High-Risk Lesions at MRI-Guided Breast Biopsy: Frequency and Rate of Underestimation

Transcription

1 Women s Imaging Original Research Lourenco et al. MRI-Guided Biopsy of High-Risk Breast Lesions Women s Imaging Original Research Ana P. Lourenco Hanan Khalil Matthew Sanford,2 Linda Donegan Lourenco AP, Khalil H, Sanford M, Donegan L Keywords: biopsy, breast, high-risk lesion, MRI, upgrade DOI:0.224/AJR Received September 3, 203; accepted after revision December 2, 203. Presented in part as an oral presentation at the RSNA 20 annual meeting, Chicago, IL. Department of Diagnostic Imaging, Alpert Medical School of Brown University, Rhode Island Hospital, Main Bldg, 3rd Fl, 593 Eddy St, Providence, RI Address correspondence to A. P. Lourenco (alourenco@lifespan.org). 2 Present address: Department of Radiology, Sanford Health of Northern Minnesota, Bemidji, MN. AJR 204; 203: X/4/ American Roentgen Ray Society High-Risk Lesions at MRI-Guided Breast Biopsy: Frequency and Rate of Underestimation OBJECTIVE. The purpose of this article is to determine the underestimation rate of highrisk lesions diagnosed at MRI-guided breast biopsy. MATERIALS AND METHODS. This was a retrospective review of 446 MRI-guided breast biopsies from January 2006 through December 200. Data were collected on examination indication, lesion size and type, and pathology results. Biopsies were performed with a 9-gauge vacuum-assisted device. Biopsy results of atypical ductal hyperplasia (ADH), papillary lesion, radial scar, lobular neoplasia, and atypia were identified and compared with final excisional pathology results. Underestimation rates were calculated and data were compared by patient and lesion characteristics using chi-square analysis. RESULTS. Of the 446 MRI-guided biopsies, 96 (2.5%) were high-risk lesions. Fortytwo of 96 lesions (44%) were masses, and 54 (56%) showed nonmass enhancement. Twenty of 96 lesions (20.8%) were ADH, nine (9.4%) were lobular neoplasia, 27 (28.%) were papillary lesions, 20 (20.8%) were radial scar, and 20 (20.8%) were other atypias. Sixty-nine of 96 lesions (7.9%) had surgical excisional pathology results available. Sixteen of 69 (23.2%) lesions were upgraded to malignancy; of the 6 (68.8%) were upgraded to ductal in situ () and five (3.2%) were upgraded to invasive. The underestimation rate was 3.6% (6/9) for ADH, 5.9% (/7) for papillary lesions, 23.% (3/3) for radial scar, 28.6% (2/7) for lobular neoplasia, and 30.8% (4/3) for other atypias (p = 0.43). There was no statistically significant difference in underestimation rate by lesion type, size, or history of newly diagnosed breast cancer. CONCLUSION. MRI-guided breast biopsy yielded high-risk lesions in 2.5% of cases, and the underestimation rate was 23.2%. No patient or lesion characteristics correlated with underestimation rate. C ontrast-enhanced breast MRI is a sensitive method for identifying breast cancer and can detect lesions not readily depicted by mammography or ultrasound []. Percutaneous core-needle sampling of lesions detected by mammography, ultrasound, or MRI often obviates surgical excision if the results are benign [2 7]. However, high-risk lesions may be histologically upgraded to ductal in situ () or invasive cancer at the time of surgical excision [3, 8, 9]. For this reason, surgical excision is often recommended after a needle biopsy finding of a high-risk lesion. High-risk histologic lesions, such as atypical ductal hyperplasia (ADH), lobular neoplasia, papillary lesions, and radial scar, have poorly understood and variable biologic profiles [0 3]. Upgrade rates for stereotactic and ultrasound-guided core biopsies have been studied extensively [3, 8, 3 27]. The most common high-risk lesion, ADH, has variable upgrade rates that are most dependent on the needle gauge and number of samples obtained at the time of stereotactic vacuum-assisted sampling [3, 6, 9 24]. Specifically, ADH is upgraded approximately 50% of the time with a 4-gauge needle and 20% of the time with an -gauge needle [3, 9 24]. When ADH is sampled with a 9-gauge needle at the time of MRI-guided biopsy, it is upgraded to or invasive cancer up to 50% of the time [4, 5, 28]. Only a few studies have attempted to identify the pathologic upgrade rates for the most commonly encountered high-risk lesions at the time of MRI biopsy [4, 5, 28, 29]. However, because breast MRI and MRIguided biopsy are increasingly used in screening high-risk patients or evaluating 682 AJR:203, September 204

2 MRI-Guided Biopsy of High-Risk Breast Lesions patients with newly diagnosed breast cancer, information about the upgrade rate for highrisk lesions is critical to informed patient counseling. To our knowledge, this is the largest single study of high-risk lesions identified at MRI-guided biopsy. The purpose of this study is to evaluate the upgrade rate for high-risk lesions at MRI-guided biopsy and to correlate upgrade rates with lesion and patient characteristics. Materials and Methods Patient Population This institutional review board approved HIPAA-compliant retrospective study reviewed 446 consecutive MRI-guided breast biopsies performed from January 2006 through December 200 at two institutions with dedicated breast imaging centers. The institutional review board waived patient consent. Breast MRI Technique and Interpretation Dynamic contrast-enhanced breast MRI was performed at our institutions on closed.5-t (January 2006 through October 200) and 3-T (from October 200 onward) systems. A dedicated seven-channel breast coil was used from January 2006 through October 200, and a 6-channel coil was used from October 200 onward. Gadoliniumbased contrast agent was administered at a concentration of 0. mmol/kg. From 2006 through 2007, gadodiamide (Omniscan, Amersham Health) was used. From 2007 (n = 93 patients) through 200, gadopentetate dimeglumine (Magnevist, Bayer HealthCare Pharmaceuticals) was used (n = 353 patients). Premenopausal women were imaged on days 7 4 of the menstrual cycle unless the MRI was done for a newly diagnosed cancer, when patients were scheduled as soon as possible. Breast MRI technique at.5 T included bilateral axial dynamic 3D T-weighted fat-suppressed gradient-recalled echo pulse sequences obtained before and after contrast agent was administered. Dynamic images were obtained every 70 seconds for a total of six acquisition times. Section thickness was 2 mm. Sagittal contrast-enhanced images of each breast were obtained after the dynamic axial acquisition. Subtraction images were obtained from the dynamic series. Fluid-sensitive T2-weighted images were also obtained. At 3 T, bilateral axial dynamic 3D T-weighted gradient-recalled echo pulse sequences were obtained before and at, 2, and 6 minutes after administration of contrast agent. Section thickness was mm. In addition, a higher-spatial-resolution axial sequence was performed between the early and delayed phases of the dynamic scans, with a 0.8-mm slice thickness. Sagittal reformatted contrast-enhanced images were created from this higher resolution sequence. Subtraction images were created from the dynamic series. STIR images were also obtained. All datasets were reviewed on a PACS workstation (Centricity, GE Healthcare; or Horizon Medical Imaging, McKesson Medical Imaging) with computer-aided diagnosis (DynaCAD version , Invivo) immediately adjacent and available for review of lesion enhancement kinetics, multiplanar reconstructions, and maximumintensity-projection images. Examinations were interpreted in conjunction with any available prior breast imaging and clinical history. All breast MRI studies were interpreted by one of seven radiologists with 3 0 years of experience in breast MRI and 3 5 years of experience in MRI-guided biopsy. The same radiologists performed all MRI-guided biopsies. Lesions detected by MRI were categorized using the BI-RADS MRI lexicon as outlined by the American College of Radiology [30]. Patients with BI-RADS category 4 or 5 lesions identified by MRI underwent a targeted ultrasound before MRI biopsy at the discretion of the reporting radiologist. In our practice, targeted ultrasound is recommended most commonly for masses and is infrequently used for nonmass lesions. If a sonographic correlate is identified, the biopsy is done with ultrasound guidance. In cases where ultrasound is negative, MRI-guided biopsy is done. Only cases where the biopsy was done with MRI guidance are included in this study. This was done to avoid the inevitable ambiguity that exists about whether the lesion detected on MRI corresponds to the lesion identified and biopsied at ultrasound. Biopsy Technique After we obtained standard informed consent and performed patient identification, the patient was positioned prone in a dedicated breast biopsy coil with the breast in compression using a grid device. Unenhanced and contrast-enhanced fat-suppressed gradient-echo T-weighted images were obtained in the axial and sagittal planes. After lesions were identified on the MRI console, the biopsy images were transferred to a standalone computer with dedicated breast MRI biopsy software (DynaCAD version , Invivo) and the lesion was targeted. Using standard biopsy technique, which has been described elsewhere [3], a minimum of six core biopsy samples were obtained using a 9-gauge vacuum-assisted device (Suros ATEC, Suros Surgical Systems, Hologic). A biopsy marking clip was placed into the biopsy cavity at the conclusion of the biopsy, and a mammogram was performed to verify clip placement. Diagnostic MR images, biopsy MR images, and postprocedure mammograms were reviewed in conjunction to determine whether the lesion of interest was accurately targeted. Core needle biopsy specimens were placed in formalin for histopathologic analysis. Surgical excision is routinely recommended for high-risk lesions in our practice. Patients who did not undergo mastectomy instead underwent preoperative mammographic needle localization of the biopsy marking clip. Specimen radiography was subsequently done to confirm excision of the biopsy marking clip. In patients undergoing mastectomy, the tissue was imaged mammographically and the biopsy marking clip was localized for pathology. Pathologists noted the presence of the biopsy site or marking clip in their reports. Data Collection and Analysis Data were collected from the MRI report and electronic medical record on patient age, indication for MRI, lesion type (mass vs nonmass) and size, number of biopsy cores obtained, presence or absence of known breast cancer, and pathology results from needle biopsy and surgical excision. All cases were included in the study if the MRI biopsy yielded ADH, atypia (flat epithelial atypia or atypia not otherwise specified), atypical lobular hyperplasia, lobular in situ, radial scar, papillary lesion (both with and without atypia), or papilloma. Biopsies with malignant findings and those with benign findings without high-risk lesions were excluded. Cases with high-risk lesions on needle biopsy, but without available surgical excisional pathology, were also excluded for purposes of upgrade rate calculation. Surgical excisional pathology reports were compared with needle biopsy reports, and upgrade rates to and invasive cancer were calculated for each high-risk lesion. Data were also analyzed comparing upgrade rates by lesion type (mass vs nonmass lesion enhancement), lesion size, and presence or absence of newly diagnosed breast cancer, using chi-square analysis (Stata version 2, Stata- Corp). Chi-square analysis was also used to compare the upgrade rates. Results Patient and MRI Lesion Characteristics Of the 446 MRI-guided needle biopsies, 96 (2.5%) found high-risk lesions. Three hundred twenty-nine MRI examinations and 94 biopsies were done at.5 T, and 7 MRI examinations and two biopsies were done at 3 T. A mean of 0 core biopsy samples (range, 6 3 samples) were obtained. The mean patient age was 52 years (range, years). The average lesion size on MRI was.6 cm (range, cm). Indications for breast MRI were preoperative staging in 45 of 96 (46.9%) cases, high-risk screening (lifetime risk of breast cancer > 20% as calculated by AJR:203, September

3 Lourenco et al. 9 ADH 2 Invasive 4 7 Lobular neoplasia Invasive TABLE : Distribution of High-Risk Lesions Found at 96 MRI-Guided Biopsies High-Risk Lesion the referring provider, most frequently using the Gail or Tyrer-Cuzick models) in 4 (42.7%) cases, and other reasons (including abnormalities on mammogram or ultrasound, 6-month follow-up of probably benign findings on MRI or after benign MRI-guided biopsy, and palpable abnormality with negative mammogram and ultrasound) in 0 (0.4%) cases. For 45 patients with newly diagnosed breast cancers, the MRI biopsy target was in the contralateral breast in 28 (62.2%) cases, in a different quadrant of the ipsilateral breast in 5 (33.3%) cases, and in the same quadrant of the ipsilateral breast in two (4.4%) cases. Both of these biopsies in the same quadrant of the ipsilateral breast were done for linear nonmass enhancement extending anteriorly toward the nipple over more than 5 cm. Forty-two of the 96 high-risk lesions (44%) were masses and 54 (56%) were nonmass enhancement. The distribution of high-risk lesions is shown in Table. Pathology findings from surgical excision were available in 69 of 96 (7.9%) cases. Of these, 6 (23.2%) lesions were upgraded to malignancy ( [68.8%] were upgraded to and five [3.2%] were upgraded to invasive ). Details are shown in Figure. Of the 27 women without excisional pathology findings available, 2 had benign follow-up breast imaging findings (one with mammography alone, six with MRI alone, and five with both mammography and MRI). The mammography followup interval after the high-risk MRI biopsy was 2 6 years (mean, 4.2 years). The range of MRI follow-up was 6 months to 5 years (mean, 2.6 years). Clinical notes in the electronic medical record reported that three patients underwent surgery at an outside institution, and none were upgraded to malignancy. Twelve patients did not have any available follow-up information. No. (%) of Cases Atypical ductal hyperplasia 20 (20.8) Lobular neoplasia 9 (9.4) Papillary lesion 27 (28.) Radial scar 20 (20.8) Other atypia 20 (20.8) TABLE 2: Underestimation Rate for 69 High-Risk Lesions With Available Surgical Excisional Pathology Results High-Risk Lesion Underestimation Rate, No. of Lesions/Total (%) Atypical ductal hyperplasia 6/9 (3.6) Lobular neoplasia 2/7 (28.6) Papillary lesion /7 (5.9) Radial scar 3/3 (23.) Other atypia 4/3 (30.8) Note p = 0.43 for all underestimation rates. 96 High-risk MRI biopsy results 69 With corresponding excisional pathology 7 Papillary lesion 0 Invasive Underestimation and Correlation to Lesion and Patient Characteristics The underestimation rate by high-risk lesion is shown in Table 2. The underestimation rate for masses was 2.9% (7/32) and that for nonmass enhancement was 24.3% (9/37) (p = 0.8). The underestimation rate was 2.4% (6/28) for lesions smaller than cm and 24.4% (0/4) for lesions cm or larger (p = 0.77). Underestimation rates were 3 Radial scar Invasive 2 3 Other atypia Invasive 3 Ductal Mixed ductal and lobular Lobular Tubular Lobular Fig. Pathology results of high-risk lesions identified at MRI-guided biopsy. ADH = atypical ductal hyperplasia, = ductal in situ. 684 AJR:203, September 204

4 MRI-Guided Biopsy of High-Risk Breast Lesions 27.3% (9/33) for patients with newly diagnosed breast cancer and 9.4% (7/36) for those without a known cancer (p = 0.44). Discussion MRI-guided breast biopsies are now frequently performed as breast MRI has become more widely available. Data regarding the risk of upgrade to malignancy is important for patient counseling and for appropriate management recommendations when high-risk lesions are encountered during an MRI-guided breast biopsy. Although several studies [4, 5, 28, 29] have looked at upgrade rates for high-risk lesions, the numbers of patients in all studies is small, making it important to continue evaluating additional data as practice patterns evolve. Our study found high-risk lesions in 2.5% of MRIguided biopsies, which is at the higher end of the range previously reported (3 2%) [29]. This may be related to the high-risk status of most patients in this study, where nearly 90% of all examinations were performed for either newly diagnosed breast cancer (46.9%) or for high-risk screening (42.7%). In prior studies [4, 5], 80 83% of breast MRI examinations were done for newly diagnosed breast cancer and high-risk screening, which is slightly lower than in our study. The underestimation rate in our study, 23.2% for all high-risk lesions combined, is within the range previously reported [29]. The rate of underestimation was highest for ADH (3.6%) and lowest for papillary lesions (5.9%). Our study differs from previously published reports in that it contains the highest number of papillary lesions and radial scars reported in studies evaluating MRI-detected high-risk lesions. Some studies of high-risk lesions have excluded papillary lesions [6, 7, 4, 32, 33], but others have reported upgrade rates for papillary lesions ranging from 0% to 29% [28, 34, 35]. Our relatively low upgrade rate for papillary lesions raises questions about the need for surgical excision in patients with this result at needle biopsy. The upgrade rate for radial scars was 23.% in this study, which supports surgical excision. Importantly, our study adds to the number of previously published reports of radial scar at MRI-guided biopsy, which was quite low, ranging from one to three [5, 6, 4, 28]. The increased numbers of radial scars and the high upgrade rate in this study may be, at least in part, related to the high-risk patient population. Our study found no significant difference in upgrade rates when comparing mass versus nonmass enhancement, lesion size, and patient history of newly diagnosed breast cancer versus no known breast cancer. This is in keeping with other published studies [4, 28, 29]. This is not surprising, given the biologic heterogeneity of high-risk lesions and the inherent sampling limitations in MRI-guided needle biopsy. It may be useful to evaluate the number of biopsy cores obtained in relation to underestimation rate in future studies. This study is limited by lack of surgical follow-up in 27 of 96 (28%) of cases, though this is within the range reported by others, which ranged up to 36% [4]. However, half of these patients without surgical follow-up had benign follow-up breast imaging, making the likelihood of occult malignancy in these patients relatively low. Additionally, pathology results were taken from the reports and slides were not rereviewed for this study. Because some patients were evaluated by MRI at higher field strength (3 vs.5 T), a selection bias is possible. A smaller percentage of radial scars (65% [3/20]) and papillary lesions (65% [3/20]) were excised as compared with ADH (95% [9/20]) and lobular neoplasia (78% [7/9]), which is another limitation of the study. However, this may be related to the controversial management of these particular high-risk lesions, because there remains considerable debate about the need for subsequent surgical excision when these lesions are identified on needle biopsy. Also, pathologists were not blinded to the clinical history or prior biopsies at the time of their interpretation, which may have biased their interpretations. However, this is reflective of actual clinical practice, where pathologists have such information accessible at the time of interpretation. Finally, there is interobserver variability in pathology reporting for atypias [36] and for distinguishing ADH from [37]. In conclusion, this study shows highrisk lesions are identified in 2.5% of all MRI-guided breast biopsies with an overall underestimation rate of 23.2%, which supports prior studies. We report the largest number of radial scars and papillary lesions identified at MRI-guided breast biopsy, with a relatively high underestimation rate of 23.% for radial scars and a relatively low underestimation rate for papillary lesions of 5.9%. This information may help guide patient management decisions about surgical excision. Acknowledgment We thank Steven E. Reinert (Life Span Corporation) for assistance with statistical analysis. References. Kriege M, Brekelmans CT, Boetes C, et al. Efficacy of MRI and mammography for breast-cancer screening in women with a familial or genetic predisposition. N Engl J Med 2004; 35: Jackman RJ, Nowels KW, Rodriguez-Soto J, Marzoni FA Jr, Finkelstein SI, Shepard MJ. Stereotactic, automated, large-core needle biopsy of nonpalpable breast lesions: false-negative and histologic underestimation rates after long-term follow-up. Radiology 999; 20: Kohr JR, Eby PR, Allison KH, et al. Risk of upgrade of atypical ductal hyperplasia after stereotactic breast biopsy: effects of number of foci and complete removal of calcifications. Radiology 200; 255: Lehman CD, Deperi ER, Peacock S, McDonough MD, Demartini WB, Shook J. Clinical experience with MRI-guided vacuum-assisted breast biopsy. AJR 2005; 84: Liberman L, Bracero N, Morris E, Thornton C, Dershaw DD. MRI-guided 9-gauge vacuum-assisted breast biopsy: initial clinical experience. AJR 2005; 85: Orel SG, Rosen M, Mies C, Schnall MD. MR imaging-guided 9-gauge vacuum-assisted core-needle breast biopsy: initial experience. Radiology 2006; 238: Perlet C, Heywang-Kobrunner SH, Heinig A, et al. Magnetic resonance-guided, vacuum-assisted breast biopsy: results from a European multicenter study of 538 lesions. Cancer 2006; 06: Foster MC, Helvie MA, Gregory NE, Rebner M, Nees AV, Paramagul C. Lobular in situ or atypical lobular hyperplasia at core-needle biopsy: is excisional biopsy necessary? Radiology 2004; 23: Liberman L, Dershaw DD, Glassman JR, et al. Analysis of cancers not diagnosed at stereotactic core breast biopsy. Radiology 997; 203: Brem RF, Lechner MC, Jackman RJ, et al. Lobular neoplasia at percutaneous breast biopsy: variables associated with at surgical excision. AJR 2008; 90: Brenner RJ, Jackman RJ, Parker SH, et al. Percutaneous core needle biopsy of radial scars of the breast: when is excision necessary? AJR 2002; 79: Douglas-Jones AG, Denson JL, Cox AC, Harries IB, Stevens G. Radial scar lesions of the breast diagnosed by needle core biopsy: analysis of cases containing occult malignancy. J Clin Pathol 2007; 60: Jackman RJ, Birdwell RL, Ikeda DM. Atypical duc- AJR:203, September

5 Lourenco et al. tal hyperplasia: can some lesions be defined as prob- AF, Hann LE, Rosen PP. Atypical ductal hyper- 29. Heller SL, Moy L. Imaging features and manage- ably benign after stereotactic -gauge vacuum-as- plasia diagnosed at stereotaxic core biopsy of ment of high-risk lesions on contrast-enhanced sisted biopsy, eliminating the recommendation for breast lesions: an indication for surgical biopsy. dynamic breast MRI. AJR 202; 98: surgical excision? Radiology 2002; 224: AJR 995; 64: D Orsi CJ, Mendelson EB, Ikeda DM, et 4. Strigel RM, Eby PR, Demartini WB, et al. Fre- 22. Liberman L, Kaplan JB, Morris EA, Abramson al. Breast Imaging Reporting and Data System: quency, upgrade rates, and characteristics of high- AF, Menell JH, Dershaw DD. To excise or to sam- ACR BI-RADS breast imaging atlas. Reston, risk lesions initially identified with breast MRI. AJR 200; 95: Liberman L, Holland AE, Marjan D, et al. Underestimation of atypical ductal hyperplasia at MRIguided 9-gauge vacuum-assisted breast biopsy. AJR 2007; 88: Lourenco AP, Mainiero MB, Lazarus E, Giri D, Schepps B. Stereotactic breast biopsy: comparison of histologic underestimation rates with - and 9-gauge vacuum-assisted breast biopsy. AJR 2007; 89:[web]W275 W Niell B, Specht M, Gerade B, Rafferty E. Is excisional biopsy required after a breast core biopsy yields lobular neoplasia? AJR 202; 99: Liberman L, Tornos C, Huzjan R, Bartella L, Morris EA, Dershaw DD. Is surgical excision warranted after benign, concordant diagnosis of papilloma at percutaneous breast biopsy? AJR 2006; 86: Liberman L. Percutaneous image-guided core breast biopsy. Radiol Clin North Am 2002; 40: [vi] 20. Jackman RJ, Nowels KW, Shepard MJ, Finkelstein SI, Marzoni FA Jr. Stereotaxic large-core needle biopsy of 450 nonpalpable breast lesions with surgical correlation in lesions with cancer or atypical hyperplasia. Radiology 994; 93: Liberman L, Cohen MA, Dershaw DD, Abramson ple the mammographic target: what is the goal of stereotactic -gauge vacuum-assisted breast biopsy? AJR 2002; 79: Winchester DJ, Bernstein JR, Jeske JM, et al. Upstaging of atypical ductal hyperplasia after vacuumassisted -gauge stereotactic core needle biopsy. Arch Surg 2003; 38:69 622; discussion, Pandelidis S, Heiland D, Jones D, Stough K, Trapeni J, Suliman Y. Accuracy of -gauge vacuumassisted core biopsy of mammographic breast lesions. Ann Surg Oncol 2003; 0: Chang JM, Moon WK, Cho N, et al. Management of ultrasonographically detected benign papillomas of the breast at core needle biopsy. AJR 20; 96: Youk JH, Kim EK, Kim MJ. Atypical ductal hyperplasia diagnosed at sonographically guided 4-gauge core needle biopsy of breast mass. AJR 2009; 92: Jang M, Cho N, Moon WK, Park JS, Seong MH, Park IA. Underestimation of atypical ductal hyperplasia at sonographically guided core biopsy of the breast. AJR 2008; 9: Crystal P, Sadaf A, Bukhanov K, McCready D, O Malley F, Helbich TH. High-risk lesions diagnosed at MRI-guided vacuum-assisted breast biopsy: can underestimation be predicted? Eur Radiol 20; 2: VA: American College of Radiology, Liberman L, Morris EA, Dershaw DD, Thornton CM, Van Zee KJ, Tan LK. Fast MRI-guided vacuum-assisted breast biopsy: initial experience. AJR 2003; 8: Rauch GM, Dogan BE, Smith TB, Liu P, Yang WT. Outcome analysis of 9-gauge MRI-guided vacuum-assisted core needle breast biopsies. AJR 202; 98: Tozaki M, Yamashiro N, Sakamoto M, Sakamoto N, Mizuuchi N, Fukuma E. Magnetic resonance-guided vacuum-assisted breast biopsy: results in 00 Japanese women. Jpn J Radiol 200; 28: Tseng HS, Chen YL, Chen ST, et al. The management of papillary lesion of the breast by core needle biopsy. Eur J Surg Oncol 2009; 35: Daniel BL, Gardner RW, Birdwell RL, Nowels KW, Johnson D. Magnetic resonance imaging of intraductal papilloma of the breast. Magn Reson Imaging 2003; 2: Darvishian F, Singh B, Simsir A, Ye W, Cangiarella JF. Atypia on breast core needle biopsies: reproducibility and significance. Ann Clin Lab Sci 2009; 39: Page DL, Rogers LW. Combined histologic and cytologic criteria for the diagnosis of mammary atypical ductal hyperplasia. Hum Pathol 992; 23: AJR:203, September 204