Nonsurgical Management of High- Risk Lesions Diagnosed at Core Needle Biopsy: Can Malignancy Be Ruled Out Safely With Breast MRI?

Transcription

1 Women s Imaging Original Research Linda et al. MRI of High-Risk Breast Lesions Women s Imaging Original Research FOCUS ON: Anna Linda 1 Chiara Zuiani 1 Alessandro Furlan 2 Michele Lorenzon 1 Viviana Londero 1 Rossano Girometti 1 Massimo Bazzocchi 1 Linda A, Zuiani C, Furlan et al. Keywords: borderline lesions, breast MRI, core needle biopsy, high-risk lesions, underestimation DOI: /AJR Institute of Diagnostic Radiology, Azienda Ospedaliero Universitaria Santa Maria della Misericordia, P.le S. Maria della Misericordia, Udine, Italy. Address correspondence to A. Linda (annalinda33@gmail.com). 2 Department of Radiology, University of Pittsburgh, Pittsburgh, PA. AJR 2012; 198: X/12/ JOURNAL Received April 15, 2011; accepted after revision July 22, American Roentgen Ray Society CLUB Nonsurgical Management of High- Risk Lesions Diagnosed at Core Needle Biopsy: Can Malignancy Be Ruled Out Safely With Breast MRI? OBJECTIVE. The purpose of this study was to investigate whether breast MRI can be used to rule out malignancy in patients with high-risk lesions diagnosed at imaging-guided core needle biopsy. SUBJECTS AND METHODS. The subjects were women consecutively registered between October 2004 and April 2010 who had high-risk lesions diagnosed at mammographically or sonographically guided core needle biopsy and subsequently underwent MRI and surgical excision. MR images were reviewed by two experienced breast radiologists. Lesions assessed as BI-RADS category 1 3 were considered negative for malignancy, and BI-RADS 4 and 5 lesions were considered malignant. Histologic findings at surgical excision were the reference standard. The sensitivity, specificity, and positive and negative predictive values of MRI in the detection of associated malignancy were calculated for the entire set of lesions and for each histologic subtype. RESULTS. The final sample consisted of 169 high-risk lesions in 166 patients. At MRI analysis, 116 (68.6%) lesions were considered negative for malignancy, and the other 53 (31.4%) malignant. At surgical excision, 22 malignant lesions were found. The overall sensitivity, specificity, and positive and negative predictive values of MRI were 72.7% (16/22), 74.8% (110/147), 30.2% (16/53), and 94.8% (110/116). The negative predictive values for papilloma, radial scar, lobular neoplasia, and atypical ductal hyperplasia were 97.4% (38/39), 97.6% (41/42), 88.0% (22/25), and 90.0% (9/10). CONCLUSION. Patients with high-risk lesions associated with the lowest likelihood of malignancy (papilloma and radial scar) and without suspicious MRI findings can safely undergo follow-up instead of surgery. Because of the low negative predictive value, however, MRI is not helpful in cases of lobular neoplasia and atypical ductal hyperplasia, and all these lesions should be excised. T he pathologic diagnosis highrisk lesion accounts for as many as 9% of all imaging-guided core needle biopsies (CNBs) [1 6]. These lesions include lobular neoplasia (LN) (lobular carcinoma in situ, atypical lobular hyperplasia), atypical ductal hyperplasia (ADH), papilloma, and radial scar. Because the reported risk of associated malignancy at surgical excision seems to vary widely (0 35%) [2 7], these lesions are usually managed surgically. However, surgical excision implies increased cost, patient anxiety [8], and morphologic alteration (i.e., scarring) of the breast that can hamper interpretation of later images [9]. The role of conventional imaging (mammography and sonography) in the management of high-risk lesions has been investi- gated, but published data are contradictory and nonconclusive [10 18]. In two studies [19, 20], breast MRI was evaluated for preoperative detection of malignancy associated with high-risk lesions, and the results were promising. The studies, however, were limited by retrospective design [19] and a relatively small sample size, which did not allow analysis by lesion [19, 20]. We undertook a large prospective study to investigate whether breast MRI can be used to safely rule out malignancy in patients with high-risk lesions diagnosed at imaging-guided CNB. Subjects and Methods The institutional review board approved this study, which was performed at a large university referral hospital for breast diseases. All patients provided written informed consent to participate in the study. 272 AJR:198, February 2012

2 MRI of High-Risk Breast Lesions Study Sample This prospective study was performed between October 2004 and April Included were women 18 years old and older with a diagnosis of highrisk lesion (LN, ADH, papilloma, radial scar) at imaging-guided CNB who subsequently underwent breast MRI and surgical excision of the lesion. The exclusion criteria were presence of synchronous or metachronous carcinoma (ductal carcinoma in situ [DCIS], invasive carcinoma) in the same breast; contraindications to MRI (pacemaker, claustrophobia, vascular clip); severe limitation to MR image interpretation (e.g., movement artifacts); lack of final pathologic result after surgical excision because patient declined surgical treatment or was referred to an outside hospital; and surgical excision performed more than 3 months after CNB. Patient enrollment and consent took place at the time of diagnosis of a high-risk lesion. Percutaneous Core Needle Biopsy Percutaneous CNB was performed under mammographic or sonographic guidance according to the judgment of the attending radiologist. The biopsy procedures were performed by radiologists with years experience in breast imaging. In cases of sonographically guided CNB, an automated biopsy gun (Magnum Biopsy Instrument, Bard) or a semiautomated biopsy gun (Precisa, HS Hospital Service) with a 14-gauge, 15-cm-long needle (throw of needle, 23 mm) was used. A mean of five core samples (range, three to eight) were obtained per lesion. In cases of small lesions, an amagnetic, sonographically visible clip (GelMark, UltraCor) was left in place to mark the biopsy site. In cases in which a mammographically detected lesion was not identified at sonographic examination, CNB was performed under mammographic guidance on a digital prone table (Mammobed Giotto, IMS) with a directional vacuum- biopsy device (Mammotome, Mammotome) with an 11-gauge needle. On average, 12 core samples (range, nine to 18) were obtained per lesion. In patients with mammographically detected microcalcifications, a specimen radiograph was obtained to confirm the presence of calcifications in each sample, and a clip (MammoMark, Mammotome) was left in place to mark the biopsy site. MRI Technique Breast MRI studies were performed with a 1.5- T system (Magnetom Avanto, Siemens Healthcare) with a dedicated bilateral surface breast coil and the patient prone. For premenopausal patients, MRI was performed during the second week of the menstrual cycle. From October 2004 through May 2009, coronal dynamic contrast-enhanced images were obtained with a T1-weighted 3D FLASH pulse sequence with the following parameters: TR/ TE, 15/4.7; flip angle, 25 ; matrix size, ; FOV, mm; section thickness, 1.8 mm; acquisition time, 1:28 minutes. From June 2009 through April 2010, axial dynamic contrast-enhanced images were obtained with a T1-weighted 3D FLASH pulse sequence with the following parameters: TR/TE, 9/4.7; flip angle, 25 ; matrix size, ; FOV, mm; section thickness, 2 mm; acquisition time, 1:20 minutes. Gadobenate dimeglumine (MultiHance, Bracco) was administered IV by automated bolus injection at a dose of 0.1 mmol/kg body weight with a flow rate of 2 ml/s immediately followed by flushing of 20 ml of saline solution with an automatic injector (Spectris Solaris, MedRad). Images were acquired sequentially once before and five times after injection of the contrast agent, beginning 12 seconds after initiation of the contrast injection and with no delay between images. Postprocessing and construction of dynamic curves were performed at a workstation (Syngo MultiModality Workplace, Leonardo, Siemens Healthcare) by one of five radiologists (more than 2 years of experience in breast MRI). Postprocessing included temporal subtraction (all contrast-enhanced images minus unenhanced images) and generation of multiplanar reconstruction and maximum intensity projections. Dynamic signal intensity time curves were constructed for regions of interest (3 3 pixels) positioned within the lesion on subjectively determined areas of maximal enhancement. MRI Analysis The study coordinator (more than 5 years of experience in breast imaging), who was not involved in further MRI evaluation, reviewed all images available and annotated on a breast map the location of each biopsied lesion using clock-face referents and relative distance from the nipple. The maps were distributed to the readers before image interpretation. At acquisition, two independent radiologists with more than 10 years of experience in breast imaging, including breast MRI, prospectively evaluated the images at the workstation. Any discrepancy in opinion was resolved by consensus. The readers were aware of the lesion location and of the histologic diagnosis at CNB. So that information on the performance of MRI would be obtained independently of mammographic and sonographic features, mammographic and sonographic images were not available to the readers at MR image interpretation. Unenhanced and contrast-enhanced MR images were evaluated. Morphologic and kinetic evaluation and lesion assessment were based on the BI-RADS MRI lexicon [21]. In keeping with the BI-RADS lexicon, an assessment of BI-RADS category 1 was used to indicate absence of contrast enhancement in the area of the lesion; BI-RADS 2, a benign finding; BI-RADS 3, a probably benign finding; BI-RADS 4, a suspicious finding; and BI-RADS 5, a finding highly suggestive of malignancy. Because of the selection criteria, BI-RADS category 6 (known cancer) was not considered an option. Reference Standard The reference standard was the result at histopathologic examination of the surgically excised specimen. Invasive cancer and DCIS were classified as malignant, and all other pathologic findings as benign. Histopathologic examinations were performed by a breast pathologist with more than 10 years of experience. Statistical Analysis Rates of underestimation of malignancy (number of high-risk lesions upgraded to malignancy at surgical excision divided by the number of highrisk lesions surgically excised) were calculated overall, for each high-risk lesion type (radial scar, papilloma, LN, ADH), and for biopsy type (sonographically guided CNB, mammographically guided vacuum- biopsy). For the purposes of computing the diagnostic yield of MRI, results of image evaluation were dichotomized as negative for malignancy (BI- RADS categories 1 3) and positive for malignancy (BI-RADS categories 4 and 5). Malignant lesions assessed as positive on MR images were considered true-positive cases and those assessed as negative were considered false-negative cases. Benign lesions assessed as positive on MR images were considered false-positive cases, and those assessed as negative were considered true-negative cases. On the basis of these data, the sensitivity, specificity, and positive and negative predictive values of MRI in the detection of malignancy were calculated. Exact 95% CI was computed for each performance measure. The performance parameters were specifically calculated for each high-risk lesion and for biopsy type. The Pearson chi-square test was used to compare performance parameters among groups. A value of p < 0.05 was considered indicative of a statistically significant difference. All statistical analyses were performed with commercially available software (MedCalc , MedCalc Software). Results Patients and Lesions A total of 3243 imaging-guided CNBs of the breast were performed in our department during the study period. Among them, 2514 AJR:198, February

3 TABLE 1: Imaging Features of 22 Lesions Diagnosed as High-Risk at Core Needle Biopsy With Finding of Malignancy After Surgical Excision Core Needle Biopsy Final Pathologic Finding Conventional Imaging Finding MRI Finding Mass Nonmasslike Enhancement BI-RADS Category Signal Intensity Time Curve Type Internal enhancement Internal Enhancement Distribution Shape Margin Size (mm) a Mammography Ultrasound Diagnosis Type Histologic Type, Grade Assessment of Finding Lesion No. 1 b False-negative Papilloma Ultrasound core needle DCIS, low 4 Occult Nodule Lobulated Irregular Homogeneous True-positive Papilloma Ultrasound core needle DCIS, low NA NA Dilated ducts Irregular Smooth Homogeneous True-positive Papilloma Ultrasound core needle DCIS, high 30 NA Dilated ducts Lobulated Irregular Homogeneous Regional Inhomogeneous True-positive Papilloma Ultrasound core needle DCIS, intermediate NA Occult Nodule Oval Irregular Inhomogeneous True-positive Radial scar Ultrasound core needle ILC, intermediate 5 Distortion Nodule Irregular Irregular Inhomogeneous False-negative Radial scar Ultrasound core needle DCIS, low 8 Occult Nodule 1 ILC, intermediate 4 Microscopic Occult Irregular Irregular Inhomogeneous True-positive LN Mammographic vacuum DCIS, intermediate 4 Microscopic Occult 1 8 False-negative LN Mammographic vacuum DCIS, high 2 Microscopic Occult Irregular Irregular Inhomogeneous True-positive LN Mammographic vacuum 20 Occult Nodule Lobulated Irregular Inhomogeneous True-positive LN Ultrasound core needle Invasive tubulolobular carcinoma, intermediate Linda et al. 11 False-negative LN Ultrasound core needle ILC, intermediate 4 Occult Nodule Oval Regular Homogeneous True-positive LN Ultrasound core needle ILC, intermediate 10 Occult Nodule Irregular Irregular Inhomogeneous True-positive LN Ultrasound core needle ILC, intermediate 11 Occult Nodule Oval Irregular Inhomogeneous False-negative LN Mammographic vacuum DCIS, low 5 Microscopic Occult 1 15 c True-positive LN Mammographic vacuum DCIS, low NA Microscopic Occult Ductal Inhomogeneous True-positive LN Ultrasound core needle ILC, intermediate 5 Occult Nodule Oval Irregular Inhomogeneous True-positive LN Ultrasound core needle DCIS, low NA Occult Nodule Irregular Irregular Inhomogeneous True-positive LN Mammographic vacuum DCIS, intermediate NA Microscopic Occult Oval Irregular Inhomogeneous False-negative ADH Ultrasound core needle DCIS, intermediate NA Occult Hypoechoic 1 area 20 True-positive ADH Mammographic vacuum DCIS, low 5 Microscopic Occult Segmental Inhomogeneous True-positive ADH Ultrasound core needle DCIS, low 10 Occult Hypoechoic Focal Inhomogeneous 3 4 area 7 Microscopic Occult Round Irregular Inhomogeneous 2 4 Invasive ductallobular carcinoma, intermediate 274 AJR:198, February True-positive ADH Mammographic vacuum Note Note Dash ( ) indicates not assessable. DCIS = ductal carcinoma in situ, NA = not available, ILC = invasive lobular carcinoma, LN = lobular neoplasia, ADH = atypical ductal hyperplasia. a Neoplastic component. b Figure 4. c Figure 1.

4 MRI of High-Risk Breast Lesions (77.6%) were performed under sonographic guidance and 729 (22.4%) under stereotactic guidance. Two hundred forty-one (7.4%) high-risk lesions were diagnosed consecutively in 236 women. Three patients (three lesions) were excluded because of the presence of ipsilateral breast cancer. Four patients (four lesions) had contraindications to MRI. All other patients consented to participation in the study and were enrolled. However, 63 patients (65 lesions) did not complete the study because they were later found to be ineligible owing to the presence of artifacts limiting MRI interpretation (eight patients with eight lesions) or because they declined surgery (55 patients with 57 lesions). The final sample included 166 patients (mean age, 52.2 years; range, years) with 169 high-risk lesions. Of these patients, 164 had one high-risk lesion, one patient had two synchronous ipsilateral high-risk lesions, and one patient had three synchronous highrisk lesions, two in one breast and one in the other breast. Of the 169 high-risk lesions, 122 (72.2%) were biopsied under sonographic guidance and 47 (27.8%) under mammographic guidance. At examination of the biopsy specimen, 64 (37.9%) lesions were found to be papilloma without atypia; 54 (31.9%), radial scar; 35 (20.7%), LN; and 16 (9.5%), ADH. The mean interval between biopsy and MRI was 11 days (range, 3 25 days). Histologic Results at Surgical Excision Surgical excision was performed days after biopsy (mean, 23 days) and 6 69 days after MRI (mean, 19 days). Histopathologic examination of the surgical specimens revealed 22 malignant lesions. Of these, 14 (63.6%) were DCIS (eight low grade, four intermediate grade, and two high grade), and the other eight (36.4%) were invasive carcinoma (six intermediate-grade invasive lobular carcinoma, one intermediate-grade invasive ductal-lobular carcinoma, one intermediate-grade invasive tubulolobular carcinoma) (Table 1). The overall biopsy underestimation rate was 13% (22/169). The underestimation rates among lesions diagnosed under sonographic guidance and lesions diagnosed under mammographic guidance were 11.5% (14/122) and 17.0% (8/47) (p = 0.559) (Table 2). MRI Analysis Against Reference Standard At breast MRI analysis, 116 (68.6%) lesions were assessed as negative: 59 (34.9%) BI-RADS category 1, 12 (7.1%) BI-RADS TABLE 2: Rate of Underestimation of Malignancy at Biopsy Pathologic Diagnosis No. of Lesions at Biopsy Type of Biopsy a category 2, and 45 (26.6%) BI-RADS category 3. The other 53 (31.4%) lesions were considered positive: 51 (30.2%) BI-RADS category 4 and 2 (1.2%) BI-RADS category 5. Of the 53 lesions assessed as positive at MRI, 16 (30.2%) lesions (all BI-RADS 4) proved malignant (true-positive) (Fig. 1) at surgical excision, and 37 (69.8%) lesions (35 BI-RADS 4, two BI-RADS 5) proved benign (falsepositive) (Fig. 2). Of 116 lesions classified as negative at MRI, 110 (94.8%) lesions (55 BI-RADS 1, 12 BI-RADS 2, 43 BI-RADS 3) were confirmed as benign at final pathologic examination (true-negative) (Fig. 3). Malignancy Underestimation Rate (%) b By Biopsy Type Overall Papilloma 64 Ultrasound core needle (60) 6.7 (4/60) 6.2 (4/64) Mammographic vacuum (4) 0 (0/4) Radial scar 54 Ultrasound core needle (41) 4.9 (2/41) 3.7 (2/54) Mammographic vacuum (13) 0 (0/13) Lobular neoplasia 35 Ultrasound core needle (15) 40.0 (6/15) 34.3 (12/35) Mammographic vacuum (20) 30.0 (6/20) Atypical ductal 16 Ultrasound core needle (6) 33.3 (2/6) 25.0 (4/16) hyperplasia Mammographic vacuum (10) 20.0 (2/10) Total 169 Ultrasound core needle (122) 11.5 (14/122) 13.0 (22/169) Mammographic vacuum (47) 17.0 (8/47) Note Values in parentheses are numbers of lesions. CNB = core needle biopsy. a Values in parentheses are numbers of lesions. b Values in parentheses are number of lesions upgraded. A The other six (5.2%) lesions (four BI-RADS 1, two BI-RADS 3) were upgraded to malignancy (false-negative) (Fig. 4). The overall sensitivity, specificity, and positive and negative predictive values of MRI in the identification of malignancy were 72.7% (95% CI, 65 79%; 16/22), 74.8% (95% CI, 67 81%; 110/147), 30.2% (95% CI, 23 38%; 16/53), and 94.8% (95% CI, 90 97%; 110/116). The performance parameters for lesions diagnosed at sonographically guided CNB and for those diagnosed at mammographically guided vacuum- biopsy are shown in Table 3. The negative predictive values for papil- Fig year-old woman with low-grade ductal carcinoma in situ and true-positive findings at MRI (lesion 15). A, Magnification mammogram (retroareolar region of left breast) shows 10-mm cluster of amorphous calcifications (arrow). Ultrasound findings were normal. Mammographically guided vacuum- biopsy result was lobular neoplasia. B, Axial T1-weighted contrast-enhanced subtracted MR image (TR/TE, 9/4.7; flip angle, 25 ) shows nonmasslike enhancing lesion (arrow) with ductal distribution and inhomogeneous enhancement in retroareolar region of left breast. Lesion had type 2 signal intensity curve and was classified BI-RADS category 4. B AJR:198, February

5 Linda et al. loma, radial scar, LN, and ADH were 97.4% (95% CI, %; 38/39), 97.6% (95% CI, %; 41/42), 88.0% (95% CI, 72 96%; 22/25), and 90.0% (95% CI, 63 99%; 9/10). All performance parameters for each highrisk lesion type are reported in Table 4. Discussion Our data showed that breast MRI can be used to identify high-risk lesions associated with a low likelihood of upgrade to malignancy at surgical excision. The overall sensitivity is 72.7%; specificity, 74.8%; positive predictive value, 30.2%; and negative predictive value, 94.8%. These results suggest a possible role for this noninvasive imaging modality in the workup of high-risk lesions. Despite differences in study methods, our findings are similar to those reported in a retrospective study [19] of 79 lesions evaluated with the Baum-Fisher score [22] and in another prospective study [20] of 32 lesions. The results of both studies confirmed that MRI has high negative predictive value for malignancy (98.2% and 96%) in the evaluation of high-risk lesions. Strigel et al. [23] evaluated 39 high-risk lesions initially detected with MRI and found no specific morphologic MRI features A Fig year-old woman with radial scar and false-positive MRI findings. A, Left craniocaudal mammogram shows area of focal asymmetric density and architectural distortion (arrow) in outer quadrants of breast. B, Correlative ultrasound scan shows hypoechoic area (arrow) with irregular shape and margins and posterior acoustic shadowing measuring 20 mm. Sonographically guided 14-gauge core needle biopsy result was radial scar. C, Coronal T1-weighted contrast-enhanced subtracted MR image (TR/TE, 15/4; flip angle, 7.25 ) shows masslike enhancing lesion (arrow) with irregular shape and margins and inhomogeneous enhancement measuring 22 mm in outer quadrants of left breast. Lesion had type 3 signal intensity curve and was classified as BI-RADS category 4. TABLE 3: Performance of MRI Based on Type of Biopsy Parameter B Mammographic Vacuum-Assisted Biopsy predictive of upgrade to malignancy. All of those lesions were enhancing, but we found that approximately one half of benign highrisk lesions were not enhancing (classified BI-RADS category 1). On the other hand, four of 22 high-risk lesions upgraded to malignancy (one radial scar, two LN, one ADH) in our series were nonenhancing at MRI. All these lesions corresponded to small DCIS of low or intermediate grade (Table 1). It is reasonable to believe that absence of enhancement may be a strong indicator of absence of invasive carcinoma. Further studies are necessary to clarify this issue. High negative predictive values for malignancy were found for papilloma without atypia and radial scar (97.4% and 97.6%), compared with LN and ADH (88% and 90%). Presumably, the variability in likelihood of malignancy among high-risk lesions may affect the usefulness of MRI in prediction of the presence or absence of malignancy. Papillomas and radial scars diagnosed at CNB are associated with much lower prevalence of malignancy at surgical excision than are LN and ADH (6.2% and 3.7% versus 34.3% and 25.0% in our study) [2 6]. Because of the low prevalence of disease, it is not surprising that MRI had a high negative predictive value in these two specific groups [24]. If on the basis of low risk of malignancy, all papillomas and radial scars with normal MRI findings in our study had been hypothetically addressed with follow-up [25], 79 unnecessary surgi- Biopsy Type Ultrasound Core Needle Biopsy Sensitivity 62.5 (47 76) 78.6 (70 85) Specificity 94.9 (83 99) 67.6 (58 76) Positive predictive value 71.4 (56 83) 23.9 (17 33) Negative predictive value 92.5 (80 98) 96.1 (90 99) Note Biopsy type values are percentages with 95% CI in parentheses. p C 276 AJR:198, February 2012

6 MRI of High-Risk Breast Lesions TABLE 4: Overall Performance of MRI Lesion Type Parameter cal procedures would have been avoided, and two false-negative cases (one in each group) would have been missed. Both these falsenegative lesions were found to be low-grade DCIS (4 and 8 mm) at pathologic examination of the surgical specimen (lesions 1 and 6, Table 1). It has been reported that absence of enhancement on MR images is observed in 20 60% of cases of low-grade DCIS [26, Papilloma Radial Scar Lobular Neoplasia Atypical Ductal Hyperplasia Sensitivity 75.0 (62 85) 50.0 (36 64) 75.0 (57 87) 75.0 (47 92) 72.7 (65 79) Specificity 63.3 (50 75) 78.8 (65 88) 95.7 (81 99) 75.0 (47 92) 74.8 (67 81) Positive predictive value 12.0 (6 23) 8.3 (3 20) 90.0 (74 97) 50.0 (25 74) 30.2 (23 38) Negative predictive value 97.4 (89 100) 97.6 (88 100) 88.0 (72 96) 90.0 (63 99) 94.8 (90 97) Note Data are percentages with 95% CI in parentheses. A 27]. On the other hand, low-grade DCIS has a benign biologic profile and if left undetected (and untreated), it is likely to remain dormant or to exhibit slow (decades long) progression to invasive carcinoma [28]. Therefore, diagnostic delay due to noninvasive management (imaging follow-up) in our two false-negative cases would probably not have a substantial effect on prognosis. Overall We propose that women with papilloma without atypia and those with radial scar with normal MRI findings may safely undergo clinical and radiologic follow-up. A possible interval time for follow-up might be every 6 months for 2 years with both conventional imaging and MRI for prompt recognition of interval onset of contrast enhancement [20]. MRI had an unacceptably Fig year-old woman with sclerosing papilloma and true-negative MRI findings. A, Mediolateral oblique mammogram shows low-density oval opacity (arrow) with regular margins in upper retroareolar area of right breast. B, Ultrasound scan corresponding to A shows oval hypoechoic nodule with regular margins (calipers) measuring 8 mm. Sonographically guided 14-gauge core needle biopsy result was sclerosing papilloma without atypia. C, Axial T1-weighted contrast-enhanced subtracted MR image (TR/TE, 9/4.7; flip angle, 25 ) shows oval mass (arrow) with smooth margins and homogeneous enhancement in upper retroareolar area of right breast. Lesion had type 2 signal intensity curve and was classified BI-RADS category 3. A B B C Fig year-old woman with lowgrade ductal carcinoma in situ (4 mm) and false-negative findings at MRI (lesion 1). A, Ultrasound scan of left breast shows hypoechoic nodule (arrow) with lobulated shape and slightly irregular margins measuring 11 mm in upper outer quadrant of breast. Mammographic findings were normal. Sonographically guided 14-gauge core needle biopsy result was papilloma without atypia. B, Axial T1-weighted contrast-enhanced subtracted MR image (TR/TE, 9/4; flip angle, 7.25 ) shows lobulated mass (arrow) with predominantly regular margins and homogeneous enhancement in upper outer quadrant of left breast. Lesion had type 2 signal intensity curve and was classified BI-RADS 3. AJR:198, February

7 Linda et al. low negative predictive value for malignancy in patients with LN and ADH (88% and 90%), corresponding to 12% and 10% falsenegative rates. According to our data, MRI cannot be recommended in the management decision process for LN and ADH. These lesions should be excised because of the high upgrade rate. Further studies are needed to establish the most appropriate time intervals and modalities for imaging follow-up, to assess cost-effectiveness, and to gauge patients perceptions of this approach. Our study had limitations. First was the small number of ADH lesions corresponding to 9.5% of all high-risk lesions compared with the number in most other studies [2, 4 6, 20]. Selection bias can be excluded on the basis of the study design (consecutive prospective recruitment). Instead, the underrepresentation of ADH, which usually presents as microcalcifications [29, 30], may be attributed to the small number of biopsies performed under mammographic guidance in our series (47/169). In addition, the pathologic differential diagnosis of ADH and low-grade DCIS is known to be challenging and to have substantial interobserver variability, even when strict criteria are used [31, 32]. This finding further emphasizes the importance of by-lesion analysis to ensure reproducibility of the results for differing spectra of high-risk lesions. The second limitation was the small proportion of lesions diagnosed at mammographically guided biopsy (21.9%). However, our results are in agreement with those of a previous study [20] of 32 high-risk lesions all diagnosed at sonographically guided CNB. In addition, we found no statistically significant differences between mammographically guided vacuum- biopsy and sonographically guided CNB with respect to the performance of MRI in the diagnosis of lesions. Third, analysis of interobserver variability in MRI interpretation in this clinical situation was not performed. Fourth, the acquisition plane of MR images was coronal in the initial phase of the study period and in the axial plane later. However, both imaging planes are currently considered appropriate for bilateral dynamic MRI [33]. Conclusion Our results suggest that patients with highrisk lesions associated with the lowest likelihood of malignancy (papilloma without atypia and radial scar) and without suspicious MRI findings can safely undergo follow-up instead of surgical excision. Patients with ADH and LN seem not to benefit from MRI, and these lesions should be surgically excised. Acknowledgment We thank Luisa Battigelli, University of Udine, Italy, for data collection and image preparation. References 1. Schueller G, Jaromi S, Ponhold L, et al. US-guided 14-gauge core-needle breast biopsy: results of a validation study in 1352 cases. Radiology 2008; 248: Lee AH, Denley HE, Pinder SE, et al. Excision biopsy findings of patients with breast needle core biopsies reported as suspicious of malignancy (B4) or lesion of uncertain malignant potential (B3). Histopathology 2003; 42: Londero V, Zuiani C, Linda A, Battigelli L, Brondani G, Bazzocchi M. Borderline breast lesions: comparison of malignancy underestimation rates with 14-gauge core needle biopsy versus 11-gauge vacuum- device. Eur Radiol 2011; 21: Rakha EA, Lee AH, Jenkins JA, Murphy AE, Hamilton LJ, Ellis IO. Characterization and outcome of breast needle core biopsy diagnoses of lesions of uncertain malignant potential (B3) in abnormalities detected by mammographic screening. Int J Cancer 2010; 129: El-Sayed ME, Rakha EA, Reed J, Lee AH, Evans AJ, Ellis IO. Predictive value of needle core biopsy diagnoses of lesions of uncertain malignant potential (B3) in abnormalities detected by mammographic screening. Histopathology 2008; 53: Houssami N, Ciatto S, Bilous M, Vezzosi V, Bianchi S. Borderline breast core needle histology: predictive values for malignancy in lesions of uncertain malignant potential (B3). Br J Cancer 2007; 96: Georgian-Smith D, Lawton TJ. Controversies on the management of high-risk lesions at core biopsy from a radiology/pathology perspective. Radiol Clin North Am 2010; 48: Golub RM, Bennett CL, Stinson T, Venta L, Morrow M. Cost minimization study of image-guided core biopsy versus surgical excisional biopsy for women with abnormal mammograms. J Clin Oncol 2004; 22: March DE, Raslavicus A, Coughlin BF, Klein SV, Makari-Judson G. Use of breast core biopsy in the United States: results of a national survey. AJR 1997; 169: Brem RF, Lechner MC, Jackman RJ, et al. Lobular neoplasia at percutaneous breast biopsy: variables associated with carcinoma at surgical excision. AJR 2008; 190: Nagi CS, O Donnell JE, Tismenetsky M, Bleiweiss IJ, Jaffer SM. Lobular neoplasia on core needle biopsy does not require excision. Cancer 2008; 112: Lam WW, Chu WC, Tang AP, Tse G, Ma TK. Role of radiologic features in the management of papillary lesions of the breast. AJR 2006; 186: Shin HJ, Kim HH, Kim SM, et al. Papillary lesions of the breast diagnosed at percutaneous sonographically guided biopsy: comparison of sonographic features and biopsy methods. AJR 2008; 190: Youk JH, Kim EK, Kwak JY, Son EJ, Park BW, Kim SI. Benign papilloma without atypia diagnosed at US-guided 14-gauge core-needle biopsy: clinical and US features predictive of upgrade to malignancy. Radiology 2011; 258: Linda A, Zuiani C, Furlan A, et al. Radial scars without atypia diagnosed at imaging-guided needle biopsy: how often is associated malignancy found at subsequent surgical excision, and do mammography and sonography predict which lesions are malignant? AJR 2010; 194: Brenner RJ, Jackman RJ, Parker SH, et al. Percutaneous core needle biopsy of radial scars of the breast: when is excision necessary? AJR 2002; 179: Youk JH, Kim EK, Kim MJ. Atypical ductal hyperplasia diagnosed at sonographically guided 14-gauge core needle biopsy of breast mass. AJR 2009; 192: Hoang JK, Hill P, Cawson JN. Can mammographic findings help discriminate between atypical ductal hyperplasia and ductal carcinoma in situ after needle core biopsy? Breast 2008; 17: Linda A, Zuiani C, Bazzocchi M, Furlan A, Londero V. Borderline breast lesions diagnosed at core needle biopsy: can magnetic resonance mammography rule out associated malignancy? Preliminary results based on 79 surgically excised lesions. Breast 2008; 17: Pediconi F, Padula S, Dominelli V, et al. Role of breast MR imaging for predicting malignancy of histologically borderline lesions diagnosed at core needle biopsy: prospective evaluation. Radiology 2010; 257: Ikeda DM, Hylton NM, Kuhl CK, et al. BI-RADS: magnetic resonance imaging, 1st ed. In: D Orsi CJ, Mendelson EB, Ikeda DM, et al. Breast Imaging Reporting and Data System: ACR BI-RADS breast imaging atlas. Reston, VA: American College of Radiology, Baum F, Fischer U, Vosshenrich R, Grabbe E. Classification of hypervascularized lesions in CE MR imaging of the breast. Eur Radiol 2002; 12: Strigel RM, Eby PR, Demartini WB, et al. Fre- 278 AJR:198, February 2012

8 MRI of High-Risk Breast Lesions quency, upgrade rates, and characteristics of highrisk lesions initially identified with breast MRI. AJR 2010; 195: Langlotz CP. Fundamental measures of diagnostic examination performance: usefulness for clinical decision making and research. Radiology 2003; 228: 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 1999; 210: Neubauer H, Li M, Kuehne-Heid R, Schneider A, Kaiser WA. High grade and non-high grade ductal carcinoma in situ on dynamic MR mammography: characteristic findings for signal increase and morphological pattern of enhancement. Br J Radiol 2003; 76: Kuhl CK, Schrading S, Bieling HB, et al. MRI for diagnosis of pure ductal carcinoma in situ: a prospective observational study. Lancet 2007; 370: Sanders ME, Schuyler PA, Dupont WD, Page D. The natural history of low-grade ductal carcinoma in situ of the breast in women treated by biopsy only revealed over 30 years of long-term follow-up. Cancer 2005; 103: Helvie MA, Hessler C, Frank TS, Ikeda DM. Atypical hyperplasia of the breast: mammographic appearance and histologic correlation. Radiology 1991; 179: Stomper PC, Cholewinski SP, Penetrante RB, Harlos JP, Tsangaris TN. Atypical hyperplasia: frequency and mammographic and pathologic relationships in excisional biopsies guided with mammography and clinical examination. Radiology 1993; 189: Eby PR, Ochsner JE, DeMartini WB, Allison KH, Peacock S, Lehman CD. Is surgical excision necessary for focal atypical ductal hyperplasia found at stereotactic vacuum- breast biopsy? Ann Surg Oncol 2008; 15: Schnitt SJ, Connolly JL, Tavassoli FA, et al. Interobserver reproducibility in the diagnosis of ductal proliferative breast lesions using standardized criteria. Am J Surg Pathol 1992; 16: Kuhl C. The current status of breast MR imaging. Part I. Choice of technique, image interpretation, diagnostic accuracy, and transfer to clinical practice. Radiology 2007; 244: FOR YOUR INFORMATION This article has been selected for the new AJR Journal Club activity. The accompanying Journal Club study guide can be found on the following page. AJR:198, February

9 Linda et al. APPENDIX 1: AJR Journal Club Study Guide: Nonsurgical Management of High-Risk Lesions Diagnosed at Core Needle Biopsy: Can Malignancy Be Ruled Out Safely With Breast MRI? Joseph J. Budovec, Margaret Mulligan, Alan Mautz Medical College of Wisconsin, Milwaukee, WI Introduction 1. Is the research question clinically relevant? Is the topic timely? 2. What is the standard practice at your institution for management of high-risk lesions diagnosed at core needle biopsy? 3. What is the research question being asked? Is a specific hypothesis formulated? How would you write the null and alternative hypotheses? Methods 4. How were patients selected for inclusion in this study? What were the exclusion criteria? 5. In general, what are the advantages of a prospectively designed study? What are the disadvantages? 6. How did the authors attempt to limit potential biases? 7. What was the reference standard to which the imaging results were compared? Results 8. Fifty-five patients were excluded from the study because they declined surgery. How might this influence the authors results? 9. How are positive predictive value and negative predictive value calculated? What is the clinical utility of a high negative predictive value? 10. The authors noted that papillomas and radial scars diagnosed at core needle biopsy had a high negative predictive value. What may be a reason for such findings? Physics 11. Briefly explain how dynamic signal intensity-time curves are created. What is the clinical utility of such curves? Discussion 12. What are the study limitations? How did the authors address the study limitations? How would you design a similar study to overcome these limitations? Background Reading 1. Georgian-Smith D, Lawton TJ. Controversies on the management of high-risk lesions at core biopsy from a radiology/pathology perspective. Radiol Clin North Am 2010; 48: Pediconi F, Padula S, Dominelli V, et al. Role of breast MR imaging for predicting malignancy of histologically borderline lesions diagnosed at core needle biopsy: prospective evaluation. Radiology 2010; 257: FOR YOUR INFORMATION For more information on Journal Clubs, see Evidence-Based Radiology A Primer in Reading Scientific Articles in the July 2010 AJR at *Please note that the authors of the Study Guide are distinct from those of the companion article. 280 AJR:198, February 2012