy of le-sions identified with MR imaging only. MRimaging–guided percutaneous breast biopsy posesseveral challenges, including the necessity to re-move the patient from the closed magnet to per-form the biopsy, limited access to the medialbreast tissue, the transient nature of contrast en-hancement, and the difficulty in confirming lesionretrieval [38]. Dedicated MR imaging–guided bi-opsy equipment has been developed to overcomesome of these challenges, including coils, breast im-mobilization and compression devices, needleguides, localizing markers, and nonferromagneticneedles with minimal artifact.

MR imaging–guided biopsy is more expensivethan biopsies done under stereotactic or ultrasoundguidance and is generally reserved for lesions iden-tified only at MR imaging. Pioneered in Europe[39], MR imaging–guided vacuum-assisted biopsyhas been further refined in the United States [40–43]. For MR imaging–guided percutaneous biopsy,vacuum-assisted biopsy probes have several advan-tages over automated core needles for tissue acqui-sition: they are faster, acquire a larger volume oftissue, and provide more accurate characterizationof lesions such as ADH and DCIS, lesions that are

encountered more frequently among the high-riskpatients undergoing breast MR imaging examina-tion than among the general population. In pub-lished experience from the authors’ institution,the median time to perform MR imaging–guidedbiopsy of a single lesion was 33 minutes; MR imag-ing–guided biopsy histology yielded cancer in 25%(Fig. 1), with more than half of the cancers beingDCIS [41].

Percutaneous biopsy: future directions

Percutaneous biopsy provides an excellent alterna-tive to surgery for histologic diagnosis. Furtherwork is necessary to optimize lesion selection forbiopsy, to refine the equipment and techniquesfor performing percutaneous biopsy, to developevidence-based criteria to optimize the biopsymethod for specific clinical scenario, to evaluatecost-effectiveness of different biopsy procedures,and to assess long-term outcomes. These studieswill allow more women to benefit from the use ofminimally invasive techniques to diagnose benignand malignant lesions of the breast.

Staging breast cancer

Preoperative staging

Traditional preoperative planning for breast cancerinvolves clinical examination and mammogram. Itwas shown from a study of 282 mastectomy speci-mens [44] (performed for unifocal breast cancer)assessed clinically and mammographically thatmost breasts (63%) had additional sites of cancerthat were undetected by clinical examination ormammography. Additional foci of cancer werefound pathologically within 2 cm of the index can-cer in 20% and greater than 2 cm away from the in-dex cancer in 43%. Seven percent had additionalfoci of carcinoma more than 4 cm away from the in-dex cancer, which likely represent cancer withina separate breast quadrant. The presence of unde-tected residual disease that is not removed entirelyat surgery is the rational for performing postopera-tive radiation therapy in patients who are treatedwith breast conservation therapy.

It is known that disease may or may not be leftbehind in the breast; however, it has not been pos-sible without the addition of MR imaging to reli-ably identify which patients have additionalmultifocal or multicentric cancer. Many studieshave shown that MR imaging is able to detect addi-tional foci of cancer in the breast that has been over-looked by conventional techniques. Severalinvestigators have shown that MR imaging is ableto detect additional foci of disease in up to onethird of patients [45,46], which may possibly result

Bartella et al52

Fig. 1. A 62-year-old asymptomatic woman who has a strong family history of breast cancer (sister at ages 53 and69 years, daughter at age 33 years) and moderately dense breasts without suspicious findings at mammography(not shown). (A) Sagittal, fat-suppressed, T1-weighted MR image of the right breast obtained within 2 minutesafter injection of intravenous gadolinium at high-risk screening MR imaging examination shows an intenselyenhancing 0.7-cm mass in the right breast upper inner quadrant. (B) Sagittal, T1-weighted, fat-suppressed, de-layed image of the right breast after contrast injection during the high-risk screening MR imaging examinationshows washout of contrast from the right breast upper inner quadrant mass. This lesion, which had no mammo-graphic or sonographic correlate, was interpreted as suspicious for carcinoma, and biopsy was recommended.(C) On the day of MR imaging–guided vacuum-assisted biopsy, a sagittal, T1-weighted, fat-suppressed scout im-age demonstrates that the lesion is still present. (D) Images obtained after tissue acquisition and clip placementshow low signal artifact from air and clip at the biopsy site. The lesion has been sampled. (E) Collimated mam-mographic image of the right breast after biopsy and clip placement demonstrates air and the clip that has de-ployed in the breast. Histologic analysis yielded invasive ductal carcinoma. (F) Mediolateral mammographicimage of the right breast on the day of breast conserving surgery demonstrates preoperative localizationof the clip under mammographic guidance. Surgery yielded invasive ductal carcinoma, 0.5 cm, adjacent to theneedle biopsy site. The sentinel nodes were free of tumor.

in a treatment change [47]. MR imaging can poten-tially provide valuable information for preoperativeplanning in the single-stage resection of breast can-cer [48,49]. By using breast MR imaging as a com-plementary test to the conventional imagingtechniques, more precise information can be ob-tained about the extent of breast cancer, ultimatelyimproving patient care.

Patient selection for preoperative breast MR im-aging may include the young patient, the patientwho has dense or moderately dense breasts, andthe patient who has difficult tumor histology,

such as infiltrating lobular carcinoma and tumorswith extensive intraductal component (EIC) inwhich tumor size assessment is difficult [50]. Infil-trating lobular carcinoma is known to be difficultto detect on mammography; for this neoplasm,MR imaging has been shown to assess the extentof disease more accurately than mammography[51,52]. MR imaging has also been shown to dem-onstrate unsuspected DCIS, which can be helpfulwhen assessing extent of disease in preoperativetesting [53,54]. EIC is associated with a known in-vasive carcinoma when greater than 25% of the


tumor is DCIS. EIC can also be associated with re-sidual carcinoma and positive margins after lump-ectomy, and there is some evidence that thepresence of EIC may indicate an increased risk oflocal recurrence.

MR imaging defines the anatomic extent of dis-ease more accurately than mammography, particu-larly in tumors with difficult histologies, asdiscussed previously. Breast MR imaging can givehelpful information for staging on tumor size, pres-ence or absence of multifocal or multicentric dis-ease, and whether the chest wall or pectoralismuscle is invaded [55]. Chest wall involvement isan important consideration for the surgeon beforesurgical planning. Mammography does not imagethe ribs, intercostal muscles, and serratus anteriormuscle that compose the chest wall. Tumor involve-ment of the chest wall changes the patient’s stage toIIIB, indicating that the patient may benefit fromneoadjuvant chemotherapy before surgery. Tumorinvolvement of the pectoralis muscle does not alterstaging, and surgery can usually proceed. Knowl-edge that the muscle is involved, however, may alterthe surgeon’s plan; for example, when the full thick-ness of the pectoralis major muscle is involved withtumor, the surgeon may be more inclined to per-form a radical instead of a modified radical mastec-tomy (Fig. 2).

Controversy exists regarding the use of MR imag-ing to stage breast cancer. MR imaging may identifycancer, especially additional DCIS that is currentlytreated with adjuvant chemotherapy and radiationtherapy. It is being argued that staging with MR im-aging results in surgical overtreatment of the pa-tient’s breast cancer. For example, many women

who may be candidates for breast conservationtherapy may be overtreated with mastectomy basedon the MR imaging results that additional diseasewas found elsewhere. The challenge is in knowingwhat is and what is not clinically significant diseaseon MR imaging. At this time, identification of sig-nificant disease that will not be treated with radia-tion therapy is not possible and all additionaldisease is treated surgically. In a recent study by Lib-erman [56], 666 nonpalpable, mammographicallyoccult MR imaging–detected lesions were reviewed;the frequency of malignancy was found to increasesignificantly with lesion size (P<.001), and only 3%of lesions smaller than 0.5 cm were found to be ma-lignant. Trials that involve radiologists in additionto radiation oncologists and surgeons are neededto answer these perplexing questions.

Sentinel lymph node biopsy

For patients who have invasive breast cancer, lymphnode status is one of the most important prognosticfactors [57]. Traditionally, lymph node status hasbeen assessed with axillary dissection. Advances inbreast cancer screening and increased public aware-ness have meant that many women are now diag-nosed at an earlier stage when the axillary lymphnodes are free of metastasis. For many of these pa-tients, axillary dissection, with its complicationssuch as lymphedema, has no benefit. Sentinellymph node biopsy has emerged as a minimally in-vasive alternative to axillary dissection, where a neg-ative sentinel node obviates the need for the latter.

The sentinel node is the node in the tumor bedthat is most likely to harbor tumor cells because itis the first to receive lymphatic drainage from the

Fig. 2. (A) Posteriorly located breast cancer. Sagittal, fat-suppressed, T1-weighted image demonstrates the chestwall to be free of tumor. The patient went on to have successful surgical excision and conservation treatment forthis invasive ductal cancer. (B) Posteriorly located irregular heterogeneously enhancing breast carcinoma. An en-hancing mass is noted involving the chest wall. Appearances demonstrate extension of the tumor into the in-tercostal muscles, which would make the patient IIIB. Primary management now is chemotherapy, ratherthan surgical. This tumor was not detected on the patient’s mammogram.

Bartella et al54

tumor. A number of studies have shown that thefindings in the sentinel node accurately predictthe status of the other axillary nodes [58,59]. Thetechnique that is used varies between institutions.A radiotracer that is taken up by macrophages andallows visualization of the lymphatics is injectedinto the breast. In the United States, technetium99m sulfur colloid (filtered or unfiltered) is used.In Europe and Australia, other tracers such as tech-netium 99m nanocolloid are widely available. Thesite of injection also varies between institutions.Most investigators favor peritumoral or sub/intra-dermal injections or a combination of these. Lym-phoscintigraphy is performed the afternoonbefore or the morning of surgery, depending onthe size of the colloid particle injected and whetherit is filtered. Images are acquired in two planes, usu-ally anterior and oblique, using a high-resolutioncamera, and the site of the sentinel node shouldbe marked on the patient’s skin, making sure thearm is abducted in the same position as in the op-erating room, allowing the surgeon to focus atten-tion on the correct spot in the axilla (Figs. 3 and 4).

At surgery, localization of the sentinel node isperformed by external counting with a gammaprobe. Intraoperative lymphatic mapping using iso-sulfur dye, injected intradermally, peritumorally,subdermally, or in the periareolar region is also of-ten performed to locate the sentinel node. Usinga combined technique of blue dye and radioisotopemapping, success rates of 97% in identifying thesentinel lymph node have been reported [60].With increasing experience in the radioisotope tech-nique, the blue dye technique only marginally im-proves radio-guided identification of the sentinel

node. After the sentinel nodes have been removed,a thorough histopathologic examination of the no-des, including multisectioning and immunohisto-chemistry analysis, is performed. Completionaxillary node dissection is then performed if thesentinel node is positive.

Overall, the sensitivity of sentinel node biopsy(SNB) for node involvement ranges from 71% to100% and the average false-negative rate is 8.4%[61]. The American Society of Breast Surgeons rec-ommends that a sentinel lymph node identificationsensitivity of 85% with a false-negative rate of 5% orless is required to abandon axillary dissection.

More recently, investigators have been evaluatingother imaging techniques to assess the axilla. PET isa noninvasive imaging modality that can detectlymph nodes. A number of studies have comparedfludeoxyglucose F 18 (FDG)-PET to SNB or axillarylymph node dissection and have found that al-though the sensitivity is relatively low, the specific-ity is greater than 94% [62,63], suggesting thatFDG-PET cannot replace histologic staging in earlybreast cancer but may be able to identify womenwho can forego SNB, requiring axillary node dissec-tion instead. The use of MR imaging in evaluatingthe axilla has also been assessed [64]. Presently,there is no imaging modality able to detect micro-scopic nodal metastasis detected at SNB.

Breast MR imaging

Breast MR imaging has become an important andpowerful tool in breast imaging. The performanceand clinical uses of breast MR imaging are morestandardized and defined than they were several

Fig. 3. Lymphoscintigraphy performed in a 47-year-old woman who subsequently underwent left breast conser-vation therapy and sentinel lymph node biopsy for a 1.3-cm moderately to poorly differentiated invasive ductalcarcinoma. The images demonstrate a chain of draining nodes in the left axilla (arrow in left image). Radiotraceris also demonstrated at the intradermal injection site overlying the primary tumor (arrow in right image). Sevenlymph nodes removed at surgery were negative for metastasis.


decades ago. In the past few years, great strides havebeen made in the realm of breast intervention, withnew coils and needles now available. More se-quences that have an increase in image qualityand speed of acquisition are available from manu-facturers. The use of breast MR imaging for cancerdetection is changing the current algorithms inthe detection and treatment of breast cancer. By be-ing able to detect cancer that is occult on conven-tional imaging such as mammography andsonography, MR imaging can provide valuable in-formation about breast cancer that was, up to thispoint, unimaginable [65,66]. This section addressescurrent and evolving trends in breast MR imagingfor cancer detection. Preoperative staging andscreening are addressed in respective sections inthis article.

Indications for the use of breast MR imaging

Many studies have shown that breast MR imaging isbest used in situations in which there is a knowncancer, a suspected cancer, or a high probability offinding cancer. For example, in the preoperativeevaluation of the patient with a known cancer, theability of MR imaging to detect multifocal (withinthe same quadrant of the breast) and multicentric(within different quadrants) disease that was previ-ously unsuspected facilitates accurate staging [45–48,50,67,68]. Incidental synchronous contralateralcarcinomas have also been detected when screening

Fig. 4. Lymphoscintigraphy performed in a 40-year-old woman, pre left mastectomy for recurrent ductalcarcinoma. She previously had left breast conserva-tion therapy and sentinel lymph node biopsy fora node-negative invasive ductal carcinoma. The im-ages demonstrate radiotracer activity at the intrader-mal injection site overlying the primary tumor (whitearrow) and tracer activity in draining nodes in the leftaxilla, in the left intramammary lymph nodes, and ina contralateral draining right axillary node (black ar-rows). The left axillary sentinel node was removed atsurgery and yielded metastatic disease.

the contralateral breast in patients who have knowncancer [46,49,69]. In the patient who has positivemargins following an initial attempt at breast con-servation, MR imaging can detect residual disease[69–71] and, in the patient who has inoperable lo-cally advanced breast cancer, MR imaging can assessresponse to neoadjuvant chemotherapy [72–75].Suspected recurrence can be confirmed with MRimaging in the previously treated breast [76,77],and breast MR imaging is absolutely indicated inthe patient who has axillary node metastases withunknown primary [78–82]. Another indicationthat is promising is the use of MR imaging forhigh-risk screening [83–87].

During the past few decades, as breast MR imag-ing has been incorporated into the clinical evalua-tion of the breast, it has become apparent thatstandardization of image acquisition and terminol-ogy is important. The American College of Radiol-ogy committee on standards and guidelines haspublished a document for the performance ofbreast MR imaging [88], and the recent publicationof the BI-RADS lexicon [89] includes a sectionabout breast MR imaging so that standardized ter-minology can be used when describing findingson breast MR imaging. The existence of standard-ized guidelines in image acquisition and interpreta-tion may help disseminate this technology fromacademic centers to the community.

Neoadjuvant chemotherapy response

Response to neoadjuvant chemotherapy for locallyadvanced breast cancer can be assessed with MR im-aging. A complete pathologic response (elimina-tion of tumor) following neoadjuvant therapy isstrongly predictive of excellent long-term survival.Minimal response suggests a poor long-term sur-vival regardless of postoperative therapy. MR imag-ing may find a role in being able to predict at anearlier time, perhaps after several cycles of chemo-therapy, which patients are responding to neoadju-vant chemotherapy. Early knowledge of suboptimalresponse may allow switching to alternative treat-ment regimens earlier rather than later. Unless theresponse is dramatic, it currently takes longer topredict response because the mammogram andphysical examination may be compromised dueto fibrosis. Studies have shown MR imaging to besuperior to mammography, sonography, and clini-cal examination in measuring residual disease afterneoadjuvant chemotherapy [90–92]. MR imagingshowed 89% to 97% correlation with pathology,whereas clinical examination demonstrated 55%correlation and mammography demonstrated52% correlation. There are limitations with MR im-aging in assessing early treatment response. A recentstudy has shown that although early changes of

Bartella et al56

tumor measurements were observed with MR imag-ing after one cycle of chemotherapy (2 weeks),these changes were not statistically significant. Ithas been shown, however, that the final change inMR imaging volume was a significant predictor ofrecurrence-free survival. A multicenter trial (Ameri-can College of Radiology Imaging Network Proto-col 6657) is under way to further assess thepredictive value of early changes in tumor volumeand changes in tumor vascularity, as measured byMR imaging [93].

Assessment of residual disease

For patients who have undergone lumpectomy withpositive margins, MR imaging can be helpful in theassessment of residual tumor load. Postoperativemammography is able to detect residual calcifica-tions, although it is limited in the evaluation of re-sidual uncalcified DCIS or residual mass. MRimaging is able to detect bulky residual disease atthe lumpectomy site and residual disease in thesame quadrant (multifocal) or a different quadrant(multicentric). MR imaging can be helpful in thedetermination of whether the patient would bestbe served with directed re-excision (residual diseaseat the lumpectomy site or multifocal disease) orwhether the patient warrants mastectomy (multi-centric disease) (Fig. 5). Evaluation for microscopicresidual disease directly at the lumpectomy site isnot the role of MR imaging because the surgeon

Fig. 5. A 33-year-old woman presented with positive-margins status post lumpectomy for a palpable massthat was mammographically occult. The MR imagedemonstrated a large residual mass surrounding thelumpectomy bed, in addition to enhancing skin nod-ules and thickening and enlarged axillary nodes; thepatient had advanced local disease and was thereforetreated with chemotherapy.

will perform re-excision based on pathologic mar-gins and not based on MR imaging results. If multi-centric disease is identified on MR imaging beforemastectomy, then it is important to sample and ver-ify this impression. A study [70] has shown that MRimaging was able to detect residual disease in 23 of33 cases (70%) and alone identified multifocal ormulticentric disease in 9 of 33 cases (27%). The au-thors looked at 100 women who underwent MR im-aging for positive margins [94]. Fifty-eight patientshad residual disease at surgery: 20 multicentric, 15multifocal, and 23 unifocal. MR identified 18 of20 cases (90%) of multicentric, 14 of 15 cases(93%) of multifocal, and 18 of 23 cases (78%) ofunifocal residual disease. Eight false-negative find-ings included 2 cases of multicentric DCIS occultto MR imaging and 6 cases of residual disease (4subcentimeter invasive and 2 microscopic DCIS)directly at the lumpectomy site. Overall sensitivityfor detection of residual cancer was 86% (50/58)and specificity was 68% (28/41).

Tumor recurrence at the lumpectomy site

Tumor recurrence after breast conservation occursat an overall rate of 1% to 2% per year. Recurrencedirectly at the lumpectomy site occurs earlier thanelsewhere in the breast and usually peaks severalyears following conservation therapy. Evaluationof the lumpectomy site by mammography is ex-tremely limited due to postoperative scarring, andphysical examination may have greater sensitivitythan mammography in the detection of recurrence.Mammography is able to detect 25% to 45% ofrecurrences and is more likely to detect recurrenttumors associated with calcifications than recur-rences without calcifications. In one study [76], allrecurrences of invasive carcinoma (enhanced withnodular enhancement and linear enhancement)were observed in the cases of DCIS recurrence.Most scars showed no enhancement.

Knowing when to image for potential recurrenceis problematic because scar tissue can enhance foryears following surgery. Because recurrence peaksin the first few years following surgery and themost likely site of recurrence is the lumpectomysite, the usefulness of the information obtainedfrom a costly MR imaging study needs to beweighed against the benefit obtained from a poten-tially less expensive needle biopsy of the area.

Occult primary breast cancer

Patients presenting with axillary metastases suspi-cious for breast primary and a negative physical ex-amination and negative mammogram shouldundergo breast MR imaging [81,82]. In patients


who have this rare clinical presentation, MR imag-ing has been able to detect cancer in 90% to100% of cases when a tumor is present. The tumorsare generally small (under 2 cm); thus, they mayevade detection by conventional imaging and phys-ical examination.

The identification of the site of malignancy is im-portant therapeutically. Patients traditionally un-dergo mastectomy because the site of malignancyis unknown. Whole-breast radiation can be given,although it is generally not recommended becausesurvival is equal but the recurrence rate is higher(up to 23%). Thus, when a site of malignancy canbe identified, the patient can be spared mastectomyand offered breast conservation therapy, therebyhaving a significant impact on patient management(Fig. 6). In one study, the results of the MR exami-nation changed therapy in approximately one halfof cases, usually allowing conservation in lieu ofmastectomy [78]. If a site of malignancy is not

identified on MR imaging, it may be reasonablefor the patient to receive full breast radiation withcareful follow-up with MR imaging examinationrather than mastectomy.

Proton MR spectroscopy of the breast

Proton MR spectroscopy (1H MRS) provides bio-chemical information about the tissue under inves-tigation. The diagnostic value of 1H MRS in canceris typically based on the detection of elevated levelsof choline compounds, which is a marker of activetumor [95]; these compounds include choline,phosphocholine, glycerophosphocholine, and tau-rine. At the field strengths used for in vivo work(1.5–4 T), these multiple resonances cannot bespectrally resolved and thus appear as a singlepeak, termed ‘‘total choline-containing com-pounds’’ (tCho). For the brain and prostate, 1H

Fig. 6. (A, B) A 60-year-old woman status post right mastectomy and left reduction mammoplasty presented witha palpable mass in the left breast. Postreduction changes were seen on the mammogram. The patient refusedbiopsy for the palpable mass, so MR imaging was performed. (C) T1-weighted, fat-suppressed sagittal imagethrough the left breast demonstrated a peripherally enhancing lesion at the site of the palpable abnormality.(D) The non–fat-suppressed image confirmed the presence of fat in the lesion, consistent with fat necrosis.

Bartella et al58

MRS is approved by the US Food and Drug Admin-istration (FDA) and is widely used.

The use of proton MR spectroscopyin the breast

1H MRS has been suggested as an adjunct to breastMR imaging [96]. Studies performed on 1.5-T MRscanners have reported sensitivities of 70% to100% and specificities of 67% to 100% for breast1H MRS. Fewer studies performed on 4-T scannershave reported sensitivities of 46% to 61% and spec-ificities of 83% to 94% [97,98]. The use of 1H MRSin the breast has predominantly been performedusing a single-voxel technique (SVS), but the useof MR spectroscopy has also been tried [99]. Onedisadvantage is that the voxel size has been a limit-ing factor in SVS. It has been shown that voxels lessthan 1 cm3 yield large errors in tCho measure-ments, even at 4 T [98]. MR spectroscopy providesinformation about the spatial distribution ofmetabolites, which is useful for studying multiplelesions. This method is technically more challeng-ing than SVS, encountering more problems in thequantification of metabolite levels. Qualitativeand quantitative approaches have been used incholine measurements. Scanning at 4 T allowsdetection of tCho in benign lesions and normalsubjects, therefore magnifying the need for a quanti-tative method for choline detection. Quantificationhas been performed in studies at 1.5 T [100] and at4 T [98], with similar absolute choline levels (0.4–5.8 mmol/L and 0.4–10 mmol/L, respectively) be-ing reported. Higher signal-to-noise ratio has beenreported on previous MR spectroscopy studies at4 T, enabling the detection of tCho in smallerlesions and the occasional detection of other me-tabolites such as taurine, creatine, and glycine [98].

Differentiating benign from malignantbreast lesions

Multiple in vivo 1H MRS studies aimed at improv-ing the discrimination between benign and malig-nant breast lesions have been done at severalcenters (Table 5) [98–107]. In a group of patientsstudied at the Memorial Sloan-Kettering CancerCenter (MSKCC) [96], the sensitivity of 1H MRSwas 100% and the specificity was 88%, which com-pared favorably to prior reports of breast 1H MRS.The use of 1H MRS as an adjunct to breast MR im-aging would have significantly (P<.01) increasedthe positive predictive value of biopsy from 35%to 82%. If 1H MRS had been used as an adjunctto MR imaging in the 40 lesions with unknown his-tology, then biopsy may have been spared in 58%of the lesions and none of the cancers would havegone undetected. These data suggest that 1H MRSmay be a useful supplement to breast MR imaging,thereby reducing the number of benign biopsy sam-ples without compromising the diagnosis of breastcancer (Fig. 7).

All cancers in the study reported by Bartella andcolleagues [96] were identified by 1H MRS(Fig. 8); there were no false-negative cases. A cho-line peak was identified at 1H MRS in a varietyof cancer histologies, including 16 invasive can-cers (infiltrating ductal, infiltrating lobular, andinfiltrating mixed ductal and lobular) and 1 DCIS.Three false-positive cases are included in thisreport: a fibroadenoma, a chronic inflammatory le-sion with atypia (Fig. 9), and ADH with columnarcell alteration. A false-positive choline peak has pre-viously been reported with a fibroadenoma[101,102]. To the authors’ knowledge, which is lim-ited by the small number of published series onbreast single-voxel 1H MRS, the other two lesions

Table 5: In vivo 1H MRS studies performed on 1.5-T magnets in order to differentiate benign frommalignant lesions

StudyMalignantlesions

Benignlesions

Sens(%)

Spec(%) TP TN FN FP

PPV(%)

Roebuck et al, 1998 [100] 10 7 70 86 7 6 3 1 88Kvistad et al, 1999 [102] 11 11 82 82 9 9 2 2 82Cecil et al, 2001 [103] 23 15 83 87 19 13 4 2 90Yeung et al, 2001 [101] 24 6 92 83 22 5 2 1 97Jagannathan et al, 2001 [110] 32 14 81 86 26 12 6 2 93Tse et al, 2003 [106] 19 21 89 100 17 21 2 0 100Huang et al, 2004 [107] 18 12 100 67 18 8 0 4 82Bartella et al, 2006 [96] 31 26 100 88 31 23 0 3 91Sum 168 112 87 85 149 97 19 15 90

Abbreviations: FN, false-negative cases; FP, false-positive cases; PPV, positive predictive value; Sens, sensitivity; Spec, spec-ificity; TN, true-negative cases; TP, true-positive cases.Adapted from Bartella L, Morris EA, Dershaw DD, et al. Proton MR spectroscopy with choline peak as malignancy markerimproves positive predictive value for breast cancer diagnosis: preliminary study. Radiology 2006;239(3):691; withpermission.


have not previously been reported. In view of thepresence of atypia in these two lesions, excisionwould have been the standard of care. Furtherwork is necessary to evaluate the frequency andcharacteristics of false-positive findings on 1H MRS.

Characterization of histopathologic subtypes

MRS has been shown to be highly sensitive for allinvasive breast cancers, regardless of histology

Fig. 7. A 52-year-old woman with a mammographi-cally detected, biopsy-proven invasive ductal carci-noma in the left breast. (A) Sagittal, T1-weighted,MR image of left breast immediately after injectionof intravenous gadolinium shows a 1.5-cm rim-en-hancing mass. (B) MR spectroscopy of this lesion dem-onstrated a positive choline peak at a frequency of3.2 ppm, with a signal-to-noise ratio greater than 2;this is a true positive finding. Cho, choline; Lac, lac-tate; Lip, lipid. (From Bartella L, Morris EA, DershawDD, et al. Proton MR spectroscopy with choline peakas malignancy marker improves positive predictivevalue for breast cancer diagnosis: preliminary study.Radiology 2006;239(3):689; with permission.)

[98,108]. Based on preliminary results from an on-going study at MSKCC, 1H MRS at 1.5 T was per-formed on 30 invasive breast cancers [108].Significantly, a choline peak was detected in onecase of DCIS [96]. This lesion is of interest, in lightof prior reports, and suggests that DCIS may notmanifest a choline peak [100,104]. Further studywith additional DCIS lesions is essential. At 4 T,the tCho levels did not appear to be related to dif-ferent histologic subtypes [98].

Fig. 8. A 43-year-old woman presented with a newpalpable mass in the right breast. MR imaging–guidedbiopsy followed by surgical excision yielded benignfibrosis and ductal hyperplasia. (A) Postcontrast sagit-tal, T1-weighted MR image of the right breast demon-strates an irregular 4.2-cm mass (arrow). (B) Spectrumobtained from MR spectroscopy did not demonstratea positive choline resonance peak, and there wasonly noise level at a frequency of 3.2 ppm; this isa true negative finding. Cho, choline; Lac, lactate;Lip, lipid. (From Bartella L, Morris EA, Dershaw DD,et al. Proton MR spectroscopy with choline peak asmalignancy marker improves positive predictive valuefor breast cancer diagnosis: preliminary study. Radiol-ogy 2006;239(3):690; with permission.)

Bartella et al60

Fig. 9. A 51-year-old woman who had a positive fam-ily history for breast cancer presented with a suspi-cious lesion detected on a high-risk screening breastMR imaging examination. (A) Postcontrast sagittal,T1-weighted MR image of the left breast shows duc-tal clumped enhancement in the retroareolar region.(B) A magnified spectrum of this lesion demonstrateda positive choline resonance peak, with a signal tonoise ratio greater than 2. Excision of this lesionyielded an atypical chronic inflammatory lesion; thisis a false-positive finding. Cho, choline; Lac, lactate;Lip, lipid. (From Bartella L, Morris EA, Dershaw DD,et al. Proton MR spectroscopy with choline peak asmalignancy marker improves positive predictive valuefor breast cancer diagnosis: preliminary study. Radiol-ogy 2006;239(3):689; with permission.)

Evaluation of normal and lactatingbreast parenchyma

In a small series performed at MSKCC, normalbreast glandular parenchyma had no choline signaldetected at 1.5 T, regardless of menstrual status orstage in the menstrual cycle [109]. Yet, choline sig-nal can be detected in normal breast tissue at higherfield strengths [98]. Choline signal has also beendocumented in the lactating breast [102,110]. De-spite the limited existing data, this detection sug-gests the increasing need for the quantification ofcholine concentrations [97,98]. A recent study per-formed at 7 T on normal volunteers detected tChoand taurine in normal glandular parenchyma [111].

Predicting response to neoadjuvantchemotherapy

An early pilot study has shown that using a 1.5-Tmagnet, a change in tCho was observed after thecompletion of treatment, which was confirmedwith pathology [110]. In a more recent pilot studyperformed on a 4-T system, 1H MRS was able topredict clinical response in patients who had locallyadvanced breast cancer within 24 hours of receivingthe first dose of neoadjuvant chemotherapy. Theseresults suggest that the addition of 1H MRS may of-fer a substantial advantage over MR imaging alonein the prediction of response to neoadjuvant che-motherapy [112].

MR spectroscopy of axillary lymphnodes in breast cancer patients

Yeung and colleagues [113] reported the role ofin vivo MR spectroscopy in the evaluation of axillarynodes using choline as a marker for metastases.This study was performed on a 1.5-T magnet usinga 14-cm circular receive-only surface coil that wasplaced against the affected breast to improve the sig-nal-to-noise ratio. They detected axillary nodal me-tastases with a sensitivity of 82% and a specificity of100%. In vitro MR spectroscopy was used to charac-terize the metabolic profile of axillary nodes anddistinguish metastatic lymph nodes from nonin-volved ones [114,115]. Over 40 metabolites wereidentified. The specificity of in vitro MR spectros-copy in detecting axillary node metastases usingthe glycerophosphocholine-phosphocholine/thre-onine ratio was 80%, and the specificity was 91%.

High-resolution magic angle spinning MRspectroscopy of breast tissue

Studies have been performed looking at ex vivo MRspectroscopic imaging of breast tissue to establishthe metabolic profile of breast tumor and nonin-volved breast tissue [116]. Metabolite composition


has been investigated in perchloric acid extractsfrom tissue and in intact tissue using high-resolu-tion magic angle spinning MR spectroscopy usingnuclear MR spectrometers (DRX-400/DRX-600BRUKER, Rheinsteiten, Germany). More than 30different metabolites were detected and assigned[116]. The total amount of choline on average wasfound to be more than 10 times higher in tumor tis-sue than in noninvolved tissue [116]. The relativeintensities of the resonances of these compoundswere also found to be different in cancerous andin noninvolved tissue.

Positron emission tomographyand breast cancer

PET is different than other imaging modalities inthat it gives metabolic information in addition toanatomic detail. FDG-PET has outperformed con-ventional imaging modalities in determining extentof disease and in evaluating disease recurrence(Fig. 10) [117,118]. Currently, this modality isused for treatment monitoring and is approved bythe FDA as an adjunct to conventional imaging mo-dalities. Studies have also suggested that FDG-PET

Fig. 10. FDG PET/CT performed on a 50 year old woman with newly diagnosed invasive ductal carcinoma of theright breast with palpable right axillary nodes. Coronal PET (A), axial PET (B), axial CT (C), and fusion PET/CT (D)demonstrate hypermetabolic activity in the primary tumor (thick arrow, SUV 12.2), within right axillary nodesand right intramammary lymph nodes (thin arrow, SUV 4.2). Sagittal PET (E), axial PET (F), axial CT (G), and fusionaxial PET/CT (H) demonstrate hypermetabolic activity in L1 vertebral body (thick arrow, SUV 5). A biopsy of theL1 lesion confirmed mestatic disease. Of note, a bone scan performed at the same time demonstrated no abnor-mality. The patient underwent chemotherapy and bisphosphonates therapy and a PET/CT performed 3 monthslater showed a marked decrease in metabolic activity in all the lesions.

Bartella et al62

Fig. 10 (continued)

enables early assessment of treatment response inpatients undergoing chemotherapy [119,120]. In-vestigators have also looked at the role of PET inthe evaluation of suspicious breast lesions [121].It seems that presently, PET is inadequate for the de-tection of small primary tumors; however, the highpositive predictive value may be helpful in cases inwhich conventional imaging is inconclusive.

Dedicated PET mammography units have beendeveloped with a higher spatial resolution to try todetect smaller breast cancers [122,123]. New tracersare also being investigated for use in breast cancer.The use of antibodies as potential tracers for specificcell receptors has also been increasing. For example,68-Ga-Dota-F(ab’)2-Herceptin allows sequentialPET imaging of HER 2 expression and may be usefulin predicting early tumor response [124].

Other new techniques

Other new imaging modalities being assessedfor their ability to detect breast cancer includescintimammography, tomosynthesis, and contrast-enhanced mammography. Previously, scintimam-mography was performed with a general-purposegamma camera. Recently, however, investigatorshave begun to use high-resolution, small field-of-view cameras specific to breast imaging, with en-couraging results [125,126]. Tomosynthesis of thebreast involves the mathematic processing of a setof planar images of the breast acquired at differentangles. It is hoped that this method of evaluatingthe breast will reduce the number of false-positiveand false-negative mammograms due to overlap-ping tissue.


The use of contrast agents has also been used inMR imaging to improve the sensitivity of mammog-raphy. Two small studies using nonionic iodinatedcontrast with temporal subtraction [127] and dualenergy subtraction [128] have performed well,demonstrating cancers with few false positives.

Summary

Advances in breast imaging have revolutionizedbreast cancer management. Screening mammogra-phy remains the ‘‘gold standard’’ for breast cancerdetection. Modalities such as digital mammogra-phy, breast ultrasound, and breast MR imagingplay increasingly important roles in breast imaging.Percutaneous needle biopsy provides a faster, lessinvasive, and less expensive method than surgeryfor breast diagnosis. SNB provides a less invasivemethod than axillary dissection for staging the ax-illa. MR spectroscopy may improve specificity ofbreast cancer diagnosis. New modalities such asPET, breast tomosynthesis, and contrast-enhancedmammography may also contribute to the breastimaging evaluation. Further work is necessary to re-fine and evaluate these newer modalities so thatwomen can be offered the most accurate and leastinvasive techniques for breast cancer diagnosis,staging, monitoring, and treatment.

References

[1] Edwards BK, Brown ML, Wingo PA, et al. An-nual report to the nation on the status of cancer,1975–2002, featuring population-based trendsin cancer treatment. J Natl Cancer Inst 2005;97(19):1407–27.

[2] Page DL, Dupont WD, Rogers LW, et al. Intra-ductal carcinoma of the breast: follow-up afterbiopsy only. Cancer 1982;49(4):751–8.

[3] Rosen PP, Kosloff C, Lieberman PH, et al. Lobu-lar carcinoma in situ of the breast. Detailed anal-ysis of 99 patients with average follow-up of24 years. Am J Surg Pathol 1978;2(3):225–51.

[4] Yen MF, Tabar L, Vitak B, et al. Quantifying thepotential problem of overdiagnosis of ductalcarcinoma in situ in breast cancer screening.Eur J Cancer 2003;39(12):1746–54.

[5] Rosenberg RD, Hunt WC, Williamson MR, et al.Effects of age, breast density, ethnicity, andestrogen replacement therapy on screeningmammographic sensitivity and cancer stage atdiagnosis: review of 183,134 screening mam-mograms in Albuquerque, New Mexico. Radiol-ogy 1998;209(2):511–8.

[6] Barlow WE, Lehman CD, Zheng Y, et al. Perfor-mance of diagnostic mammography for womenwith signs or symptoms of breast cancer. J NatlCancer Inst 2002;94(15):1151–9.

[7] Eltahir A, Jibril JA, Squair J, et al. The accuracyof ‘‘one-stop’’ diagnosis for 1,110 patients pre-senting to a symptomatic breast clinic. J RColl Surg Edinb 1999;44(4):226–30.

[8] Dee KE, Sickles EA. Medical audit of diagnosticmammography examinations: comparison withscreening outcomes obtained concurrently. AJRAm J Roentgenol 2001;176(3):729–33.

[9] Gennaro G, di Maggio C. Dose comparisonbetween screen/film and full-field digital mam-mography. Eur Radiol 2006;16(11):2559–66.

[10] Pisano ED, Gatsonis C, Hendrick E, et al. Diag-nostic performance of digital versus film mam-mography for breast-cancer screening. N Engl JMed 2005;353(17):1773–83.

[11] ACR practice guideline for the performance ofa breast ultrasound examination. In: ACR prac-tice guidelines and technical standards: ultra-sound research (vol. 10). Reston (VA): ACR;2002. p. 763–5.

[12] Stavros AT, Thickman D, Rapp CL, et al. Solidbreast nodules: use of sonography to distin-guish between benign and malignant lesions.Radiology 1995;196(1):123–34.

[13] Zonderland HM, Coerkamp EG, Hermans J,et al. Diagnosis of breast cancer: contributionof US as an adjunct to mammography. Radiol-ogy 1999;213(2):413–22.

[14] Bassett LW, Kimme-Smith C, Sutherland LK,et al. Automated and hand-held breast US: ef-fect on patient management. Radiology 1987;165(1):103–8.

[15] Deurloo EE, Tanis PJ, Gilhuijs KG, et al. Reduc-tion in the number of sentinel lymph node pro-cedures by preoperative ultrasonography of theaxilla in breast cancer. Eur J Cancer 2003;39(8):1068–73.

[16] Heywang-Kobrunner SHD, Dershaw D, SchreerI. Diagnostic breast imaging. 2nd edition. NewYork: Thieme; 2001.

[17] Duffy SW, Tabar L, Chen HH, et al. The impactof organized mammography service screeningon breast carcinoma mortality in seven Swedishcounties. Cancer 2002;95(3):458–69.

[18] Paci E, Duffy SW, Giorgi D, et al. Quantificationof the effect of mammographic screening onfatal breast cancers: the Florence Programme1990–96. Br J Cancer 2002;87(1):65–9.

[19] Smith RA, Saslow D, Sawyer KA, et al. AmericanCancer Society guidelines for breast cancerscreening: update 2003. CA Cancer J Clin2003;53(3):141–69.

[20] Dershaw DD. Mammographic screening of thehigh-risk woman. Am J Surg 2000;180(4):288.

[21] Bazzocchi M, Zuiani C, Panizza P, et al. Con-trast-enhanced breast MRI in patients with sus-picious microcalcifications on mammography:results of a multicenter trial. AJR Am J Roent-genol 2006;186(6):1723–32.

[22] Kriege M, Brekelmans CT, Boetes C, et al. Differ-ences between first and subsequent rounds ofthe MRISC breast cancer screening program

Bartella et al64

for women with a familial or genetic predispo-sition. Cancer 2006;106(11):2318–26.

[23] Liberman L. Centennial dissertation. Percutane-ous imaging-guided core breast biopsy: state ofthe art at the millennium. AJR Am J Roentgenol2000;174(5):1191–9.

[24] Liberman L, Goodstine SL, Dershaw DD, et al.One operation after percutaneous diagnosis ofnonpalpable breast cancer: frequency and asso-ciated factors. AJR Am J Roentgenol 2002;178(3):673–9.

[25] Liberman L, Feng TL, Dershaw DD, et al. US-guided core breast biopsy: use and cost-effec-tiveness. Radiology 1998;208(3):717–23.

[26] Liberman L, Sama MP. Cost-effectiveness of ste-reotactic 11-gauge directional vacuum-assistedbreast biopsy. AJR Am J Roentgenol 2000;175(1):53–8.

[27] Liberman L. Percutaneous image-guided corebreast biopsy. Radiol Clin North Am 2002;40(3):483–500 [vi].

[28] Jackman RJ, Rodriguez-Soto J. Breast microcal-cifications: retrieval failure at prone stereotacticcore and vacuum breast biopsy—frequency,causes, and outcome. Radiology 2006;239(1):61–70.

[29] Jackman RJ, Birdwell RL, Ikeda DM. Atypicalductal hyperplasia: can some lesions be definedas probably benign after stereotactic 11-gaugevacuum-assisted biopsy, eliminating the recom-mendation for surgical excision? Radiology2002;224(2):548–54.

[30] Lee CH, Carter D, Philpotts LE, et al. Ductal car-cinoma in situ diagnosed with stereotactic coreneedle biopsy: can invasion be predicted? Radi-ology 2000;217(2):466–70.

[31] Liberman L, Benton CL, Dershaw DD, et al.Learning curve for stereotactic breast biopsy:how many cases are enough? AJR Am J Roent-genol 2001;176(3):721–7.

[32] Pfarl G, Helbich TH, Riedl CC, et al. Stereotactic11-gauge vacuum-assisted breast biopsy: a vali-dation study. AJR Am J Roentgenol 2002;179(6):1503–7.

[33] Parker SH, Klaus AJ, McWey PJ, et al. Sono-graphically guided directional vacuum-assistedbreast biopsy using a handheld device. AJRAm J Roentgenol 2001;177(2):405–8.

[34] Philpotts LE, Hooley RJ, Lee CH. Comparisonof automated versus vacuum-assisted biopsymethods for sonographically guided core bi-opsy of the breast. AJR Am J Roentgenol 2003;180(2):347–51.

[35] Soo MS, Baker JA, Rosen EL. Sonographic detec-tion and sonographically guided biopsy ofbreast microcalcifications. AJR Am J Roentgenol2003;180(4):941–8.

[36] Berg WA. Rationale for a trial of screening breastultrasound: American College of Radiology Im-aging Network (ACRIN) 6666. AJR Am J Roent-genol 2003;180(5):1225–8.

[37] LaTrenta LR, Menell JH, Morris EA, et al. Breastlesions detected with MR imaging: utility andhistopathologic importance of identificationwith US. Radiology 2003;227(3):856–61.

[38] Liberman L. Percutaneous magnetic resonanceimaging guided biopsy. In: Morris EA,Liberman L, editors. Breast MRI: diagnosisand intervention. New York: Springer, Inc.;2005. p. 297–315.

[39] Perlet C, Heywang-Kobrunner SH, Heinig A,et al. Magnetic resonance-guided, vacuum-assis-ted breast biopsy: results from a European mul-ticenter study of 538 lesions. Cancer 2006;106(5):982–90.

[40] Lehman CD, DePeri ER, Peacock S, et al. Clini-cal experience with MRI-guided vacuum-assis-ted breast biopsy. AJR Am J Roentgenol 2005;184(6):1782–7.

[41] Liberman L, Bracero N, Morris E, et al. MRI-guided 9-gauge vacuum-assisted breast biopsy:initial clinical experience. AJR Am J Roentgenol2005;185(1):183–93.

[42] Liberman L, Morris EA, Dershaw DD, et al. FastMRI-guided vacuum-assisted breast biopsy: ini-tial experience. AJR Am J Roentgenol 2003;181(5):1283–93.

[43] Orel SG, Rosen M, Mies C, et al. MR imaging-guided 9-gauge vacuum-assisted core-needlebreast biopsy: initial experience. Radiology2006;238(1):54–61.

[44] Holland R, Veling SH, Mravunac M, et al. His-tologic multifocality of Tis, T1–2 breast carci-nomas. Implications for clinical trials ofbreast-conserving surgery. Cancer 1985;56(5):979–90.

[45] Orel SG, Schnall MD, Powell CM, et al. Stagingof suspected breast cancer: effect of MR imagingand MR-guided biopsy. Radiology 1995;196(1):115–22.

[46] Fischer U, Kopka L, Grabbe E. Breast carci-noma: effect of preoperative contrast-enhancedMR imaging on the therapeutic approach. Radi-ology 1999;213(3):881–8.

[47] Bedrosian I, Mick R, Orel SG, et al. Changes inthe surgical management of patients with breastcarcinoma based on preoperative magnetic res-onance imaging. Cancer 2003;98(3):468–73.

[48] Liberman L, Morris EA, Dershaw DD, et al. MRimaging of the ipsilateral breast in women withpercutaneously proven breast cancer. AJR Am JRoentgenol 2003;180(4):901–10.

[49] Lee SG, Orel SG, Woo IJ, et al. MR imagingscreening of the contralateral breast in patientswith newly diagnosed breast cancer: prelimi-nary results. Radiology 2003;226(3):773–8.

[50] Berg WA. Imaging the local extent of disease.Sem Breast Dis 2001;4:153–73.

[51] Quan ML, Sclafani L, Heerdt AS, et al. Magneticresonance imaging detects unsuspected diseasein patients with invasive lobular cancer. AnnSurg Oncol 2003;10(9):1048–53.


[52] Weinstein SP, Orel SG, Heller R, et al. MR imag-ing of the breast in patients with invasive lobu-lar carcinoma. AJR Am J Roentgenol 2001;176(2):399–406.

[53] Hwang ES, Kinkel K, Esserman LJ, et al. Mag-netic resonance imaging in patients diagnosedwith ductal carcinoma-in-situ: value in the diag-nosis of residual disease, occult invasion, andmulticentricity. Ann Surg Oncol 2003;10(4):381–8.

[54] Viehweg P, Lampe D, Buchmann J, et al. In situand minimally invasive breast cancer: morpho-logic and kinetic features on contrast-enhancedMR imaging. MAGMA 2000;11(3):129–37.

[55] Morris EA, Schwartz LH, Drotman MB, et al.Evaluation of pectoralis major muscle in pa-tients with posterior breast tumors on breastMR images: early experience. Radiology 2000;214(1):67–72.

[56] Liberman L. Does size matter? Positive predic-tive value of MRI-detected breast lesions asa function of lesion size. AJR Am J Roentgenol2006;182(2):426–30.

[57] Morrow M. Role of axillary dissection in breastcancer management. Ann Surg Oncol 1996;3(3):233–4.

[58] Krag D, Weaver D, Ashikaga T, et al. The senti-nel node in breast cancer—a multicenter vali-dation study. N Engl J Med 1998;339(14):941–6.

[59] Veronesi U, Paganelli G, Viale G, et al. A ran-domized comparison of sentinel-node biopsywith routine axillary dissection in breast cancer.N Engl J Med 2003;349(6):546–53.

[60] Derossis AM, Fey J, Yeung H, et al. A trend anal-ysis of the relative value of blue dye and isotopelocalization in 2,000 consecutive cases of senti-nel node biopsy for breast cancer. J Am CollSurg 2001;193(5):473–8.

[61] Lyman GH, Giuliano AE, Somerfield MR, et al.American Society of Clinical Oncology guide-line recommendations for sentinel lymphnode biopsy in early-stage breast cancer. J ClinOncol 2005;23(30):7703–20.

[62] Guller U, Nitzsche EU, Schirp U, et al. Selectiveaxillary surgery in breast cancer patients basedon positron emission tomography with 18F-flu-oro-2-deoxy-D-glucose: not yet!. Breast CancerRes Treat 2002;71(2):171–3.

[63] van der Hoeven JJ, Hoekstra OS, Comans EF,et al. Determinants of diagnostic performanceof [F-18]fluorodeoxyglucose positron emissiontomography for axillary staging in breast cancer.Ann Surg 2002;236(5):619–24.

[64] Stadnik TW, Everaert H, Makkat S, et al. Breastimaging. Preoperative breast cancer staging:comparison of USPIO-enhanced MR imagingand 18F-fluorodeoxyglucose (FDC) positronemission tomography (PET) imaging for axil-lary lymph node staging-initial findings. EurRadiol 2006;16(10):2153–60.

[65] Kaiser WA, Zeitler E. MR imaging of the breast:fast imaging sequences with and without Gd-DTPA. Preliminary observations. Radiology 1989;170(3 Pt 1):681–6.

[66] Heywang SH, Wolf A, Pruss E, et al. MR imagingof the breast with Gd-DTPA: use and limita-tions. Radiology 1989;171(1):95–103.

[67] Mumtaz H, Hall-Craggs MA, Davidson T, et al.Staging of symptomatic primary breast cancerwith MR imaging. AJR Am J Roentgenol 1997;169(2):417–24.

[68] Schelfout K, Van Goethem M, Kersschot E, et al.Contrast-enhanced MR imaging of breast le-sions and effect on treatment. Eur J Surg Oncol2004;30(5):501–7.

[69] Frei KA, Kinkel K, Bonel HM, et al. MR imagingof the breast in patients with positive marginsafter lumpectomy: influence of the time intervalbetween lumpectomy and MR imaging. AJR AmJ Roentgenol 2000;175(6):1577–84.

[70] Orel SG, Reynolds C, Schnall MD, et al. Breastcarcinoma: MR imaging before re-excisional bi-opsy. Radiology 1997;205(2):429–36.

[71] Faverly DR, Hendriks JH, Holland R. Breast car-cinomas of limited extent: frequency, radio-logic-pathologic characteristics, and surgicalmargin requirements. Cancer 2001;91(4):647–59.

[72] Knopp MV. MR mammography with pharmo-cokinetic mapping for monitoring of breastcancer treatment during neoadjuvant therapy.Magn Reson Imaging Clin N Am 1997;2:633–58.

[73] Esserman L, Kaplan E, Partridge S, et al. MRIphenotype is associated with response to doxo-rubicin and cyclophosphamide neoadjuvantchemotherapy in stage III breast cancer. AnnSurg Oncol 2001;8(6):549–59.

[74] Wasser K, Klein SK, Fink C, et al. Evaluation ofneoadjuvant chemotherapeutic response ofbreast cancer using dynamic MRI with high tem-poral resolution. Eur Radiol 2003;13(1):80–7.

[75] Delille JP, Slanetz PJ, Yeh ED, et al. Invasiveductal breast carcinoma response to neoadju-vant chemotherapy: noninvasive monitoringwith functional MR imaging pilot study. Radiol-ogy 2003;228(1):63–9.

[76] Gilles R, Guinebretiere JM, Shapeero LG, et al.Assessment of breast cancer recurrence withcontrast-enhanced subtraction MR imaging:preliminary results in 26 patients. Radiology1993;188(2):473–8.

[77] Dao TH, Rahmouni A, Campana F, et al. Tumorrecurrence versus fibrosis in the irradiatedbreast: differentiation with dynamic gadoli-nium-enhanced MR imaging. Radiology 1993;187(3):751–5.

[78] Olson JA Jr, Morris EA, Van Zee KJ, et al. Mag-netic resonance imaging facilitates breast con-servation for occult breast cancer. Ann SurgOncol 2000;7(6):411–5.

Bartella et al66

[79] Schorn C, Fischer U, Luftner-Nagel S, et al. MRIof the breast in patients with metastatic diseaseof unknown primary. Eur Radiol 1999;9(3):470–3.

[80] Campana F, Fourquet A, Ashby MA, et al. Pre-sentation of axillary lymphadenopathy withoutdetectable breast primary (T0 N1b breast can-cer): experience at Institut Curie. RadiotherOncol 1989;15(4):321–5.

[81] Morris EA, Schwartz LH, Dershaw DD, et al. MRimaging of the breast in patients with occult pri-mary breast carcinoma. Radiology 1997;205(2):437–40.

[82] Orel SG, Weinstein SP, Schnall MD, et al. BreastMR imaging in patients with axillary nodemetastases and unknown primary malignancy.Radiology 1999;212(2):543–9.

[83] Kuhl CK, Schmutzler RK, Leutner CC, et al.Breast MR imaging screening in 192 womenproved or suspected to be carriers of a breastcancer susceptibility gene: preliminary results.Radiology 2000;215(1):267–79.

[84] Warner E, Plewes DB, Shumak RS, et al.Comparison of breast magnetic resonance im-aging, mammography, and ultrasound for sur-veillance of women at high risk for hereditarybreast cancer. J Clin Oncol 2001;19(15):3524–31.

[85] Morris EA, Liberman L, Ballon DJ, et al. MRI ofoccult breast carcinoma in a high-risk popula-tion. AJR Am J Roentgenol 2003;181(3):619–26.

[86] Warner E, Plewes DB, Hill KA, et al. Surveillanceof BRCA1 and BRCA2 mutation carriers withmagnetic resonance imaging, ultrasound, mam-mography, and clinical breast examination.JAMA 2004;292(11):1317–25.

[87] Kriege M, Brekelmans CT, Boetes C, et al. Effi-cacy of MRI and mammography for breast-can-cer screening in women with a familial orgenetic predisposition. N Engl J Med 2004;351(5):427–37.

[88] Morris EA, Harms S. ACR practice guideline forthe performance of magnetic resonance imag-ing (MRI) of the breast. Reston (VA): AmericanCollege of Radiology; 2004.

[89] ACR breast imaging reporting and data system,breast imaging atlas. Reston (VA): AmericanCollege of Radiology; 2003.

[90] Abraham D, Jones R, Jones S, et al. Evaluationof neoadjuvant chemotherapeutic response oflocally advanced breast cancer by magnetic res-onance imaging. Cancer 1998;78(1):91–100.

[91] Partridge SC, Gibbs JE, Lu Y, et al. Accuracy ofMR imaging for revealing residual breast cancerin patients who have undergone neoadjuvantchemotherapy. AMJ Am J Roentgenol 2002;179(5):1193–9.

[92] Partridge SC, Gibbs JE, Lu Y, et al. MRI measure-ments of breast tumor volume predict responseto neoadjuvant chemotherapy and recurrence-

free survival. AJR Am J Roentgenol 2005;184(6):1774–81.

[93] ACRIN. Contrast-enhanced breast MRI for eval-uation of patients undergoing neoadjuvanttreatment for locally advanced breast cancer:PROTOCOL 6657. Available at: www.acrin.org/6657_protocol.html. Accessed June 19,2006.

[94] Morris EA. Assessment of residual disease. In:Morris EA, Liberman L, editors. Breast MRI: di-agnosis and intervention. New York: Springer;2005. p. 225.

[95] Negendank W. Studies of human tumors byMRS: a review. NMR Biomed 1992;5:303–24.

[96] Bartella L, Morris EA, Dershaw DD, et al.Proton MR spectroscopy with choline peakas malignancy marker improves positivepredictive value for breast cancer diagnosis:preliminary study. Radiology 2006;239(3):686–92.

[97] Meisamy S, Bolan PJ, Baker EH, et al. Adding invivo quantitative 1H MR spectroscopy toimprove diagnostic accuracy of breast MR imag-ing: preliminary results of observer perfor-mance study at 4.0 T. Radiology 2005;236(2):465–75.

[98] Bolan PJ, Meisamy S, Baker EH, et al. In vivoquantification of choline compounds in thebreast with 1H MR spectroscopy. Magn ResonMed 2003;50(6):1134–43.

[99] Jacobs MA, Barker PB, Bottomley PA, et al. Pro-ton magnetic resonance spectroscopic imagingof human breast cancer: a preliminary study.J Magn Reson Imaging 2004;19(1):68–75.

[100] Roebuck JR, Cecil KM, Schnall MD, et al. Hu-man breast lesions: characterization with pro-ton MR spectroscopy. Radiology 1998;209(1):269–75.

[101] Yeung DK, Cheung HS, Tse GM. Human breastlesions: characterization with contrast-en-hanced in vivo proton MR spectroscopy—initialresults. Radiology 2001;220(1):40–6.

[102] Kvistad KA, Bakken IJ, Gribbestad IS, et al.Characterization of neoplastic and normal hu-man breast tissues with in vivo (1)H MR spec-troscopy. J Magn Reson Imaging 1999;10(2):159–64.

[103] Cecil KM, Schnall MD, Siegelman ES, et al. Theevaluation of human breast lesions with mag-netic resonance imaging and proton magneticresonance spectroscopy. Breast Cancer Res Treat2001;68(1):45–54.

[104] Yeung DK, Yang WT, Tse GM. Breast cancer: invivo proton MR spectroscopy in the character-ization of histopathologic subtypes and prelim-inary observations in axillary node metastases.Radiology 2002;225(1):190–7.

[105] Gribbestad IS, Singstad TE, Nilsen G, et al. Invivo 1H MRS of normal breast and breast tu-mors using a dedicated double breast coil.J Magn Reson Imaging 1998;8(6):1191–7.

http://www.acrin.org/6657_protocol.html




[106] Tse GM, Cheung HS, Pang LM, et al. Character-ization of lesions of the breast with proton MRspectroscopy: comparison of carcinomas, be-nign lesions, and phyllodes tumors. AJR Am JRoentgenol 2003;181(5):1267–72.

[107] Huang W, Fisher PR, Dulaimy K, et al. Detec-tion of breast malignancy: diagnostic MR proto-col for improved specificity. Radiology 2004;232(2):585–91.

[108] Bartella L, Liberman L, Thakur S, et al. Magneticresonance spectroscopy (MRS) of the breast.Does sensitivity vary with cancer histology?Radiology 2005;237:177.

[109] Bartella L, Thakur S, Morris E, et al. Proton mag-netic resonance spectroscopy (MRS) of the breast.Does normal breast parenchyma give a false pos-itive choline peak? Radiology 2005;237:178.

[110] Jagannathan NR, Kumar M, Seenu V, et al. Eval-uation of total choline from in-vivo volumelocalized proton MR spectroscopy and its re-sponse to neoadjuvant chemotherapy in locallyadvanced breast cancer. Br J Cancer 2001;84(8):1016–22.

[111] Bolan PJ, Snyder CJ, DelaBarre LJ, et al. Prelim-inary experience with breast 1H MRS at 7Tesla.Proc Int Soc Magn Reson Med 2006;14:580.

[112] Meisamy S, Bolan PJ, Baker EH, et al. Neoadju-vant chemotherapy of locally advanced breastcancer: predicting response with in vivo 1HMR spectroscopy—a pilot study at 4 T. Radiol-ogy 2004;233(2):424–31.

[113] Yeung DKW, Yang W-T, Tse GMK. Breast cancer:in vivo proton MR spectroscopy in the charac-terization of histopathologic subtypes and pre-liminary observations in axillary nodemetastases. Radiology 2002;225(1):190–7.

[114] Sharma U, Mehta A, Seenu V, et al. Biochemicalcharacterization of metastatic lymph nodes ofbreast cancer patients by in vitro 1H magneticresonance spectroscopy: a pilot study. Magn Re-son Imaging 2004;22:697–706.

[115] Seenu V, Pavan Kumar M, Sharma U, et al. Po-tential of magnetic resonance spectroscopy todetect metastasis in axillary lymph nodes inbreast cancer. Magn Reson Imaging 2005;23:1005–10.

[116] Bathen TF, Sensen LR, Sitter B, et al. MAS MRS:a tool for breast cancer grading? Proc Intl SocMag Reson Med 2005;13;130.

[117] Port ER, Yeung H, Gonen M, et al. 18F–2-fluoro-2-deoxy-D-glucose positron emissiontomography scanning affects surgical manage-ment in selected patients with high-risk,

operable breast carcinoma. Ann Surg Oncol2006;13(5):677–84.

[118] Suarez M, Perez-Castejon MJ, Jimenez A, et al.Early diagnosis of recurrent breast cancer withFDG-PET in patients with progressive elevationof serum tumor markers. Q J Nucl Med 2002;46(2):113–21.

[119] Jansson T, Westlin JE, Ahlstrom H, et al. Posi-tron emission tomography studies in patientswith locally advanced and/or metastatic breastcancer: a method for early therapy evaluation?J Clin Oncol 1995;13(6):1470–7.

[120] Schelling M, Avril N, Nahrig J, et al. Positronemission tomography using [(18)F]Fluoro-deoxyglucose for monitoring primary chemo-therapy in breast cancer. J Clin Oncol 2000;18(8):1689–95.

[121] Avril N, Rose CA, Schelling M, et al. Breast im-aging with positron emission tomography andfluorine-18 fluorodeoxyglucose: use and limita-tions. J Clin Oncol 2000;18(20):3495–502.

[122] Murthy K, Aznar M, Thompson CJ, et al. Resultsof preliminary clinical trials of the positronemission mammography system PEM-I: a dedi-cated breast imaging system producing glucosemetabolic images using FDG. J Nucl Med 2000;41(11):1851–8.

[123] Rosen EL, Turkington TG, Soo MS, et al. Detec-tion of primary breast carcinoma with adedicated, large-field-of-view FDG PET mam-mography device: initial experience. Radiology2005;234(2):527–34.

[124] Smith-Jones PM, Solit D, Afroze F, et al. Earlytumor response to Hsp90 therapy using HER2PET: comparison with 18F-FDG PET. J NuclMed 2006;47(5):793–6.

[125] Brem RF, Rapelyea JA, Zisman G, et al. Occultbreast cancer: scintimammography with high-resolution breast-specific gamma camera inwomen at high risk for breast cancer. Radiology2005;237(1):274–80.

[126] Coover LR, Caravaglia G, Kuhn P. Scintimam-mography with dedicated breast camera detectsand localizes occult carcinoma. J Nucl Med2004;45(4):553–8.

[127] Jong RA, Yaffe MJ, Skarpathiotakis M, et al.Contrast-enhanced digital mammography: ini-tial clinical experience. Radiology 2003;228(3):842–50.

[128] Lewin JM, Isaacs PK, Vance V, et al. Dual-energycontrast-enhanced digital subtraction mam-mography: feasibility. Radiology 2003;229(1):261–8.

03 - radiol clin n am 2007 - imaging breast cancer

Documents