Woo Kyung Moon, MDJung-Gi Im, MDYoung Hwan Koh, MDDong-Young Noh, MDIn Ae Park, MD

Index terms:Breast, US, 00.1298Breast neoplasms, calcifications,

00.3221, 00.3241Breast neoplasms, diagnosis,

00.1298, 00.301

Radiology 2000; 217:849–854

Abbreviations:ACR 5 American College of

RadiologyBI-RADS 5 Breast Imaging Reporting

and Data SystemDCIS 5 ductal carcinoma in situ

1 From the Departments of Radiology(W.K.M., J.G.I., Y.H.K.), Surgery (D.Y.N.),and Pathology (I.A.P.) and Clinical Re-search Institute, Seoul National Univer-sity Hospital and the Institute of Radia-tion Medicine, SNUMRC, 28 Yongon-Dong, Chongno-Gu, Seoul 110-744,Korea. Received December 22, 1999;revision requested January 25, 2000;revision received February 29; ac-cepted March 24. Address corre-spondence to W.K.M. (e-mail: moonwk@radcom.snu.ac.kr).© RSNA, 2000

Author contributions:Guarantor of integrity of entire study,W.K.M.; study concepts, W.K.M., J.G.I.,D.Y.N., Y.H.K.; study design, W.K.M.,J.G.I., D.Y.N., I.A.P.; definition of intel-lectual content, W.K.M., J.G.I., Y.H.K.;literature research, W.K.M., Y.H.K.; clin-ical studies, W.K.M., Y.H.K.; data acqui-sition, W.K.M., Y.H.K.; data analysis,W.K.M., J.G.I., D.Y.N., I.A.P.; statisticalanalysis, W.K.M., Y.H.K.; manuscriptpreparation, W.K.M., J.G.I., D.Y.N., I.A.P.;manuscript editing, W.K.M., J.G.I.; manu-script review, all authors.

US of MammographicallyDetected ClusteredMicrocalcifications1

PURPOSE: To determine whether ultrasonography (US) can depict breast massesassociated with mammographically detected clustered microcalcifications andwhether the visibility at US is different between benign and malignant lesions.

MATERIALS AND METHODS: Ninety-four patients with 100 mammographicallydetected microcalcification clusters prospectively underwent US with a 10- or 12-MHz transducer before mammographically guided presurgical hook-wire localiza-tion. The visibility of breast masses at US was correlated with histologic and mam-mographic findings.

RESULTS: Surgical biopsy revealed 62 benign lesions, 30 intraductal cancers, andeight invasive cancers. At US, breast masses associated with microcalcifications wereseen in 45 (45%) of 100 cases. US depicted more breast masses associated withmalignant (31 [82%] of 38) than with benign (14 [23%] of 62) microcalcifications(P , .001). In malignant microcalcification clusters larger than 10 mm, US depictedassociated breast masses in all 25 cases. There was no statistically significant differ-ence in shape and distribution of calcific particles, as well as in breast composition,at mammography between US visible and invisible groups.

CONCLUSION: Given a known mammographic location, US with a high-frequencytransducer can depict breast masses associated with malignant microcalcifications,particularly clusters larger than 10 mm. US can be used to visualize large clusters ofmicrocalcifications that have a very high suspicion of malignancy.

Clustered microcalcifications may be the only detectable manifestation of early breastcancer. Mammography is very sensitive in the detection of microcalcifications, but be-cause benign calcifications cannot always be distinguished from those indicating malig-nancy, the specificity of mammography remains low. Only 20%–35% of the cases willprove to be cancerous after hook-wire–guided surgical biopsy (1–3). Currently, to ourknowledge, no imaging modalities other than mammography have an accepted role in thedetection and characterization of microcalcifications (4,5).

The advances in ultrasonographic (US) equipment and the refinement of breast imagingtechniques enable radiologists to detect and characterize small lesions better (6–9) and toprovide efficient and economical US guidance for percutaneous procedures (10,11). Aspractitioners gain hands-on experience using US in patients with nonpalpable lesions,both localization procedures and percutaneous biopsies are being performed with increas-ing frequency by means of US guidance as opposed to mammographic guidance (11–13).However, the low capability of US to depict microcalcifications remains a major limitation(14,15). Microcalcifications cannot be depicted with US when they are located insideechogenic, fibroglandular breast tissue because of the difficulty in differentiating themfrom the echogenic interfaces among tissues.

After using a high-frequency transducer, some investigators have reported that USdepicted clustered microcalcifications in breast cancers (16–18). Calcifications associatedwith malignant tumors are more likely to be seen sonographically because most malignantcalcifications occur within the masses as opposed to within echogenic breast parenchyma.A hypoechoic background of tumor enhances the ability of US to enable identification ofthe hyperechoic punctate calcifications. In contrast, benign calcifications, especially infibrocystic diseases, are less likely to be seen at US because most benign calcifications do

849

not occur within masses. Those studieswere, however, retrospective in nature(16,18) and lacked reproducibility of thefindings. In nonpalpable lesions, difficul-ties also arise in ensuring that a lesionthat was visible at US was the same asthat seen on the mammograms (13,18). US demonstration of microcalcifica-tions is considered by many to be unre-liable if not impossible.

The objectives of this prospective studywere to determine whether (a) US per-formed with a high-frequency transducercan demonstrate breast masses associatedwith mammographically detected clus-tered microcalcifications without massdensity and (b) the visibility at US is dif-ferent between benign and malignant le-sions.

MATERIALS AND METHODS

Between July 1997 and March 1999, 94consecutive patients, aged 29–70 years(mean age, 52 years), who had beenscheduled to undergo surgical breast bi-opsy on the basis of clustered microcalci-fications detected at mammography un-derwent US just before and after hook-wire localization. All patients gave fullinformed consent for the study, whichwas approved by our institutional reviewboard. These patients represented 89%(94 of 106) of the women eligible for thestudy during that period. Twelve patientswere excluded from the study because ofvasovagal reactions during the wire local-ization in one patient, bleeding after thewire localization in another patient, andpreviously confirmed malignancy or be-nignity by means of core needle biopsyin four patients. The final six patients,who had microcalcification clusters indiffuse or regional distributions measur-ing greater than 4 cm in maximal diam-eter, were excluded because of the diffi-culty in the correlation of mammo-graphic, US, and histologic findings. Per-cutaneous needle biopsy is rarely used forthe diagnosis of microcalcifications atour institution.

Mammography was performed by us-ing a conventional screen-film techniqueand dedicated equipment (Senographe,600T; GE Medical Systems, Milwaukee,Wis). Routine mediolateral oblique andcraniocaudal mammograms were obtainedin all patients, and additional spot com-pression magnification and true lateralimages were available in all but two pa-tients. Multicentric lesions were seen infour patients, and bilateral lesions wereseen in two patients. Of 100 lesions, 48

lesions were in heterogeneously or ex-tremely dense breasts and 52 lesions werein entirely or predominantly fatty breasts.According to the American College of Ra-diology (ACR) Breast Imaging Reportingand Data System (BI-RADS) (19), the finalassessment was probably benign lesions(category 3) in four cases, suspicious le-sions (category 4) in 67 cases, and highlysuspicious lesions (category 5) in 29 cases.In four cases of probably benign lesions,the patient underwent biopsy because ofpatient or referring clinician preferencebased on clinical grounds (Fig 1).

US examinations were performed byone of two radiologists (W.K.M., J.G.I.)who had at least 3 years experience inbreast US. A 10-5–MHz linear-array probe(HDI 3000; Advanced Technology Labo-ratories, Bothell, Wash) was used in 80patients (83 lesions) and a 12-5–MHz lin-ear-array probe (HDI 5000; AdvancedTechnology Laboratories) was used in 14patients (17 lesions). All US examinationswere performed with the patient inthe supine position with the arms raised.If necessary, the patient was shifted intoan appropriate contralateral posterioroblique position to scan the lateral andinferior parts of the breast. US was fo-cused only on the suspicious area in thebreast. Scanning was done in radial andantiradial planes, as well as longitudinaland transverse planes (6). The examina-tion took approximately 18 minutes(range, 10–40 minutes). At US performedafter hook-wire localization, a sterile gel

(Aquasonic 100; Parker Laboratories, Fair-field, NJ) was used to avoid infection ofthe wound site. The pre- and postlocal-ization US examinations were performedby the same radiologist.

In all 100 cases, hook-wire localizationwas performed by using a fenestratedcompression plate and a 21-gauge hook-wire needle (Kopans spring hook localiza-tion needle; Cook, Bloomington, Ind).All procedures were carefully done toavoid tissue damage or bleeding in thebreast. Final wire placement mammo-grams in craniocaudal, oblique, and truelateral projections were obtained with aradiopaque marker at the wire entrypoint on the skin. The localizing wire wasplaced within 2 mm of the lesions in 89lesions and within 5 mm in 11 lesions.After placement of the hook wire, thepatient moved to the US room for thesecond US examination. PostlocalizationUS was performed to confirm that thelesion seen at prelocalization US was cor-rect. In nonpalpable lesions, difficultiescan arise in ensuring that a lesion that isvisible at US is the same as that seen onthe mammograms. US scans obtained af-ter hook-wire localization were not usedfor image analysis. All patients under-went surgery within 24 hours of the USexaminations. Mammography of thespecimen was performed in all patients,and microcalcifications were confirmedin all 100 lesions.

The 100 clusters of microcalcificationswere categorized into two groups accord-

Figure 1. Fibroadenoma in the right breast of a 36-year-old woman. (a) Mediolateral obliquespot compression and magnification mammogram shows a 5-mm cluster of microcalcifications(arrow) in the upper breast. Most calcifications are coarse, although they vary in size and shape.This case was interpreted as a probably benign lesion, but surgical biopsy was performed becausethe patient had a family history of breast cancer. (b) US scan shows an 11-mm, ovoid, well-defined hypoechoic mass (arrows) in the area corresponding to the microcalcifications at mam-mography. Echogenic dots (arrowheads) within the mass that are suggestive of calcifications areseen, without acoustic shadowing.

850 z Radiology z December 2000 Moon et al

ing to the visibility at US: sonographi-cally visible lesions and invisible lesions.The lesion was considered to be sono-graphically visible when a mass with orwithout sonographically visible micro-calcification was definitely seen at thesuspicious area determined at mammog-raphy and confirmed at US after needlelocalization. We did not attempt to sono-graphically identify microcalcificationswithout an associated US mass.

All mammographic and US imageswere assessed preoperatively by means ofconsensus between the two radiologists.At mammography, the size of the calcificcluster was measured at the greatest di-mension. The shape and distribution ofmicrocalcifications were described ac-cording to the ACR BI-RADS (19). At US,the lesions were described according tosize, shape, orientation, echogenicity,echotexture, margin, boundary echo,acoustic transmission, and US evidenceof calcifications (6). The US maximumdiameter of the mass was defined as thesize of the lesion. The shape of masseswas classified as round, lobular, or irreg-ular. The orientation of masses was clas-sified as wider than tall or taller thanwide according to the anteroposterior totransverse dimension ratio. The echoge-nicity of masses was compared with thatof the subcutaneous fat and classified ashyperechoic, isoechoic, mildly hypo-echoic, markedly hypoechoic, or ane-choic. The echotexture of masses wasclassified as homogeneous or heteroge-neous. Mass margins were classified aswell defined, microlobulated, ill defined,or spiculated. If an echogenic boundarywas seen, it was defined as a thin capsuleor thick halo. Posterior shadowing or en-hancement was considered to be presentwhen an area had relatively less or more

through transmission of sound than waspresent in the surrounding tissue at thesame depth. If punctate echogenic dotssuggestive of calcifications were seenwithin the mass, they were reported. Thetwo radiologists who performed the USexaminations also correlated US findingswith histologic and mammographic find-ings.

To determine whether the visibility atUS is affected by the pathologic condi-tion (benign or malignant disease), breastcomposition (fatty or dense breast), and/ormammographic findings (size, shape, anddistribution of microcalcifications), statisti-cal analysis was performed with a statisticalsoftware system (SAS for Windows, version6.12; SAS Institute, Cary, NC). The Fisherexact test and x2 test were used for inde-pendent samples, and the Mann-WhitneyU test was used for the variables with non-normal distributions. Findings with a Pvalue of less than .05 were considered asstatistically significant. By using the pres-ence of the mass at US as the diagnosticcriterion for malignancy in ACR BI-RADScategory 4 or 5 microcalcifications, a diag-nostic accuracy, including sensitivity, spec-ificity, and positive and negative predictivevalues, was calculated. Because the numberof cases in which we used the 12-5–MHztransducer was small, the results of US per-formed with 10-5–MHz and with 12-5–MHz transducers were reported together.

RESULTS

Surgical biopsy revealed 62 benign le-sions in 56 patients and 30 intraductaland eight invasive carcinomas in 38 pa-tients (Table 1). All invasive carcinomaswere histologically ductal carcinomasand not otherwise specified.

At US, breast masses associated withmammographically detected microcalci-fications were seen in 45 (45%) of 100cases: 23% (14 of 62) in benign and 82%(31 of 38) in malignant microcalcifica-tions, both before and after hook-wirelocalization. In malignant microcalcifica-tion clusters larger than 10 mm, US de-picted associated breast masses in all 25cases (Table 2). The correlations of visi-bility at US with histologic findings andwith mammographic findings are sum-marized in Tables 1 and 2, respectively.The US findings of 45 masses associatedwith microcalcifications are summarizedin Table 3.

Of 14 benign lesions found at US, 10lesions were seen as a solid mass and fouras a cystic mass. In all four cases of fibro-adenoma, US depicted an oval or round,well-defined hypoechoic mass with echo-genic dots within the mass that was sug-gestive of calcifications (Fig 1). In threecases of ductal hyperplasia without atypiaand in one case of microcysts, US depictedmultiple well-defined cysts, 3–6 mm insize, in the area that corresponded to mi-crocalcifications at mammography. In twocases of atypical ductal hyperplasia, USshowed irregular hypoechoic masses with-out posterior shadowing.

Of 31 malignancies found at US, 23cases were DCIS and eight cases were in-filtrating ductal carcinomas (Table 1). Allof the malignant lesions were seen assolid masses with heterogeneous echo-texture but with variable findings. At US,a lobular shape (n 5 14), mild hypoecho-genicity (n 5 16), ill-defined margin (n 513), normal acoustic transmission (n 518), and lack of associated calcifications(n 5 13) were more commonly seen inDCIS (Fig 2), whereas an irregular shape(n 5 6) and calcifications within the mass(n 5 7) were more frequently seen ininvasive cancers (Table 3).

US depicted more breast masses associ-ated with malignant microcalcificationsthan masses associated with benign mi-crocalcifications: 82% (31 of 38) in ma-lignancies and 23% (14 of 62) in benignlesions (P , .001). Breast masses associ-ated with microcalcifications were morefrequently seen in invasive cancers thanin DCIS: 100% (eight of eight) in invasivecancers and 77% (23 of 30) in DCIS (P ,.01). According to subtypes of DCIS,breast masses associated with microcalci-fications were more frequently seen inthe comedo type than in the noncomedotype: 81% (13 of 16) in comedo and 71%(10 of 14) in noncomedo DCIS (P , .05).According to the breast composition,more masses were visible in the dense

TABLE 1Correlation of US Visibility and Histologic Diagnosis of MammographicallyDetected Microcalcifications

Histologic Diagnosis

Visibility at US

Visible(n 5 45)

Nonvisible(n 5 55)

Benign lesions (n 5 62)Microcysts 1 4Sclerosing adenosis 2 5Intraductal papilloma 1 2Ductal hyperplasia without atypia 4 34Atypical ductal hyperplasia 2 3Fibroadenoma 4 0

Ductal carcinoma in situ (n 5 30)Comedo type 13 3Noncomedo type 10 4

Infiltrating ductal carcinoma (n 5 8) 8 0

Volume 217 z Number 3 US of Mammographically Detected Clustered Microcalcifications z 851

breast than in the fatty breast—48% (23of 48) in dense breast and 42% (22 of 52)in fatty breast—but the difference wasnot statistically significant (P . .05). Ac-cording to the mammographic findings,breast masses were more frequently seenin calcific clusters larger than 10 mmthan in calcific clusters smaller than 10mm: 52% (34 of 65) in calcific clusterslarger than 10 mm and 31% (11 of 35) incalcific clusters smaller than 10 mm (P ,.05). There was no significant differencein the shape and distribution of calcificparticles at mammography between thesonographically visible and invisiblegroups (P . .05).

By using the presence of a US mass asthe diagnostic criterion for malignancyin ACR BI-RADS category 4 (n 5 67) or 5(n 5 29) microcalcifications, the sensitiv-ity, specificity, and positive and negativepredictive values of US were 82% (31 of38), 83% (48 of 58), 76% (31 of 41), and87% (48 of 55). For ACR BI-RADS cate-gory 4 (n 5 40) or 5 (n 5 22) lesions largerthan 10 mm, the sensitivity, specificity,and positive and negative predictive val-ues of US were 100% (31 of 31), 84% (31of 37), 81% (25 of 31), and 100% (31of 31).

DISCUSSION

Our study results showed that, even withhigh-frequency transducers, US depictedbreast abnormalities associated with mi-crocalcifications in only 45% (45 of 100)of the cases. However, US depicted breastmasses associated with malignant micro-calcifications in the majority of the cases;breast masses were seen in 82% (31 of 38)of malignant calcifications comparedwith in 23% (14 of 62) of benign calcifi-cations. In malignant microcalcificationclusters larger than 10 mm, US depictedassociated breast masses in all 25 cases. Inthis study, postlocalization US was usedto confirm that the lesion seen on theprelocalization US scan was correct.

US is less sensitive for demonstrationof microcalcifications than is mammog-raphy (20,21). The smaller the calcifica-tions, the lower the sensitivity of US fordepicting them. The low capability to vi-sualize microcalcifications remains a ma-jor limitation for using US as a screeningor diagnostic tool for breast cancers.However, the currently used high-fre-quency transducers can yield a higherpercentage of mammographically visiblecalcifications than could the previouslyused lower-frequency transducers (22,23).With use of high-frequency, correctly fo-

TABLE 2Correlation of US Visibility and Mammographic Findings of MammographicallyDetected Microcalcifications

Mammographic Findings

Benign DCIS* Invasive Cancer

Visible(n 5 14)

Nonvisible(n 5 48)

Visible(n 5 23)

Nonvisible(n 5 7)

Visible(n 5 8)

Nonvisible(n 5 0)

Overall breast compositionFatty 7 26 10 4 5 0Dense 7 22 13 3 3 0

Size of calcific cluster (mm)†

,10 5 17 4 7 2 010–19 8 24 17 0 3 0$20 1 7 2 0 3 0Mean 6 SD 8.9 6 3.5 10.2 6 4.6 15.3 6 4.2 6.6 6 3.9 16.0 6 4.7 0

Shape of microcalcificationRound 1 3 0 0 0 0Punctate 2 3 0 1 0 0Amorphous 3 8 3 2 1 0Heterogeneous 8 32 8 3 3 0Linear branching 0 2 12 1 4 0

Distribution of microcalcificationClustered 11 41 11 3 5 0Linear 1 2 5 1 0 0Segmental 2 5 7 3 3 0

* DCIS 5 ductal carcinoma in situ.† The size of the calcific cluster was measured at the greatest dimension.

TABLE 3US Findings of Breast Masses Associated with MammographicallyDetected Microcalcifications

US FindingsBenign

(n 5 14)DCIS

(n 5 23)Invasive Cancer

(n 5 8)

Size of mass (mm),10 3 3 310–19 10 14 4$20 1 6 1Mean 6 SD 11.1 6 5.0 18.3 6 5.9 13.6 6 4.1

ShapeRound 8 2 0Lobular 2 14 2Irregular 4 7 6

OrientationWider than tall 10 12 3Taller than wide 4 11 5

EchogenicityHyperechoic 1 0 0Isoechoic 4 5 0Mildly hypoechoic 5 16 3Markedly hypoechoic 0 2 5Anechoic 4 0 0

EchotextureHomogeneous 8 0 0Heterogeneous 6 23 8

MarginWell defined 7 0 0Microlobulated 3 10 2Ill defined 4 13 4Spiculated 0 0 2

Boundary echoThin echogenic capsule 2 0 0Thick echogenic halo 1 3 3Absent 11 20 5

Acoustic transmissionShadowing 1 5 4Enhancement 2 0 0Normal 11 18 4

CalcificationsPresent 6 10 7Absent 8 13 1

852 z Radiology z December 2000 Moon et al

cused 10–12-MHz probes, tiny echogenicspots without acoustic shadowing thatcorrespond to the mammographic imagefindings can be seen.

According to the results of a phantomexperiment with a 7.5-MHz transducer,US can depict and enable identificationof minute glass beads even 100 mm indiameter, provided they are scattered inan ideal hypoechoic area (24). The visu-alization of typical malignant microcalci-fications as small as 100–500 mm shouldincrease with state-of-the-art US equip-ment and a higher frequency transducer(25). In our study, echogenic dots withinthe mass that were suggestive of calcifi-cations were seen in 55% (17 of 31) ofmalignant cases. The assumption that USwill more likely depict malignant ratherthan benign calcification was proved inthe current study. However, the visibilityof calcifications within a mass cannot beused to distinguish between benign andmalignant disease; calcifications are sono-graphically visible in some benign lesionsand invisible in some malignancies.

The main benefit of identifying a USabnormality in women with mammo-graphically detected microcalcificationsis to allow the use of US to guide inter-ventional procedures, such as needle bi-opsy and needle localization (18,22).However, US cannot be routinely used toguide biopsy or needle localization ofsuspicious calcifications because it fails todepict the calcifications or an associatedmass in many cases. Our study resultsshowed that the majority of large malig-nant clusters of microcalcifications were

visible at US, whereas small malignantclusters or benign clusters of any sizewere frequently invisible at US. There-fore, it would be reasonable to use US totry to visualize large (.10-mm) clustersof microcalcifications with a high suspi-cion of malignancy (estimated likelihoodof malignancy 75% or higher, using mam-mographic assessment criteria). US-guidedprocedures are less expensive and fasterthan stereotactically guided procedures(11). In addition, for those institutions thatdo not have stereotactic equipment, theuse of US in selected cases (large-area high-suspicion microcalcifications) would ex-tend the role of percutaneous biopsy atthese sites.

Common US features of invasive breastcancers include irregular shape, tallerthan wide orientation, marked hypo-echogenicity, heterogeneous echotex-ture, spiculated margins, and posterioracoustic shadowing (6–9,26). Little isknown about the US features of DCIS be-cause this entity usually manifests as puremammographic calcifications, which, toour knowledge, rarely have been evalu-ated with US. In our study, a lobularshape, mild hypoechogenicity, and nor-mal acoustic transmission were the mostcommon findings of DCIS, and thesefindings are consistent with those of pre-vious reports about DCIS at US (27,28).Similar US findings were also seen insome benign lesions, such as sclerosingadenosis and atypical ductal hyperplasia,and were reported in a radial scar (29).Therefore, the US findings of DCIS seenin our study seem to be nonspecific. Ab-

normal distention of central ducts with-out an associated mass or intracystic pap-illary nodule, which is usually seen inpatients with a nipple discharge or a pal-pable mass, was not observed in our se-ries.

Benign US features were typically seenin some fibrocystic lesions and fibroade-nomas. Multiple small cysts were seen inthe area that corresponded to microcalci-fications at mammography in three casesof ductal hyperplasia without atypia andin one case of microcysts, and a well-defined ellipsoid mass was seen in all fi-broadenomas. On mammograms, the mi-crocalcifications in these cases were oflow to intermediate suspicion in all cases(Fig 1). Whether US can increase thespecificity of mammography and helpsto reduce the number of surgical or corebiopsies performed in women with mi-crocalcifications needs further investiga-tion.

Our study seemed to include a smallerproportion of small DCIS lesions thanexpected; 37% (11 of 30) of DCIS lesionsin our series were smaller than 10 mmcompared with 72% (47 of 65) in theSickles series (30). Thus, our overall re-sults showed a larger percentage of DCISlesions to be visible at US than would beobserved in a practice that routinely de-tects DCIS at a smaller size. In our study,US visibility of mammographic calcifica-tions in dense breasts (48% [23 of 48])was higher than that in fatty breasts(42% [22 of 52]), but the difference wasnot statistically significant. We speculatethat mammographic calcifications withsubtle or small associated masses mightbe obscured by the dense parenchymaand misinterpreted as pure calcifications.

In summary, given a known mammo-graphic location, US performed with ahigh-frequency transducer can depictbreast masses associated with microcalci-fications in a minority (45%) of cases.However, the visibility at US is muchhigher in malignant microcalcifications,particularly clusters larger than 10 mm,than in benign microcalcifications. There-fore, US can be used to visualize largeclusters of microcalcifications that have avery high suspicion of malignancy. Thepotential benefit of US examination forsuspicious breast microcalcifications is toidentify a mass lesion associated with thecalcifications and to guide the needle bi-opsy or hook-wire localization in casesfor which stereotactic biopsy or localiza-tion cannot be performed or is unavail-able.

Figure 2. DCIS, cribriform type, in the left breast of a 43-year-old woman. (a) Craniocaudal spotcompression and magnification mammogram shows a 5-mm cluster of microcalcifications (ar-row) in the deep portion of the breast. The calcifications are fine, irregular, and variable in sizeand shape. (b) US scan shows an 11-mm, ill-defined, hypoechoic mass (arrows) in the areacorresponding to microcalcifications at mammography. Punctate echogenic dots (arrowheads)within the mass are probably due to the calcifications.

Volume 217 z Number 3 US of Mammographically Detected Clustered Microcalcifications z 853

References1. Ciatto S, Cataliotti L, Distante V. Nonpal-

pable lesions detected with mammogra-phy: review of 512 consecutive cases. Ra-diology 1987; 169:99–102.

2. Mayer JE, Eberlein TJ, Stomper PC, Son-nenfeld MR. Biopsy of occult breast le-sions: analysis of 1261 abnormalities.JAMA 1990; 263:2341–2343.

3. Orel SG, Kay N, Reynolds C, Sullivan DC.BI-RADS categorization as a predictor ofmalignancy. Radiology 1999; 211:845–850.

4. Jackson VP. The current role of ultra-sonography in breast imaging. RadiolClin North Am 1995; 33:1161–1170.

5. Westerhof JP, Fischer U, Moritz JD, Oest-mann JW. MR imaging of mammographi-cally detected clustered microcalcifica-tions: is there any value? Radiology 1998;207:675–681.

6. Stavros AT, Thickman D, Rapp CL, Den-nis MA, Parker SH, Sisney GA. Solid breastnodules: use of sonography to distinguishbetween benign and malignant lesions.Radiology 1995; 196:123–134.

7. Gordon PB, Goldenberg SL. Malignantbreast masses detected only by ultra-sound. Cancer 1995; 76:626–630.

8. Zonderland HM, Coerkamp EG, HermansJ, van de Vijver MJ, van Voorthuisen AE.Diagnosis of breast cancer: contributionof US as an adjunct to mammography.Radiology 1999; 213:413–422.

9. Kolb TM, Lichy J, Newhouse JH. Occultcancer in women with dense breasts: de-tection with screening US—diagnosticyield and tumor characteristics. Radiol-ogy 1998; 207:191–199.

10. Parker SH, Jobe WE, Dennis MA, et al.US-guided automated large-core biopsy.Radiology 1993; 187:507–511.

11. Liberman L, Feng TL, Dershaw DD, Mor-ris EA, Abramson AF. US-guided corebreast biopsy: use and cost-effectiveness.Radiology 1998; 208:717–723.

12. Gordon PB, Goldenberg SL, Chan NH.Solid breast lesions: diagnosis with US-guided fine-needle aspiration biopsy. Ra-diology 1993; 189:573–580.

13. Cleverley JR, Jackson AR, Bateman AC.Pre-operative localization of breast micro-calcification using high-frequency ultra-sound. Clin Radiol 1997; 52:924–926.

14. Schwartz GF, Goldberg BB, Rifkin MD,D’Orazio SE. Ultrasonographic localiza-tion of non-palpable breast masses. Ultra-sound Med Biol 1988; 14(suppl 1):23–25.

15. Cilotti A, Bagnolesi P, Moretti M, et al.Comparison of the diagnostic perfor-mance of high-frequency ultrasound as afirst- or second-line diagnostic tool innon-palpable lesions of the breast. EurRadiol 1997; 7:1240–1244.

16. Yang WT, Suen M, Ahuja A, Metreweli C.In vivo demonstration of microcalcifica-tion in breast cancer using high resolu-tion ultrasound. Br J Radiol 1997; 70:685–690.

17. Ranieri E, D’Andrea MR, D’Alessio A, etal. Ultrasound in the detection of breastcancer associated with isolated clusteredmicrocalcifications, mammographicallyidentified. Anticancer Res 1997; 17:2831–2835.

18. Rickard MT. Ultrasound of malignantbreast microcalcifications: role in evalua-tion and guided procedures. Australas Ra-diol 1996; 40:26–31.

19. American College of Radiology. Breastimaging reporting and data system (BI-RADS). 3rd ed. Reston, Va: American Col-lege of Radiology, 1998.

20. Sickles EA. Sonographic detectability ofbreast calcifications. Proc SPIE 1983; 419:51–52.

21. Potterton AJ, Peakman DJ, Young JR. Ul-trasound demonstration of small breastcancers detected by mammographicscreening. Clin Radiol 1994; 49:808–813.

22. Rizzatto G, Chersevani R, Abbona M,

Lombardo VL, Macorig D. High-resolu-tion sonography of breast carcinoma. EurJ Radiol 1997; 24:11–19.

23. Kenzel PP, Hadijuana J, Schoenegg W,Minguillon C, Hosten N, Felix R. Can thehigh-resolution 13-MHz ultrasonographyfacilitate the preoperative localizationand marking of microcalcification clus-ters? Rofo Fortschr Geb Rontgenstr NeuenBildgeb Verfahr 1996; 165:551–556. [Ger-man]

24. Kasumi F. Can microcalcifications lo-cated within breast carcinomas be de-tected by ultrasound imaging? Ultra-sound Med Biol 1988; 14(suppl 1):175–182.

25. Anderson ME, Soo MS, Bentley RC, Tra-hey GE. The detection of breast microcal-cifications with medical ultrasound. JAcoust Soc Am 1997; 101:29–39.

26. Fornage BD, Sneige N, Faroux MJ, AndryE. Sonographic appearance and ultra-sound guided fine-needle aspiration bi-opsy of breast carcinomas smaller than 1cm3. J Ultrasound Med 1990; 9:559–560.

27. Teboul M. Ductal echography. In: MadjarH, Teubner J, Hackeloer BJ, eds. Breastultrasound update. Basel, Switzerland:Karger, 1994; 110–126.

28. Kasumi F, Sakuma H. Ultrasonic image ofnoninvasive carcinomas. In: Madjar H,Teubner J, Hackeloer BJ, eds. Breast ultra-sound update. Basel, Switzerland: Karger,1994; 168–173.

29. Stavros AT. Ultrasound of ductal carci-noma in situ. In: Silverstein MJ, ed. Duc-tal carcinoma in situ of the breast. Balti-more, Md: Williams & Wilkins, 1997;135–158.

30. Sickles EA. Quality assurance: how to au-dit your own mammography practice. Ra-diol Clin North Am 1992; 30:265–275.

854 z Radiology z December 2000 Moon et al