Solitary Dilated Duct Identified at Mammography: Outcomes Analysis

Transcription

1 Women s Imaging Original Research Mammography of Solitary Dilated Duct Chang et al. Women s Imaging Original Research FOCUS ON: C. Belinda Chang 1 Natalya M. Lvoff 2 Jessica W. Leung 3 R. James Brenner 1,4 Bonnie N. Joe 1 Hilda H. Tso 1 Edward A. Sickles 1 Chang CB, Lvoff NM, Leung JW, et al. Keywords: BI-RADS, ductal carcinoma in situ, mammography, solitary dilated duct DOI: /AJR Received April 21, 2009; accepted after revision September 22, Supported by a National Cancer Institute Breast Cancer Surveillance Consortium cooperative agreement (U01CA3740). The collection of cancer incidence data used in this study was supported in part by the California Department of Public Health under Health and Safety Code Section ; the National Cancer Institute Surveillance, Epidemiology, and End Results Program under contract N01-PC awarded to the Northern California Cancer Center, contract N0-PC awarded to the University of Southern California, and contract N02-PC awarded to the Public Health Institute; and the Centers for Disease Control and Prevention Nation Program of Cancer Registries under agreement U55/CCR awarded to the Public Health Institute. The ideas and opinions expressed herein are those of the authors and endorsement by the State of California, Department of Public Health; the National Cancer Institute; and the Centers for Disease Control and Prevention or their contractors or subcontractors is not intended and should not be inferred. 1 Department of Radiology and Biomedical Imaging, University of California, San Francisco, Mount Zion Medical Center, 1600 Divisadero St., Rm. C-250, Box 1667, San Francisco, CA Address correspondence to C. B. Chang (belinda.chang@radiology.ucsf.edu). 2 Breast Diagnostic Center of Marin, Greenbrae, CA. 3 Breast Health Center, California Pacific Medical Center, San Francisco, CA. 4 Bay Imaging Consultants, Carol Ann Read Breast Health Center, Oakland, CA. AJR 2010; 194: X/10/ American Roentgen Ray Society WOMEN S IMAGING Solitary Dilated Duct Identified at Mammography: Outcomes Analysis OBJECTIVE. The purpose of this study is to review the clinical and pathologic outcomes for cases of solitary dilated duct identified at mammography. MATERIALS AND METHODS. For all screening mammography examinations during a 22-year period and all diagnostic mammography examinations during the last 10 of these years, the radiologists recorded the principal finding of each abnormal mammographic examination during image interpretation. Only examinations with the recorded finding of solitary dilated duct were studied. We examined radiology records to determine imaging followup and pathology records to determine histologic diagnosis, and we performed linkage with our regional tumor registry to identify cancers not biopsied at our institution. RESULTS. The finding of solitary dilated duct was recorded for nine (0.0038%) of 235,209 consecutive screenings and for 12 (0.041%) of 29,267 consecutive diagnostic mammography examinations. Five screening and five diagnostic cases were stable at follow-up (minimum interval, 2 years) and did not undergo biopsy; tumor registry linkage showed no subsequent cancer diagnosis. Biopsy was performed for four (44%) of nine screening and seven (58%) of 12 diagnostic cases. One cancer each (ductal carcinoma in situ) was identified from the screening and diagnostic populations, yielding positive predictive values of 11% (1/9) and 8% (1/12), respectively. CONCLUSION. Solitary dilated duct is a rare mammographic finding, this series being the largest reported to date. Although few cases are studied, solitary dilated duct appears to have a greater than 2% likelihood of malignancy, sufficiently high to suggest that a suspicious (BI-RADS 4a) assessment may be appropriate. T he mammographic finding of solitary dilated duct is rare and poorly understood. There are anecdotal reports of solitary dilated duct as the only mammographic finding of underlying malignancy [1 10], indicating its potential importance in the early detection of breast cancer. However, some investigators have estimated that the finding of solitary dilated duct has a very low risk of malignancy [3, 9], supporting its assessment as a benign (BI-RADS category 2) or probably benign (BI-RADS category 3) lesion [11]. Solitary dilated duct also has been reported to coexist with more suspicious mammographic findings [6, 10], but in such cases the associated mass, grouped microcalcifications, architectural distortion, or developing asymmetry would itself have a sufficiently high likelihood of malignancy to prompt a suspicious (BI-RADS category 4) assessment. Solitary dilated duct is described and illustrated in the current edition of the BI-RADS atlas as the first of four mammographic findings classified as special cases [12]. The accompanying text states that if unassociated with other suspicious clinical or mammographic findings, it is usually of minor clinical significance [12]. Insofar as this statement is made under the imprimatur of the widely read BI-RADS atlas, it is likely to influence those practicing radiologists without much, if any, personal experience who encounter the rare finding of solitary dilated duct. However, to our knowledge, to date there is no large clinical series indicating the positive predictive value for malignancy of solitary dilated duct. The goal of this largescale study is to report the clinical and pathologic outcomes for the isolated finding of solitary dilated duct identified at screening or diagnostic mammography. 378 AJR:194, February 2010

2 Mammography of Solitary Dilated Duct Materials and Methods This was a retrospective single-institution study performed at a large academic center. The study was approved by our institutional review board, which waived requirements for written informed consent. We conducted a retrospective review of all 235,209 consecutive screening examinations performed from January 1, 1985, through December 31, 2006, and all 29,267 consecutive diagnostic mammographic examinations performed from January 1, 1997, through December 31, 2006, to identify those abnormal examinations reported to have the mammographic finding of solitary dilated duct, including both symptomatic and asymptomatic patients. Symptoms were identified through the referring clinician s written requisition and through questionnaire forms filled out by patients on the day of the examination. Patients also reported family history of breast cancer in the questionnaire. From study inception in 1985, the standard definition of a tubular or branching structure was used by our interpreting radiologists to identify solitary dilated duct cases [6], a definition that subsequently was adopted in BI-RADS [12]. From study inception, we also used a set of mammography assessment categories [13] that mapped precisely to the assessment categories subsequently developed for BI-RADS, to which we switched when the BI- RADS categories were published. Therefore, although the beginning of our study predates publication of the first edition of BI-RADS by several years, all identifications of solitary dilated duct cases and all study assessments conform to BI- RADS content. From 1985 to 2005 all mammography examinations were performed using film-screen recording systems. In 2006, the last year of study accrual, we began a transition to full-field digital equipment, so only a small percentage (< 5%) of study examinations involved digital mammography. At the time of interpretation, the mammographic assessment was prospectively entered by the radiologist into a computer database used for report generation, quality assurance, and clinical research [13]. The radiologist also entered the principal mammographic finding into the database for all abnormal assessments (BI-RADS category 0, 4, 5) [13]. Most examinations were interpreted by breast imaging specialists; some general radiologists read small numbers of cases. The characteristics of our mammography facilities, patient population, and interpreting radiologists have been reported previously [14, 15]. The study cohort was identified by searching our database for all examinations with the prospectively recorded principal finding of solitary dilated duct, without more suspicious associated mammographic findings. Because the text of all our mammography reports also is stored permanently in a commercial radiology reporting system (IDXrad, IDX Systems), we were able to perform a word search for dilated duct using reportingsystem software to validate our selection of solitary dilated duct cases. We excluded all diagnostic mammography examinations performed within 3 months after abnormal solitary dilated duct screening at our institution to avoid duplication of cases already identified at screening. This approach to case selection resulted in inclusion of all solitary dilated duct screening mammography examinations, whether or not they were interpreted as abnormal at diagnostic breast imaging. Furthermore, because it was standard procedure for the radiologists in our practice to assess as suspicious any solitary dilated duct case in a symptomatic woman, our study should include almost all solitary dilated duct diagnostic mammography examinations as well because almost all of our diagnostic mammography patients were symptomatic. We examined radiology department records to determine imaging follow-up and pathology records from our institution to determine the histologic diagnosis as well as the size, nodal status, and stage of all cancers. We also performed linkage with our regional tumor registry to identify any cancers not biopsied at our institution. A previous study has shown that linkage with our regional tumor registry identifies more than 94.3% of extant cancer cases [16]. For study purposes, cancer was defined as ductal carcinoma in situ (DCIS) or any type of invasive carcinoma. We stopped case accrual at the end of 2006, 2 years before ending the collection of follow-up data. This allowed a minimum of 1 year for subsequent cancer diagnosis and 1 more year for cancer cases to be entered into our regional tumor registry. All patients who did not undergo biopsy had at least 2 years of mammographic follow-up. The diagnostic workup of solitary dilated duct included spot-compression magnification mammography and (in most cases) targeted ultrasound examination to evaluate for associated calcifications, mass (outside the dilated duct), architectural distortion, and developing asymmetry. If any of these were present, the case was excluded from this study of solitary dilated duct as an isolated mammographic finding. Ultrasound was used to TABLE 1: Patient Characteristics indicate whether the solitary dilated duct seen at mammography contained fluid, debris, focal mural irregularity, or an intraluminal mass. Ductography was performed as clinically indicated. Breast MRI examination was available at our institution only for the last few years of study and even then was used only sparingly before cancer diagnosis. Recommendation for biopsy versus follow-up of solitary dilated duct cases was at the discretion of the radiologist interpreting the diagnostic breast imaging examination. Ultrasound-guided core biopsy was performed for cases in which there was the sonographic target of an intraluminal mass or focal mural irregularity. Preoperative mammographic wire localization was used for the remaining (surgical) biopsies. Descriptive statistics of means, range, and frequency were used to describe patient characteristics and most study outcomes. Positive predictive value was defined as the fraction of studied solitary dilated duct cases with subsequent diagnosis of breast cancer (true-positive cases) divided by all studied solitary dilated duct cases (true-positive cases plus false-positive cases) [17]. Results The principal finding of solitary dilated duct was recorded for 21 (0.0079%) of 264,476 consecutive mammography examinations. This involved nine (0.0038%) of 235,209 consecutive screening and 12 (0.041%) of 29,267 consecutive diagnostic mammography examinations. Patient demographics were similar for our study (21 cases) and overall (264,476 cases) populations (Table 1). For the study population, patient age ranged from 32 to 80 years (mean, 56.3 years). All of our solitary dilated duct screening mammography patients were asymptomatic. Indications for performing diagnostic mammography at our institution have been previously described [18]; all of our solitary dilated duct diagnostic mammography patients had nonspecific symptoms rather than a palpable lump or spontaneous nipple discharge: noncyclic unilateral breast tenderness (eight patients), unilateral pain (three patients), and unilateral breast enlargement (one patient). Characteristic All Patients Solitary Dilated Duct Patients Total 264, Average age (y) 55.9 (range, 18 99) 56.3 (range, 32 80) First-degree relative with breast cancer (%) 17.8 (47,112 patients) 19.0 (4 patients) Personal history of breast cancer (%) 8.9 (23,459 patients) 9.5 (2 patients) AJR:194, February

3 Chang et al. For 12 cases of solitary dilated duct, previous mammography examinations were not available for comparison (most involved baseline examinations); therefore, stability A C or interval change of the solitary dilated duct could not be assessed. Previous mammography examinations were available for comparison for the other nine solitary dilated duct B D Fig. 1 Solitary dilated duct identified at screening mammography in 64-year-old woman. Ultrasoundguided core biopsy revealed low grade ductal carcinoma in situ, 7 mm in greatest dimension. Although ultrasound is not examined in this article, this case illustrates sonographic finding of solitary dilated duct. A, Mediolateral oblique projection mammogram shows dilated retroareolar duct (arrow). B, Craniocaudal projection mammogram shows dilated retroareolar duct (arrow). C, Transverse sonogram shows dilated duct corresponding to A and B, with echogenic filling defect representing intraluminal mass (arrow). D, Transverse Doppler sonogram shows intraluminal mass (arrow) is highly vascular. cases, all of which showed the solitary dilated duct to be a new finding. Ten (48%) of the 21 solitary dilated duct cases (five screening, five diagnostic) did not undergo biopsy and were stable at follow-up (minimum interval, 2 years). Tumor registry linkage showed no subsequent cancer diagnosis for any of these cases. Biopsy was performed in 11 (52%) of the 21 cases: four (44%) of nine screening and seven (58%) of 12 diagnostic cases. Of the 11 biopsies (four screening, seven diagnostic), four were done as wire localizations using mammographic guidance (two screening, two diagnostic) and seven were done as ultrasound-guided core biopsy (two screening, five diagnostic). Two cases with malignant results were identified, one each from the screening (Fig. 1) and diagnostic populations, yielding positive predictive values of 11% (1/9) and 8% (1/12), respectively. The screening-detected malignancy had a maximum diameter of 7 mm with pathology showing DCIS, papillary and cribriform patterns, and grade I (low grade) using the Scarff-Bloom-Richardson histologic grading system [19]; the diagnostic-detected malignancy had a maximum diameter of 40 mm with pathology showing DCIS, micropapillary and cribriform types, and grade III (high grade). One screening-detected case yielded atypical ductal hyperplasia at surgical biopsy. The remaining eight biopsied cases (two screening, six diagnostic) had benign causes (Table 2) involving fibrocystic change associated with benign fluidfilled ectatic ducts. Targeted ultrasound was performed for 14 of the solitary dilated duct cases, including the two cases in which biopsy yielded DCIS. Ultrasound showed fixed focal abnormalities (intraluminal mass or focal mural irregularity) within the dilated duct in seven cases (including one DCIS case); ultrasound showed no focal abnormalities in the other 380 AJR:194, February 2010

4 Mammography of Solitary Dilated Duct TABLE 2: Solitary Dilated Duct Cases With Benign Pathology at Biopsy seven cases (including one DCIS case). For the seven cases undergoing biopsy for which there was a focal ultrasound abnormality, ultrasound-guided core biopsy targeted the focal abnormality (one DCIS, six proliferative fibrocystic changes with or without apocrine metaplasia). For the remaining four biopsies, performed using mammographic wire localization targeting the entire portion of the duct that was dilated at mammography, the extent of excision was determined intraoperatively by the surgeon and also may have included additional portions of the duct that were visibly abnormal. The contralateral breast was assessed as negative (BI-RADS category 1) or benign (BI-RADS category 2) in all cases. The mean interval between solitary dilated duct detection and biopsy was 1.5 months, range 0 days 3.5 months. Most patients (17, 81.0%) did not have a family history of breast cancer. Two (9.5%) had a first-degree relative diagnosed with breast cancer under the age of 50 years, and two (9.5%) had a first-degree relative diagnosed with breast cancer at age 50 years or older. One woman had a personal history of ipsilateral breast cancer, and one had a personal history of contralateral breast cancer. One of these two women with a personal history of breast cancer had a benign breast biopsy, and the other did not undergo biopsy and has had stable mammographic follow-up for 2 years. These two patients with a personal history of breast cancer represent 9.5% (2/21) of the study cohort, comparable with the 8.9% personal history rate of our overall patient population (Table 1). The women with biopsy-proven cancer were 47 and 64 years old (mean, 55.5 years) and did not have a family history or personal history of breast cancer. Pathology Screening Diagnostic Nonproliferative fibrocystic change 0 2 Proliferative fibrocystic change (no atypia or apocrine metaplasia) 1 3 Proliferative fibrocystic change with apocrine metaplasia 1 1 Proliferative fibrocystic change with atypical ductal hyperplasia (no malignancy at subsequent surgical excision) 1 0 Discussion The isolated finding of a solitary dilated duct identified at mammography is rare. We found only 21 (0.0079%) solitary dilated duct cases in 264,476 consecutive mammography examinations. Because of its infrequent occurrence, little is known about the clinical and pathologic significance of solitary dilated duct. Based on anecdotal reports, the current edition of the BI-RADS atlas states that uncomplicated cases of solitary dilated duct are usually of minor clinical significance [12], suggesting the appropriateness of benign (BI-RADS category 2) or probably benign (BI-RADS category 3) assessments. However, to our knowledge, no previous large-scale studies have been performed to indicate the positive predictive value of solitary dilated duct as an isolated mammographic finding. Moskowitz [3] reported on the positive predictive value of selected mammographic findings using data from a consecutive series of 40,431 screening examinations performed from 1973 to 1979 using xeroradiography. Among these cases, there were 26 described as having biopsy recommended due to duct dilatation, a mammographic finding that was neither defined nor illustrated. None of these 26 cases was found to be malignant. However, it is unlikely that the finding described as duct dilatation in the study by Moskowitz represents solitary dilated duct as we define it today because the frequency of 0.064% (26/40,431) for duct dilatation in that series was 17 times higher than the % screening frequency for solitary dilated duct reported in our series; although nipple retraction is now a much more frequent finding at mammography screening than is solitary dilated duct, Moskowitz reported duct dilatation to occur twice as frequently as nipple retraction; and all 26 cases of duct dilatation in that study were sufficiently abnormal to be recommended for biopsy compared with only four of the nine screening solitary dilated duct cases in our series. Huynh et al. [10] reported a series of 46 biopsied cases of unilateral or asymmetrically dilated ducts, 11 (24%) of which represented malignancy. However, the cases in that series were highly selected rather than consecutive, some (perhaps most) of the cases involved multiple ducts rather than a solitary dilated duct (numbers not specified), some of the cases were associated with suspicious microcalcifications, and all but one of the noncalcified lesions showed interval appearance or enlargement of the dilated ducts visible at mammography. Hence, that series also does not provide insight into the clinical and pathologic outcomes of solitary dilated duct as an isolated mammographic finding. Our study involves a consecutive series of all screening and almost all diagnostic mammography examinations showing uncomplicated solitary dilated duct. Our observation that 9.5% (2 of 21) of cases were malignant indicates a positive predictive value for solitary dilated duct that is substantially higher than the 2% ceiling defined in BI-RADS for probably benign (category 3) assessments. Although these data are limited by the rarity of solitary dilated duct as an isolated mammographic finding, our data suggest that it may be appropriate to assess all uncomplicated solitary dilated duct cases as suspicious (BI-RADS category 4a), accompanied by the recommendation for prompt tissue diagnosis. The imaging evaluation of all solitary dilated duct cases should involve spot-compression or spot-compression magnification mammography to evaluate for the presence of more suspicious associated findings, such as mass, grouped calcifications, architectural distortion, or developing asymmetry, any of which, if present, should represent an additional target for biopsy. Ultrasound examination also should be performed to identify specific biopsy targets within the solitary dilated duct, such as intraluminal mass or mural irregularity, as well as to provide a general indication of the practicality of ultrasound-guided percutaneous biopsy. None of our solitary dilated duct cases presented with the symptom of spontaneous nipple discharge. Therefore, only two patients underwent ductography (one benign, one DCIS), in each case because nipple discharge from a solitary duct opening was elicited after a dilated duct was identified at mammography. Only one study patient underwent MRI, which showed linear (presumably ductal) enhancement found to be benign at biopsy. MRI was performed infrequently, in part because most of our solitary dilated duct cases occurred before breast MRI was available at our institution. Although we do not report a number of cases sufficient to assess the roles of ductography and MRI in the evaluation of solitary dilated duct, use of these examinations should be considered as AJR:194, February

5 Chang et al. adjuncts to diagnostic mammography and ultrasound when clinically appropriate. Cardenosa et al. [20] described the clinical utility of ductography in showing the presence and location of intraductal lesions not visible at mammography. Hirose et al. [21] reported on the potential additional utility of MR ductography in evaluating cases of spontaneous nipple discharge. Our results indicate that patient characteristics such as age, family history of breast cancer, and personal history of breast cancer do not appear to correlate with the risk of malignancy associated with solitary dilated duct. Although we have limited data, the cause of solitary dilated duct when benign appears to be some form of fibrocystic change (mostly proliferative fibrocystic change, occasionally with apocrine metaplasia and atypical ductal hyperplasia). When it is malignant, solitary dilated duct appears to indicate the presence of DCIS. In almost all solitary dilated duct cases, the dilated duct appears to be filled with some debris, with or without accompanying fluid, as seen at ultrasound. Both cases of malignancy presenting as solitary dilated duct show similar pathology: DCIS with micropapillary and cribriform patterns. Each of these tumors grew within the duct in which it arose, expanding (dilating) the duct rather than forming a discrete mass as seen at mammography. Although the most common mammographic presentation of the micropapillary subtype of DCIS is grouped microcalcifications, it also may appear as a noncalcified focal asymmetry or solitary dilated duct [22, 23]. Our study has several limitations. First, given the rarity of solitary dilated duct as an isolated mammographic finding, our study population is small, although it is the largest reported to date. Second, we lack definitive pathologic correlation for 10 of our 21 solitary dilated duct cases. We assume that these cases were benign because 2-year-minimum mammographic follow-up did not reveal additional findings suggesting breast cancer and because tumor registry linkage did not identify subsequent breast cancer. However, a 2-year-minimum follow-up interval may be too short for malignancies likely to represent DCIS [24], so the actual positive predictive value for solitary dilated duct may be higher than reported in this study. Third, ours is a retrospective study performed at a single large academic medical center, and our results may not be applicable to breast imaging practices with substantially different patient and radiologist populations. The major strength of our study is the large-scale consecutive-case series on which it is based, an important attribute considering the rarity of solitary dilated duct. Given our observed % frequency of solitary dilated duct at screening mammography, among the approximately 30 million screening mammography examinations performed yearly in the United States [25], one may estimate that there will be 1,140 solitary dilated duct cases (30,000, = 1,140). We believe that the data we report may be helpful to the radiologists who encounter these cases because of the likelihood that many will have little to no previous personal experience with solitary dilated duct. The isolated mammographic finding of solitary dilated duct appears to have a substantially greater than 2% likelihood of malignancy, sufficiently high that a suspicious (BI-RADS category 4a) assessment may be more appropriate than the benign (BI-RADS category 2) or probably benign (BI-RADS category 3) assessment that is suggested in the BI-RADS atlas. References 1. Martin JE, Moskowitz M, Milbrath JR. Breast cancer missed by mammography. AJR 1979; 132: Martin JE. Atlas of mammography: histologic and mammographic correlations. Baltimore, MD: Williams & Wilkins, 1982:59 64, Moskowitz M. The predictive value of certain mammographic signs in screening for breast cancer. Cancer 1983; 51: Sadowsky N, Kopans DB. Breast cancer. Radiol Clin North Am 1983; 21: Wolfe JN. Xeroradiography of the breast, 2nd ed. Springfield, IL: Thomas, Sickles EA. Mammographic features of early breast cancer. AJR 1984; 143: Sickles EA. Mammographic features of 300 consecutive nonpalpable breast cancers. AJR 1986; 146: Ikeda DM, Andersson I. Ductal carcinoma in situ: atypical mammographic appearances. Radiology 1989; 172: Knutzen AM, Gisvold JJ. Likelihood of malignant disease for various categories of mammographically detected, nonpalpable breast lesions. Mayo Clin Proc 1993; 68: Huynh PT, Parellada JA, de Paredes ES, et al. Dilated duct pattern at mammography. Radiology 1997; 204: Sickles EA. Periodic mammographic follow-up of probably benign lesions: results in 3,184 consecutive cases. Radiology 1991; 179: American College of Radiology. ACR breast imaging reporting and data system (BI-RADS), 4th ed. Reston, VA: American College of Radiology, 2003: Sickles EA. The usefulness of computers in managing the operation of a mammography screening practice. AJR 1990; 155: Hunt KA, Rosen EL, Sickles EA. Outcome analysis for women undergoing annual versus biennial screening mammography: a review of 24,211 examinations. AJR 1999; 173: Sickles EA, Wolverton DE, Dee KE. Performance parameters for screening and diagnostic mammography: specialist and general radiologists. Radiology 2002; 224: Ernster VL, Ballard-Barbash R, Barlow WE, et al. Detection of ductal carcinoma in situ in women undergoing screening mammography. J Natl Cancer Inst 2002; 94: American College of Radiology. ACR breast imaging reporting and data system (BI-RADS), 4th ed. Reston, VA: American College of Radiology, 2003: Dee KE, Sickles EA. Medical audit of diagnostic mammography examinations: comparison with screening outcomes obtained concurrently. AJR 2001; 176: Le Doussal V, Tubiana-Hulin M, Friedman S, Hacene K, Spyratos F, Brunet M. Prognostic value of histologic grade nuclear components of Scarff- Bloom-Richardson (SBR): an improved score modification based on a multivariate analysis of 1262 invasive ductal breast carcinomas. Cancer 1989; 64: Cardenosa G, Doudna C, Eklund GW. Ductography of the breast: technique and findings. AJR 1994; 162: Hirose M, Nobusawa H, Gokan T. MR ductography: comparison with conventional ductography as diagnostic method in patients with nipple discharge. RadioGraphics 2007; 27[suppl 1]:S183 S Evans AJ, Pinder S, Ellis IO, et al. Screening-detected and symptomatic ductal carcinoma in situ: mammographic features with pathologic correlation. Radiology 1994; 191: Soo MS, Williford ME, Walsh R, Bentley RC, Kornguth PJ. Papillary carcinoma of the breast: imaging findings. AJR 1995; 164: Lev-Toaff AS, Feig SA, Saitas VL, Finkel GC, Schwartz GF. Stability of malignant breast microcalcifications. Radiology 1994; 192: Nass S, Ball J, eds. Institute of Medicine report on improving breast imaging quality standards. Washington, DC: National Academies Press, 2005: AJR:194, February 2010