Clinical Imaging 47 (2018) 4–8

Contents lists available at ScienceDirect

Clinical Imaging

j ourna l homepage: ht tp : / /www.c l in i ca l imag ing.org

Original Article

Digital breast Tomosynthesis vacuum assisted biopsy forTomosynthesis-detected Sonographically occult lesions☆,☆☆,☆☆☆

Nikki S. Ariaratnam ⁎, Sherrill T. Little, Markus A. Whitley, Kristy FergusonSouth Jersey Radiology Associates, 1307 White Horse Road, Suite A102, Voorhees, NJ 08043, USA

☆ No financial disclosures or funding information.☆☆ This project underwent IRB approval with a waiver

☆☆☆ This material has only been presented at RNovember 30, 2016.

⁎ Corresponding author.E-mail address: nariaratnam@sjra.com (N.S. Ariarat

http://dx.doi.org/10.1016/j.clinimag.2017.08.0020899-7071/© 2017 Published by Elsevier Inc.

a b s t r a c t

a r t i c l e i n f o

Article history:Received 8 June 2017Received in revised form 17 July 2017Accepted 3 August 2017Available online xxxx

Purpose: To assess the utility and pathological results from DBT VAB for lesions occult on 2Dmammography andbreast ultrasound (US).Materials and methods: A retrospective review of 1116 consecutive stereotactic biopsies was performed over27months. DBT VAB was performed for 38 non-calcified lesions which were solely detected using DBT. Imagingfindings and pathology results were reviewed.Results: Pathologic findings were malignant in 8 of 38 lesions [masses (5) and distortion (3)]. High-risk findingsfound in 14 lesions.Conclusion: DBT VAB is easily performed and the majority of cases yield actionable pathologies. Therefore, per-form DBT VAB primarily when available.

© 2017 Published by Elsevier Inc.

Keywords:Digital breast TomosythesisBreastBiopsy

1. Introduction

The latest major shift in screening mammography has been the ad-dition the digital breast tomosynthesis (DBT). This technology allowsthe reader to navigate the breast tissue in thin slice reconstructed im-ages and provides significant gains in the detection of masses and archi-tectural distortion, which may be obscured on standard 2D projections[1]. Previous studies have demonstrated an increase in cancer detectionrate which has solidified this technology in the frontline of breast imag-ing techniques [1,2].

In a typical work-up, a non-calcified lesion appreciated on DBT ei-ther at the time of screening or diagnostic evaluation, is followed byan ultrasound (US) to help localize the lesion and determine if biopsywill be required [3]. DBT-detected architectural distortions have beenshown at surgery to have a high rate of malignancywhenwire localizedand surgically excised [4]. At our institution, when a DBT-detected le-sion meets the mammographic standards for biopsy and is occult on atargeted US, we proceed to digital breast tomosynthesis vacuumassisted biopsy (DBT VAB). The purpose of this study is to present ourinitial findings over a 27-month study period, to examine the

.SNA Scientific Assembly on

nam).

presentation of DBT-detected lesions that then necessitated DBT VABand the types of pathologies found using this biopsy method.

2. Materials and methods

An IRB waiver was obtained for this retrospective review of a pro-spectively-maintained database. DBT VABwas performed at two outpa-tient facilities between December 2012 and February 2016 by twobreast fellowship trained radiologists of five and thirteen of experienceand one interventional radiologist with sixteen years of experiencereading mammography and performing breast procedures. All readerswere certified in tomosynthesis interpretation through proctored re-view of at least 100 cases and had been reading tomosynthesis for oneyear.

We reviewed 1116 consecutive stereotactic biopsies assessed asBreast Imaging Reporting And Data System (BI-RADS) category 4 or 5that were imaged with both conventional digital mammography andDBT during the second and third years that tomosynthesis imagingwas available at our institution.

Two independent breast radiologists separately reviewed each casefor technical feasibility, for sampling under DBT VAB, for lesions lackingcalcification and not seen on standard 2D views, versus conventionalstereotactic prone biopsy. They agreed that there was no appropriatesonographic target seen on targeted US, which was first performed bya breast sonographer and then followed by a radiologist-performedUS. None of the lesions were clinically palpable. All masses were mea-sured in their largest dimension on DBT images and shape and marginswere characterized and recorded according to the BI-RADS lexicon.

http://crossmark.crossref.org/dialog/?doi=10.1016/j.clinimag.2017.08.002&domain=pdfhttp://dx.doi.org/10.1016/j.clinimag.2017.08.002mailto:nariaratnam@sjra.comJournal logohttp://dx.doi.org/10.1016/j.clinimag.2017.08.002http://www.sciencedirect.com/science/journal/http://www.clinicalimaging.org

5N.S. Ariaratnam et al. / Clinical Imaging 47 (2018) 4–8

All the cases of DBT VAB were drawn from DBT-detected lesions onscreening mammography utilizing the CC and MLO Combinationmode of the Selenia Dimensions (Hologic, Bedford, MA). A 90-degreelateral view with DBT was performed for the purposes of localizationand biopsy planning prior to DBT VAB.

2.1. DBT vacuum assisted biopsy

The Eviva and Affirm Breast Biopsy systems (Hologic, Bedford, MA)were used for all DBT VAB. The lesion was targeted in the projectionwhere the target was best seen and the biopsy needle would traversethe least amount of breast tissue. Patients were positioned seated or ina lateral decubitus position (Fig. 1).

A tomosynthesis localizer series was obtained and a cursor wasplaced on the lesion in the appropriate DBT slice. Coordinates weresent to the biopsy unit. Under sterile technique and local lidocaine anes-thetic, the 9 G needle was introduced and positioning was confirmedwith pre- and post-fire 2D stereotactic pair images. Twelve core biopsysamples were obtained. A specimen radiograph was not performed. Abiopsy clip was left at the site of biopsy and confirmed with a final setof tomosynthesis images. The duration of the procedure was measuredfrom the time of the first scout image to the time of the post biopsy clipimage taken immediately after clip deployment. After the procedure,the patient underwent full CC and ML imaging of the biopsied breastto confirm clip placement at the site of the original DBT-detected lesion.

DBT VABwas performed in 38 consecutive patients for non-calcifiedlesions detected solely on DBT and not observed on targeted diagnosticUS. The age range of thewomenwho underwent the procedure was be-tween 38 and 76 years with a mean of 55 years old. In one case, ultra-sound biopsies were first attempted which did not correlate with thesite of the originally detected tomosynthesis-detected architectural dis-tortion (Fig. 2a–f). This patient then proceeded to DBT VAB on the samedate.

In another case, the biopsy target was attempted under standardprone 2D stereotactic biopsy and could not be performed due to inabil-ity to visualize the target. Therefore, the procedure was converted to aDBT VAB. The remainder of the cases proceeded directly fromdiagnosticmammogram recommendation to DBT VAB.

2.2. Data collection

All DBT images, pathology reports, patient history, and mammo-graphic, US, andMRI reportswere reviewedbyone author (NSA). Breastparenchymal density information was collected from the screening re-port and classified according to the Breast Imaging Reporting and Data

Fig. 1. 59-year old patient, seated for upright DBT VAB.

System lexicon, 4th edition. Density measurements were confirmedwith automated software (Volpara, Rochester, NY).

The radiologist who performed the biopsy reviewed the pathologyreports (CBL Pathology, Rye, NY) for concordance. All malignant caseswere considered concordant and sent for surgical excision. All high-risk lesions (radial scars, lobular carcinoma in situ, atypical papillomas,atypical ductal hyperplasia and atypical lobular hyperplasia) were con-sidered concordant and sent for surgical consultation. All benign lesionswere considered concordant and reassessed in six months and subse-quent annual screening exam to assure proper target sampling. Proce-dures were reviewed for complications, including hematoma orinfections.

BreastMRIwas obtained in 2 patients after DBTVAB for preoperativeevaluation of extent of disease.

2.3. Statistical analysis

The percentage of cases falling into malignant, high risk or benigncategorieswas calculated according to the pathology report from the bi-opsy. PPV 3 was defined as the percentage of all known biopsies donebecause of positive screening or diagnostic examinations or additionalimaging evaluations of positive screening examinations that resultedin a tissue diagnosis of cancer within 1 year. First, a lesion-level analysisof PPV 3 of DBT VAB was calculated as the number of malignancies onhistologic examination of the specimen from DBT VAB divided by thetotal number of lesions that underwent DBT VAB. Excised high-risk le-sionswere followedup for theirfinal pathological result to obtain anup-grade rate and a final PPV 3 for this DBT VAB cohort. Analyses wereperformed using statistical software (Stata version 14).

3. Results

DBT VAB procedures accounted for 3.4% of the total stereotactic pro-cedures reviewed during the exam period (38 of 1116). Technical suc-cess was obtained in 38 of the 38 lesions targeted using DBT VAB. Theaverage procedure time was 15 min (range 10–51, median 13 min).One outlier casewith duration of 51minwas due to a vasovagal episodeafter initial positioning and patient anxiety before resuming procedure.

Procedure reports revealed one moderate post biopsy hematomaafter DBT VAB. At the time of 6-month follow up in this patient, whohad a benign biopsy result, there was no significant residual collectionor symptoms at the biopsy site.

Breast density distributions are presented in Table 1.Upon review of the clinical history for past high risk lesion or breast

cancer among the DBT biopsy patients, only one patient had a previoushistory of atypical lobular hyperplasia (ALH) and one patient had a pre-vious history of DCIS. Four of 38 patients (11%) had a 1st degree relativewith history of breast cancer and seven of 38 patients (18%) reported a2nd degree relative with history of breast cancer.

There were equal numbers of masses and architectural distortions,which were pure distortions without mass. The average mass size was5.8 mm (range 2 to 8 mm) and the morphologies were wellcircumscribed oval (10), irregular spiculated (5), and irregularmicrolobulated (1). There were 6 one-view asymmetries appreciatedin only 1 DBT projection.

3.1. Pathologic evaluation

DBTVAB of 38 lesions showed 8malignancies (5 invasive ductal car-cinomas, 1 invasive lobular carcinoma, 2 DCIS). The PPV 3 was 21%.

Fourteen of the 38 DBT VAB yielded high risk lesions, atypical ductalhyperplasia (ADH), ALH, lobular carcinoma in situ (LCIS), papillomaswith atypia and radial sclerosing lesions (RSL), which were recom-mended for surgical consultation. One patient opted for 6 month fol-low-up for RSL and three additional patients did not undergo surgerydue to other medical conditions, after meeting with a surgeon. Of the

Fig. 2. a 57-year old female presents for screening mammogram. Standard CC view shown. b CC DBT image from screening reveals architectural distortion. c Scout image from DBT VABdemonstrates 2 existing biopsy clips at site of attempted US biopsies which did not correlate to the DBT-detected architectural distortion. c, d Post fire images of needle in place at thetarget, difficult to appreciate on standard stereotactic pair and further obscured given presence of lidocaine anesthetic. e Post clip deployment, hourglass clip at the site of DBT VAB. fPathology from DBT VAB revealed invasive ductal carcinoma (H&E stain).

6 N.S. Ariaratnam et al. / Clinical Imaging 47 (2018) 4–8

10 high-risk lesions that were excised, two lesions, an atypicalpapllioma and RSL with LCIS, were upgraded to malignancy, yieldingan upgrade rate of 14%. The atypical papilloma presented as a newwell circumscribedmass and the RSLwith LCIS presentedwith architec-tural distortion (Fig. 3).

When these two upgraded lesions were added to the DBT VABbiopsied malignancies, the total lesion level PPV 3 is 26%. DBT VAByielded actionable pathologies, including high risk lesions and malig-nancies, in 57.9% of cases (22 of 38), requiring a surgeon's consultationor excision.

Table 1Breast density distributions (Total number of patients = 38)

Number of patients Percentage

Fatty 2 5%Scattered fibroglandular densities 12 32%Heterogeneously dense 10 26%Extremely dense 14 37%

Fig. 4. Bar chart demonstrating pathology type and type of lesion presentation onscreening DBT on the x axis and number of patients on the y axis.

7N.S. Ariaratnam et al. / Clinical Imaging 47 (2018) 4–8

Sixteen of 38 DBT VAB resulted in benign pathology (42%). All le-sions were interpreted as concordant by the performing radiologist.Specifically, the two cases of benign architectural distortion yieldedfibroadenomatous hyperplasia with stromal sclerosis (fibrosis) and fatnecrosis respectively. None of the benign biopsies or non-resectedhigh risk lesions had evidence of additional or new malignancy at sub-sequent post biopsy follow up imaging at 6 months and 1 year.

Breast MRI was performed for two patients with malignancy foundon DBT VAB. Both MRI exams found no other additional disease beyondthe sites of DBT-detected malignancy and did not alter surgicalmanagement.

All of the malignant lesions presented as either irregular masses orarchitectural distortion. Fig. 4 demonstrates how the lesions presentedwhen contrasted with their final pathology.

As shown, all six lesions that were defined as 1 view asymmetries,seen only in one tomosynthesis projection, but not in the orthogonalview, were found to be benign.

4. Discussion

The purpose of this studywas to present the results of DBT VAB per-formance for DBT-detected, sonographically occult lesions. Amajority ofthe pathologies fromDBT VAB reviewed in this study yielded actionableresults, including malignancies and high-risk lesions, requiring asurgeon's attention, affirming this easily performed technique in the ex-pedient management of these patients. These findings are similar toother investigations of DBT-detected lesionswithout sonographic corre-late, which have found positive predictive values of up to 47%, specifi-cally for DBT-detected architectural distortions [3,4,5 6].

Fig. 3. a 42 year old female presents for screening mammogram. Standard MLO view shown.breast. Pathology from DBT VAB yielded a radial sclerosing lesion with LCIS.

The vast majority of all cancers detected on screening mammogra-phy present with masses or distortion, which are then found with acarefully performed targeted US using location data enhanced by DBT,allowing for US biopsy, or these suspicious lesions contain calcificationsallowing for conventional stereotactic biopsy [1,6,7]. Therefore, our PPV3 of 26% for DBTVAB remains reasonable considering these factors. Pub-lished guidelines have established the average PPV 3 fall within a rangeof 26–32% [8,9,10].

Themean duration of DBT VABwas 15min (median 13min), whichis comparable to other feasibility studies of DBT VAB, which have foundmean durations of 9 to 13 min with upright DBT VAB when comparedwith 29 min for conventional 2D prone biopsy [11,12]. The techniqueof DBT VAB allows radiologists to confidently target the DBT-detectedabnormalities because the scout is an identical DBT-slice. The clearscout paddle allows a larger field of view than the conventional pronestereotactic biopsy unit, improving visualization of local anatomic land-marks. A prone DBT VAB table, which can aid in approach particularlyfor medial and inferior lesions, is also now commercially available. Alllocalizations for biopsy using a DBT scout were successfully visualizedon initial attempt. In a time when patient anxiety and resource utiliza-tion surrounding breast cancer screening is under constant scrutiny

b MLO DBT image from screening reveals architectural distortion (arrow) in the superior

8 N.S. Ariaratnam et al. / Clinical Imaging 47 (2018) 4–8

[13,14], DBT VAB allows radiologists to efficiently and definitively take apatient from mammographic abnormality to diagnosis.

Othermeans to obtain pathologic diagnosis for DBT-detected lesionshave been described, but each has limitations. Tomosynthesis-guidedwire needle localization requires subsequent anesthesia and a surgicalprocedure—with all their inherent social and economic costs [5]. BreastMRI and consequent MRI biopsy is also a worthwhile consideration ifDBT VAB is not available [4]. However, a suspicious finding on MRI inthe vicinity on the DBT-detected lesions may not correlate followingMRI biopsy and also may lead to additional false positive findings thatrequire biopsy or short term MRI follow-up. In some areas, access tobreast MRI remains limited and the time required to schedule and per-form the diagnostic MRI and biopsy, contraindications such as metallicimplants, inability for patient positioning, and claustrophobia remainbarriers to this modality.

Lobular phenotypes of breast cancer, most commonly presenting asspiculated masses, asymmetry, or architectural distortion, remain elu-sive diagnoses by conventional mammography and even breast MRI,however may be better appreciated with DBT [15,16,17]. DBT uncoversarchitectural distortion, which is a frequentlymissed sign of breast can-cer that may be obfuscated by overlapping tissue, particularly in densebreasts [18]. Furthermore, DBT accentuates irregular margins of massesand spiculation caused by desmoplastic reaction which has been foundto increase the specificity for BIRADS 4 and 5 lesions [1,19]. Therefore,given a DBT-detected true architectural distortion or spiculated mass,in the absence of conventional stereotactic or US correlate, we stronglyadvocate DBT VAB as themost straightforward method to reach impor-tant diagnoses.

There are several limitations to this study. In our institution, screen-ing with DBT was offered to all patients however, was performed ac-cording to patient preference and availability. Consequently, ourretrospective cohort was subject to the possibility of selection bias. Fur-thermore, we did not assess DBT-detected lesionswith breast MRI priorto DBT VAB. MRI may have led to the detection of other important le-sions elsewhere in both breasts and in a DBT-detected lesion with lowsuspicion lesion characteristics, such as well circumscribed margins oronly seen in one DBT projection, one may have not proceeded withthe DBT VAB. The correlation of DBT-detected lesions to MRI-detectedlesions in a screening population has not yet been directly compared, al-though clinical trials are underway [20]. The potential complementarynature of these two modalities, and their respective modes of biopsy,warrant further study.

5. Conclusion

DBT VAB was easily and quickly performed in our patient popula-tion. No significant complications or discordant results were observed.The imagingfindings of themalignancies and high risk lesions in our co-hort were indistinguishable, and therefore tissue sampling was essen-tial. DBT VAB is an excellent way to establish the pathologic diagnosesfor DBT-detected suspicious lesions, particularly if US is negative andMRI is cost- or access-prohibitive. As DBT screening becomes more

prevalent, the need for this mode of biopsy will become ever morecompelling.

References

[1] Skaane P, Bandos AI, Gullien R, et al. Comparison of digital mammography alone anddigital mammography plus tomosynthesis in a population-based screening pro-gram. Radiology 2013;267(1):47–56.

[2] Rose SL, Tidwell AL, Bujnoch LJ, Lushwaha AC, Nordmann AS, Sexton Jr R. Implemen-tation of breast tomosynthesis in a routine screening practice: an observationalstudy. Am J Roentgenol 2013;200(6):1401–8.

[3] Durand MA, Wang S, Hooley RJ, Raghu M, Philpotts LE. Tomosynthesis-detected ar-chitectural distortion: management algorithm with radiologic-pathologic correla-tion. Radiographics 2016;36(2):311–21.

[4] Freer PE, Niell B, Rafferty EA. Preoperative Tomosynthesis-guided needle localizationof Mammographically and Sonographically occult breast lesions. Radiology 2015;275(2):377–83.

[5] Ray KM, Turner E, Sickles EA, Joe BN. Suspicious findings at digital breastTomosynthesis occult to conventional digital mammography: imaging featuresand pathology findings. Breast J 2015;21(5):538–42.

[6] Lamb PM, Perry NM, Vinnicombe SJ, Wells CA. Correlation between ultrasound char-acteristics, mammographic findings and histological grade in patients with invasiveductal carcinoma of the breast. Clin Radiol 2000;55(1):40–4.

[7] Knutzen AM, Gisvold JJ. Likelihood of malignant disease for various categories ofMammographically detected. Nonpalpable Breast Lesions Mayo Clin Proc 1993;68(5):454–60.

[8] Carney PA, Sickles EA, Monsees BS, et al. Identifying minimally acceptable interpre-tive performance criteria for screening mammography. Radiology 2010;255(2):354–61.

[9] Sprague BL, Arao RF, Miglioretti DL, et al. National Performance Benchmarks formodern diagnostic digital mammography: update from the breast cancer surveil-lance consortium. Radiology 2017;283(1):59–69.

[10] Lee CS, Bhargavan-Chatfield M, Burnside ES, Nagy P, Sickles EA. The National Mam-mography Database: preliminary data. Am J Roentgenol 2016;206(4):883–90.

[11] Schrading S, Distelmaier M, Dirrichs T, Detering S, Brolund L, Strobel K, et al. Digitalbreast tomosynthesis-guided vacuum-assisted breast biopsy: initial experiences andcomparison with prone stereotactic vacuum-assisted biopsy. Radiology 2015;274(3):654–62.

[12] Sabatino V et al. Use of digital breast tomosynthesis as a guide in a consecutive seriesof 178 vacuum assisted biopsies: feasibility and clinical usefulness. Presented onNov. 29, 2016 at the Radiological Society of North America 2016 scientific assemblyand annual meeting, Chicago IL, 2016.

[13] Siu AL, et al. Screening for breast cancer: preventive services task force recommen-dation statement. Ann Intern Med 2016;164(4):279–96.

[14] Welch HG, Prorok PC, O'Malley AJ, Kramer BS. Breast-cancer tumor size, Overdiagno-sis, and mammography screening effectiveness. N Engl J Med 2016;375:1438–47.

[15] Yeh ED, Slanetz PJ, EdmisterWB, Talele A,Monticciolo D, DBSO Kopans. Invasive lob-ular carcinoma: spectrum of enhancement and morphology on magnetic resonanceimaging. Breast J 2003;9(1):13.

[16] Moy L, Elias K, Patel V, Lee J, Babb J, Toth H, Mercado C. Is Breast MRI Helpful in theevaluation of inconclusive mammographic findings? Am J Roentgenol 2009;193:986–993.

[17] Evans WP, Warren Burhenne LJ, Laurie L, O'Shaughnessy KF, Castellino RA. Invasivelobular carcinoma of the breast: mammographic characteristics and computer-aideddetection. Radiology 2002;225(1):182–9.

[18] Yankaskas BC, Schell MJ, Bird RE, Desrochers DA. Reassessment of breast cancersmissed during routine screening mammography: a community-based study. Am JRoentgenol 2001;177(3):535–41.

[19] Zuley ML, Bandos AI, Ganott MA, et al. Digital breast tomosynthesis versus supple-mental diagnostic mammographic views for evaluation of noncalcified breast le-sions. Radiology 2013;266(1):89–95.

[20] ECOG-ACRIN Cancer Research Group. Comparison of abbreviated breast mri and dig-ital breast tomosynthesis in breast cancer screening in women with dense breasts.Randomized clinical trial (October 2016-present).

http://refhub.elsevier.com/S0899-7071(17)30153-5/rf0005http://refhub.elsevier.com/S0899-7071(17)30153-5/rf0005http://refhub.elsevier.com/S0899-7071(17)30153-5/rf0005http://refhub.elsevier.com/S0899-7071(17)30153-5/rf0010http://refhub.elsevier.com/S0899-7071(17)30153-5/rf0010http://refhub.elsevier.com/S0899-7071(17)30153-5/rf0010http://refhub.elsevier.com/S0899-7071(17)30153-5/rf0015http://refhub.elsevier.com/S0899-7071(17)30153-5/rf0015http://refhub.elsevier.com/S0899-7071(17)30153-5/rf0015http://refhub.elsevier.com/S0899-7071(17)30153-5/rf0020http://refhub.elsevier.com/S0899-7071(17)30153-5/rf0020http://refhub.elsevier.com/S0899-7071(17)30153-5/rf0020http://refhub.elsevier.com/S0899-7071(17)30153-5/rf0025http://refhub.elsevier.com/S0899-7071(17)30153-5/rf0025http://refhub.elsevier.com/S0899-7071(17)30153-5/rf0025http://refhub.elsevier.com/S0899-7071(17)30153-5/rf0030http://refhub.elsevier.com/S0899-7071(17)30153-5/rf0030http://refhub.elsevier.com/S0899-7071(17)30153-5/rf0030http://refhub.elsevier.com/S0899-7071(17)30153-5/rf0035http://refhub.elsevier.com/S0899-7071(17)30153-5/rf0035http://refhub.elsevier.com/S0899-7071(17)30153-5/rf0035http://refhub.elsevier.com/S0899-7071(17)30153-5/rf0040http://refhub.elsevier.com/S0899-7071(17)30153-5/rf0040http://refhub.elsevier.com/S0899-7071(17)30153-5/rf0040http://refhub.elsevier.com/S0899-7071(17)30153-5/rf0045http://refhub.elsevier.com/S0899-7071(17)30153-5/rf0045http://refhub.elsevier.com/S0899-7071(17)30153-5/rf0045http://refhub.elsevier.com/S0899-7071(17)30153-5/rf0050http://refhub.elsevier.com/S0899-7071(17)30153-5/rf0050http://refhub.elsevier.com/S0899-7071(17)30153-5/rf0055http://refhub.elsevier.com/S0899-7071(17)30153-5/rf0055http://refhub.elsevier.com/S0899-7071(17)30153-5/rf0055http://refhub.elsevier.com/S0899-7071(17)30153-5/rf0055http://refhub.elsevier.com/S0899-7071(17)30153-5/rf0060http://refhub.elsevier.com/S0899-7071(17)30153-5/rf0060http://refhub.elsevier.com/S0899-7071(17)30153-5/rf0065http://refhub.elsevier.com/S0899-7071(17)30153-5/rf0065http://refhub.elsevier.com/S0899-7071(17)30153-5/rf0070http://refhub.elsevier.com/S0899-7071(17)30153-5/rf0070http://refhub.elsevier.com/S0899-7071(17)30153-5/rf0070http://refhub.elsevier.com/S0899-7071(17)30153-5/rf0075http://refhub.elsevier.com/S0899-7071(17)30153-5/rf0075http://refhub.elsevier.com/S0899-7071(17)30153-5/rf0075http://refhub.elsevier.com/S0899-7071(17)30153-5/rf0080http://refhub.elsevier.com/S0899-7071(17)30153-5/rf0080http://refhub.elsevier.com/S0899-7071(17)30153-5/rf0080http://refhub.elsevier.com/S0899-7071(17)30153-5/rf0085http://refhub.elsevier.com/S0899-7071(17)30153-5/rf0085http://refhub.elsevier.com/S0899-7071(17)30153-5/rf0085

Digital breast Tomosynthesis vacuum assisted biopsy for Tomosynthesis-detected Sonographically occult lesions1. Introduction2. Materials and methods2.1. DBT vacuum assisted biopsy2.2. Data collection2.3. Statistical analysis

3. Results3.1. Pathologic evaluation

4. Discussion5. ConclusionReferences