Rate of Malignancy in MRI-Detected Probably Benign (BI-RADS 3) Lesions

Transcription

1 Women s Imaging Original Research Spick et al. Malignancy in MRI BI-RADS 3 Lesions Women s Imaging Original Research Claudio Spick 1,2 Dieter H. M. Szolar 1 Pascal A. Baltzer 2 Manfred Tillich 1 Pia Reittner 1 Klaus W. Preidler 1 Katja Pinker-Domenig 2 Thomas H. Helbich 2 Spick C, Szolar DHM, Baltzer PA, et al. Keywords: BI-RADS 3, breast cancer, characteristics, MRI, probably benign DOI: /AJR Received March 20, 2013; accepted after revision July 11, Diagnostikum Graz-Süd West, Weblinger Guertel 25, 8054 Graz, Austria. Address correspondence to D. H. M. Szolar (dieter.szolar@diagnostikum-graz.at). 2 Department of Biomedical Imaging and Image-Guided Therapy, Division of Molecular and Gender Imaging, Medical University of Vienna (AKH), Vienna, Austria. AJR 2014; 202: X/14/ American Roentgen Ray Society Rate of Malignancy in MRI-Detected Probably Benign (BI-RADS 3) Lesions OBJECTIVE. The purpose of this study was to evaluate the malignancy rate in MRIdetected probably benign (BI-RADS 3) lesions in women without a history of breast cancer. MATERIALS AND METHODS. In this study, 1265 patients underwent breast MRI during a 7-year period. One hundred and eight (8.5%) patients with a nonpalpable breast lesion classified as BI-RADS 3 at MRI and with a needle biopsy or adequate follow-up of at least 24 months were included. Statistical analysis included calculation of the negative predictive value with its 95% CI. RESULTS. Of 108 lesions, 107 (99.1%) were correctly assessed as probably benign, resulting in a negative predictive value of 99.1% (95% CI, %). Histopathology was requested by the patient or referring physician in 44 patients. Of these, 43 (39.8%) lesions were classified as benign and one (0.9%) as malignant. There were no changes evident in any of the remaining 64 (59.2%) lesions during follow-up (range, 2 9 years). CONCLUSION. In MRI-detected probably benign (BI-RADS 3) lesions, the malignancy rate is low and within the accepted cancer rate for mammographically or sonographically detected BI-RADS 3 lesions. Short-term follow-up MRI at intervals of 6, 12, and 24 months in MRI BI-RADS 3 lesions remains a strong tool with which to detect suspicious lesions. Interval changes in size, morphology, or enhancement are regarded as indicative of malignancy. T he category BI-RADS 3 is defined as probably benign. Rather than biopsy, such lesions undergo short-term follow-up by an appropriate methodology to exclude malignancy. Several studies have established the validity of short-term imaging surveillance for BI-RADS 3 breast lesions detected at mammography. Data from various institutions in several countries have shown a small likelihood of malignancy for BI-RADS 3 breast lesions, ranging from 0.3% to 1.7%. A malignancy rate of less than 2% is within the acceptable range established by the imaging community [1 7]. Moreover, malignant breast lesions that are initially assessed as BI-RADS 3 are reliably and promptly identified by an interval change at short-term imaging follow-up when they are still at an early stage and have a favorable prognosis. MRI of the breast is increasingly used as a complementary diagnostic modality [8 13]. To standardize the reporting of MRI of the breast and to minimize false-positive results without compromising sensitivity, simi- lar to the mammographic BI-RADS lexicon, the American College of Radiology (ACR) introduced the 2003 edition of the MRI BI- RADS lexicon [14]. Although all MRI BI- RADS categories are well established, more data and additional investigation are required for the BI-RADS 3 category to establish the specific imaging features of BI- RADS 3 breast lesions detected with MRI. To date, only a few articles have assessed the value of MRI in BI-RADS 3 breast lesions, showing cancer rates from 0.6 to 10% [15 21]. The purpose of our study was to evaluate the rate of malignancy in MRI-detected BI- RADS 3 lesions in women without a history of breast cancer. Materials and Methods Case Selection This retrospective HIPAA-compliant study was conducted after an institutional review board waiver of authorization. Written informed consent for MRI of the breast was obtained from all patients. Our prospectively populated institutional database was searched for women with nonpalpable breast lesions 684 AJR:202, March 2014

2 Malignancy in MRI BI-RADS 3 Lesions that were classified as probably benign (BI-RADS 3) at MRI and in whom needle biopsy or adequate follow-up of at least 24 months was subsequently performed. Indications for the examination were in congruence with the ACR recommendations. Patients with a history of breast cancer and patients at high risk for breast cancer were not included in this study. Case selection was based on the original written report. No additional imaging review was performed. During a 7-year period (May 2002 to May 2009), 1265 patients (mean age ± SD, 50 ± 13 years; age range, years) underwent MRI of the breast. Of this group, 108 (8.5%) patients with a probably benign lesion met our inclusion criteria. No patient was lost during follow-up. MRI MRI was performed using a dedicated breast coil and a 1.5-T system (Magnetom Symphony, Siemens Healthcare). The examination was performed in the second week of the menstrual cycle in premenopausal women. Menopausal women under hormone replacement therapy were examined 1 month after discontinuation of treatment [11]. The imaging protocol was as follows: axial T2-weighted turbo inversion recovery magnitude sequence (TR/TE. 4810/80; matrix. 256/75; FOV, 330 mm; and slice thickness, 3 mm), axial T2-weighted turbo spin-echo without fat suppression sequence (TR/TE, 3470/99; matrix, 512/70; FOV, 330; and slice thickness, 3 mm), and axial T1-weighted 3D FLASH sequence (TR/TE, 5.13/2.04; flip angle, 12 ; matrix, 512/66; FOV, 300 mm; and slice thickness, 2 mm). T1-weighted images were obtained before and five times (within 6 minutes) after IV injection of 0.1 mmol/kg of gadopentetate dimeglumine (Omniscan, GE Healthcare). Subtracted images were obtained by subtracting unenhanced images from contrast-enhanced images using the manufacturer s commercially available software. MR Image Interpretation All imaging data were read in routine clinical practice. Results were entered into a prospectively populated database within our institutional electronic information system. MRI reading was performed by one of three radiologists, each with more than 10 years of experience in breast MRI. The readers interpreted the MR images using the ACR BI-RADS lexicon [14, 22]. No additional imaging review was performed for this retrospective analysis. Criteria for distinguishing between benign and malignant contrast-enhancing lesions were based on both morphology and enhancement kinetics in accordance with the BI-RADS lexicon. For analysis of lesion enhancement kinetics, regions of interest (ROIs) were manually drawn in the most enhancing part of a lesion, and the intensity course was plotted against time. The percentage of signal intensity (SI) increase was defined as SI = [(SI post SI pre ) / SI pre ] 100, where SI post is SI after contrast enhancement and SI pre is SI before contrast enhancement. SI time curves were calculated and classified as type I (showing persistent enhancement increasing with time, frequently indicative of a benign lesion), type II (showing a plateau phase after maximum SI was reached 2 to 3 minutes after contrast media injection, frequently indicative of a borderline lesion), or type III (showing an early increase in SI followed by rapid washout of the contrast agent, usually indicative of malignancy) [14]. Applying the descriptors defined by BI-RADS, the following interpretation scheme was used to categorize the probability of malignancy for each lesion [14, 23]. BI-RADS category 1 was assigned if there was no enhancement. BI-RADS category 2 was assigned if the lesion shape (round or oval) and margins (smooth or lobulated) were suggestive of a benign lesion and if the internal architecture showed low-si internal septations (fibroadenoma) regardless of the enhancement kinetics of the lesion. BI-RADS category 2 was also assigned if the internal architecture was heterogeneous or slightly heterogeneous and if the shape (round or oval), margins (smooth or lobulated), and time course kinetics suggested a benign lesion (type I curve, persistent enhancement and slow enhancement rate). BI-RADS category 3 was assigned to a newly diagnosed lesion with a round or oval shape, smooth or lobulated margins, and if the lesion showed a type I or type II curve (fast enhancement rate of > 80% and a plateau phase). There is some overlap in the lesion characteristics between the BI-RADS 2 and 3 categories. In such cases, each radiologist was free to assess lesions as benign or probably benign as appropriate. BI-RADS category 4 was assigned if, in the same setting, a type III (fast enhancement rate and washout) curve was observed. BI-RADS category 4 was also assigned in lesions with a suspicious shape (irregular) and margins (irregular or spiculated) but benign-appearing kinetics. BI-RADS category 5 was assigned if the shape (irregular), margins (irregular or spiculated), and time-course kinetics indicated a malignant lesion. Nonmasslike enhancement was categorized as an area of enhancement 5 mm. For lesions with nonmasslike enhancement, BI-RADS category 2 was assigned if the spatial distribution was diffuse and the internal enhancement pattern was either stippled or homogeneous. BI-RADS category 3 was assigned if the spatial distribution was either focal or regional and the internal enhancement pattern was either stippled or homogeneous. BI- RADS category 4 was assigned if the spatial distribution was segmental, ductal, or linear, regardless of the internal enhancement pattern. BI-RADS category 4 was assigned if the internal enhancement pattern was clumped, regardless of the spatial distribution. BI-RADS category 5 was assigned if both the spatial distribution and the internal enhancement pattern were suspicious. Patient Management Our short interval follow-up protocol consisted of MRI of the breast at 6, 12, and 24 months. If the full set of follow-up examinations documented stability, the patient was advised to resume routine yearly mammography screening. Cases in which the patient had a benign needle biopsy result or imaging follow-up without newly diagnosed suspicious lesions were regarded benign for statistical analysis. Any change in the size of a mass, morphology, or in nonmasslike enhancement was regarded as a substantial change in lesion appearance and therefore suspicious. In these cases, biopsy or rebiopsy was recommended [8, 14]. In accordance with international guidelines, it was not our policy to recommend any histologic TABLE 1: Histologic Results of 44 Probably Benign (MRI BI-RADS 3) Breast Lesions Undergoing Invasive Tissue Diagnosis Histologic Diagnosis No. of Findings Benign lesions Fibroadenoma 11 (25) Fibrocystic changes 7 (15.9) Fibrosis 3 (6.8) Benign breast parenchyma 2 (4.5) Scar tissue 1 (2.3) Mastitis 2 (4.5) Duct hyperplasia 2 (4.5) Sclerosing adenosis 5 (11.5) Fatty tissue 1 (2.3) Lymph node 1 (2.3) Microcalcification 2 (4.5) Cyst 1 (2.3) Hamartoma 0 Papilloma 2 (4.5) High-risk lesions Atypical ductal hyperplasia 2 (4.5) Phyllodes tumor 0 Papilloma with atypia 1 (2.3) Malignant lesions Ductal carcinoma in situ 1 (2.3) Total 44 (100) Note Data in parentheses are percentages. AJR:202, March

3 Spick et al. workup (needle breast biopsy or even open surgical breast biopsy) in nonpalpable BI-RADS 3 breast lesions. Rather, we recommended periodic imaging surveillance [14]. However, 44 (40.7%) women underwent tissue diagnosis on the basis of the patient s or referring physician s preference. After needle biopsy, histologic results were correlated with imaging findings, and specific recommendations were made to the patient and the referring physician. Surgery was recommended in cases in which needle biopsy results were discordant with the imaging features or if surgical excision was suggested by the pathologist. Data Collection and Analysis A malignancy was defined as a lesion that yielded invasive carcinoma or ductal carcinoma in situ at needle breast biopsy, open surgical biopsy, or both and confirmed at subsequent pathology review. Data were entered into a computerized spreadsheet (Microsoft Excel). Statistical analyses were performed A with statistical software (SPSS Statistics, IBM). Descriptive statistical metrics were calculated. Exact 95% CIs were assessed using the Wilson method. Fig year-old woman with BI-RADS 3 lesion. A D, Axial unenhanced (A) and subtracted 2 minute (B) and 5 minute (C) T1-weighted dynamic fat-suppressed MR images of breast show smooth oval mass (diameter, cm) in lower outer quadrant of right breast with plateau (type II) curve type (D). Lesion was classified as BI-RADS 3. Follow-up MRI of breast 6 months later (not shown) did not show any change in size. B Results In this study, 107 of 108 (99.1%) lesions were correctly assessed as probably benign (BI-RADS 3) when compared with the reference standard (stable at follow-up or benign biopsy results). As a consequence, the negative predictive value was 99.1% (95% CI, %). The final histologic diagnosis of 44 lesions (40.7%) on the basis of subsequent surgical excision or follow-up after needle breast biopsy is shown in Table 1. Histopathology classified 43 (39.8%) lesions as benign and one (0.9%) as malignant. In the remaining 64 lesions (59.2%), imaging follow-up was available for at least 2 years (mean, 3.2 years; range, 2 9 years). There were no changes in imaging findings in any of these 64 (59.2%) breast lesions during follow-up. Malignancy was found in one (0.9%) of 108 lesions (95% CI, %) (Figs. 1 and 2). On initial breast MRI, this lesion was depicted as an oval mass with smooth margins (diameter, cm) with early homogeneous contrast enhancement and a plateau (type II) curve type. The lesion was classified as BI-RADS 3. At 23 months after the initial study, the lesion showed an increase in size ( cm) and shape and appeared as a round mass with irregular borders and had a washout (type III) curve type. Histology revealed this lesion as invasive ductal carcinoma. The imaging characteristics of BI-RADS 3 lesions are listed in Table 2. Discussion The results of our study show that the malignancy rate in BI-RADS 3 lesions at MRI Signal Intensity Relative Enhancement (%) No. of Time Points After Contrast Material Administration C D 686 AJR:202, March 2014

4 Malignancy in MRI BI-RADS 3 Lesions Signal Intensity Relative Enhancement (%) No. of Time Points After Contrast Material Administration is in the accepted range for malignancy detected by mammography and ultrasound. Furthermore, our malignancy rate of 0.9% is in accordance with the previously reported 0.2 2% malignancy rates for mammography and ultrasound [1 7, 24, 25]. BI-RADS 3 assessment is used to characterize the morphologic and kinetic imaging features that would suggest that the breast lesion is probably benign and the likelihood of malignancy is less than 2%. Unlike BI- RADS 3 mammographic and ultrasound lesions, BI-RADS 3 MRI lesions have not been well studied. Thus, more data and additional investigation are required for MRI-detected BI-RADS 3 lesions to establish specific imaging feature guidelines. To date, only a few studies have investigated the frequency of malignancy solely in MR-detected BI- RADS 3 lesions, which showed malignancy rates from 0.6% to 10% [15 21]. A D Table 3 shows the malignancy rate in MRIdetected BI-RADS 3 lesions in the published literature. As can be seen, the majority of cases were correctly assessed as BI-RADS 3. The malignancy rate ranged from 0.6% to 10%. The findings of the current study are in good agreement with recently published data on the frequency of malignancy in MRI of BI-RADS 3 lesions. Weinstein et al. [20] reported a frequency of malignancy of 0.9% in 106 BI-RADS 3 lesions in the contralateral breast of women with known breast cancer. Marshall et al. [21] assessed 132 BI-RADS 3 lesions in a high-risk population and found one (0.8%) malignancy. Compared with these two studies, we found a similar cancer rate of 0.9%. However, we investigated only patients without a history of breast cancer, and therefore our results are more applicable to the general population. Another strength of our study is that it offers histopathologic B Fig year-old woman with BI-RADS 3 lesion (same patient as in Fig. 1). A D, At 23 months after initial study, unenhanced (A) and subtracted 2 minute (B) and 5 minute (C) axial T1- weighted dynamic fat-suppressed MR images of breast show increase in size of lesion ( cm), irregular border, and washout (type III) curve type (D). Histology (not shown) revealed invasive ductal carcinoma. proof of 43 (39.8%) lesions correctly classified as BI-RADS 3 at MRI. In the remaining 64 (59.2%), the stability of the lesion s imaging characteristics was verified during follow-up ranging from 2 to 9 years. Studies by Liberman et al. [15] and Sadowski et al. [16] have shown a cancer rate of 6% and 10% within groups of probably benign lesions. However, these studies were conducted before the ACR MRI BI-RADS lexicon was introduced, and we agree with Weinstein et al. [20] that this might have triggered a higher frequency of malignancy in the MRI-diagnosed probably benign breast lesions. In addition, over time, there have been significant improvements both in MRI software and hardware, which might explain the more accurate diagnosis and characterization of MRI-detected breast lesions [20]. Previous studies [18 21] have shown a BI- RADS 3 category assignment in line with our C AJR:202, March

5 results and underline the importance of including specific imaging features for BI- RADS 3 lesions at MRI in a new version of the BI-RADS lexicon. The BI-RADS 3 category should be assigned to a newly diagnosed mass with a round or oval shape, smooth or lobulated margins, and a type I or type II curve (fast enhancement rate > 80% and a plateau phase). In nonmasslike enhancement, the BI-RADS category 3 may be assigned if the spatial distribution is either focal or regional and the internal enhancement pattern is either stippled or homogeneous. The curve type in such lesions may be type I or II. Any interval changes in size or morphology during follow-up as well as changes in the enhancement curve type (type III) are regarded as indicative of malignancy. Applying this approach, malignant lesions were diagnosed because of morphologic changes during follow-up at 18 months and 11 months in two previous studies (16, 19). On the other hand, the cancers described by Weinstein et al. [20] and Marshall et al. [21] were revealed because of prophylactic mastectomy. Thus, no lesions were detected during follow-up. Therefore, short-term follow-up MRI in lesions classified as probably benign remains a good way to detect suspicious lesions. Follow-up at 0, 6, 12, and 24 months is used in conventional imaging and is the current strategy used in breast MRI. In addition to the malignancy rate in BI- RADS 3 lesions, the percentage of BI-RADS 3 lesions in relation to the total number of patients evaluated at MRI seems to play an important role. In our study, the BI-RADS 3 category was assigned in 108 (8.5%) of 1265 patients. Our assignment rate is in accordance with results reported by previous breast imaging studies [1 5, 7, 18, 20, 21, 26]. In these studies, the rate of diagnosing a BI-RADS 3 lesion ranged from 3% to 10%. Overall, it seems that in recent MRI studies of the breast the malignancy rate and frequency of BI-RADS 3 assignment closely approximates the results of previously published mammographic and sonographic studies [1 5, 24, 25]. The current study has some limitations. One limitation is its retrospective character. That means additional and follow-up examinations (e.g., second-look ultrasound or follow-up mammography and ultrasound) were not necessarily performed at our institution. However, this practice may better reflect the real clinical situation compared with a strict single-center study. All imaging data were Spick et al. TABLE 2: Morphology and Contrast Enhancement Patterns of All 108 Probably Benign (MRI BI-RADS 3) Breast Lesions Including 44 Histologically Verified Lesions Characteristic Total No. of Benign Lesions Total No. of Malignant Lesions Morphology Mass 68 (62.9) 1 (0.9) Round or oval 52 (48.1) 1 (0.9) Lobulated 14 (12.9) 0 Irregular 2 (1.9) 0 Nonmasslike enhancement 40 (37.0) 0 Regional 21 (19.4) 0 Focal 19 (17.4) 0 Internal enhancement Homogeneous 91 (84.3) 0 Heterogeneous 17 (15.7) 1 (0.9) Rim enhancement 0 0 Curve types Persistent 66 (61.1) 0 Plateau 42 (38.9) 1 (0.9) Washout 0 0 Total 107 (99.1) 1 (0.9) Note Data in parentheses are percentages. read in routine clinical practice, and results were collected from original written reports entered into a prospectively populated database. The breast MRI studies of all 1265 patients in our study were read independently by three radiologists, each with more than 10 years of experience in breast MRI, and followed a standardized imaging protocol as outlined in the Materials and Methods section. Therefore, the results reflect the interobserver variability in breast MRI interpretation and a uniform quality of breast MRI studies. However, because of the lack of descriptors for probably benign lesions, the readers may have falsely placed lesions into the BI-RADS 3 category. In conclusion, in this study breast MRI showed a high negative predictive value of 99.1% in lesions classified as BI-RADS 3. Malignancy was found in one (0.9%) of 108 lesions, which is within the established acceptable rate for mammographically and sonographically detected BI-RADS 3 lesions. Short-term follow-up MRI at an interval of 6, 12, and 24 months in lesions classified as BI-RADS 3 remains a strong tool with which to detect suspicious lesions. Any interval changes in size, morphology, or en- TABLE 3: Malignancy Rate in MRI-Detected Probably Benign (BI-RADS 3) Lesions Compared With Literature Study Year of Publication No. of Probably Benign Lesions Total No. of Cancers Malignancy Rate (%) Liberman et al. [15] Sadowski and Kelcz [16] Eby et al. [18] Hauth et al. [19] Weinstein et al. [20] Marshall et al. [21] Current study Total malignancy range AJR:202, March 2014

6 Malignancy in MRI BI-RADS 3 Lesions hancement are regarded as indicative of malignancy. Further studies may show a different follow-up approach for MRI-detected BI-RADS 3 lesions. References 1. Sickles EA. Periodic mammographic follow-up of probably benign lesions: results in 3,184 consecutive cases. Radiology 1991; 179: Helvie MA, Pennes DR, Rebner M, Adler DD. Mammographic follow-up of low-suspicion lesions: compliance rate and diagnostic yield. Radiology 1991; 178: Varas X, Leborgne F, Leborgne JH. Nonpalpable, probably benign lesions: role of follow-up mammography. Radiology 1992; 184: Varas X, Leborgne JH, Leborgne F, Mezzera J, Jaumandreu S. Revisiting the mammographic follow-up of BI-RADS category 3 lesions. AJR 2002; 179: Vizcaíno I, Gadea L, Andreo L, et al. Short-term follow-up results in 795 nonpalpable probably benign lesions detected at screening mammography. Radiology 2001; 219: Wolfe JN, Buck KA, Salane M, Parekh NJ. Xeroradiography of the breast: overview of 21,057 consecutive cases. Radiology 1987; 165: Gruber R, Jaromi S, Rudas M, et al. Histologic work-up of non-palpable breast lesions classified as probably benign at initial mammography and/ or ultrasound (BI-RADS category 3). Eur J Radiol 2013; 82: Sardanelli F, Boetes C, Borisch B, et al. Magnetic resonance imaging of the breast: recommendations from the EUSOMA working group. Eur J Cancer 2010; 46: Helbich TH. Contrast-enhanced magnetic resonance imaging of the breast. Eur J Radiol 2000; 34: 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: Mann RM, Kuhl CK, Kinkel K, Boetes C. Breast MRI: guidelines from the European Society of Breast Imaging. Eur Radiol 2008; 18: Morris EA. Review of breast MRI: indications and limitations. Semin Roentgenol 2001; 36: Pinker K, Grabner G, Bogner W, et al. A combined high temporal and high spatial resolution 3 Tesla MR imaging protocol for the assessment of breast lesions: initial results. Invest Radiol 2009; 44: American College of Radiology. Breast Imaging Reporting and Data System (BI-RADS), 4th ed. Reston, VA: American College of Radiology, Liberman L, Morris EA, Benton CL, et al. Probably benign lesions at breast magnetic resonance imaging: preliminary experience in high-risk women. Cancer 2003; 98: Sadowski EA, Kelcz F. Frequency of malignancy in lesions classified as probably benign after dynamic contrast-enhanced breast MRI examination. J Magn Reson Imaging 2005; 21: Gökalp G, Topal U. MR imaging in probably benign lesions (BI-RADS category 3) of the breast. Eur J Radiol 2006; 57: Eby PR, DeMartini WB, Peacock S, et al. Cancer yield of probably benign breast MR examinations. J Magn Reson Imaging 2007; 26: Hauth E, Umutlu L, Kummel S, et al. Follow-up of probably benign lesions (BI-RADS 3 category) in breast MR imaging. Breast J 2010; 16: Weinstein SP, Hanna LG, Gatsonis C, et al. Frequency of malignancy seen in probably benign lesions at contrast-enhanced breast MR imaging: findings from ACRIN Radiology 2010; 255: Marshall AL, Domchek SM, Weinstein SP. Follow-up frequency and compliance in women with probably benign findings on breast magnetic resonance imaging. Acad Radiol 2012; 19: Ikeda DM, Hylton NM, Kinkel K, et al. Development, standardization, and testing of a lexicon for reporting contrast-enhanced breast magnetic resonance imaging studies. J Magn Reson Imaging 2001; 13: Kuhl CK, Schild HH, Morakkabati N. Dynamic bilateral contrast-enhanced MR imaging of the breast: trade-off between spatial and temporal resolution. Radiology 2005; 236: Graf O, Helbich TH, Hopf G, Graf C, Sickles EA. Probably benign breast masses at US: is follow-up an acceptable alternative to biopsy? Radiology 2007; 244: Graf O, Helbich TH, Fuchsjaeger MH, et al. Follow-up of palpable circumscribed noncalcified solid breast masses at mammography and US: can biopsy be averted? Radiology 2004; 233: Kuhl CK, Schmutzler RK, Leutner CC, et al. Breast MR imaging screening in 192 women proved or suspected to be carriers of a breast cancer susceptibility gene: preliminary results. Radiology 2000; 215: AJR:202, March