Retrospective Analysis on Malignant Calcification Previously Misdiagnosed as Benign on Screening Mammography 스크리닝유방촬영술에서양성으로진단되었던악성석회화에대한후향적분석

Transcription

1 Original Article pissn / eissn Retrospective Analysis on Malignant Calcification Previously Misdiagnosed as Benign on Screening 스크리닝유방촬영술에서양성으로진단되었던악성석회화에대한후향적분석 Su Min Ha, MD 1, Joo Hee Cha, MD 2 *, Hak Hee Kim, MD 2, Hee Jung Shin, MD 2, Eun Young Chae, MD 2, Woo Jung Choi, MD 2 1 Department of Radiology, Research Institute of Radiology, Chung-Ang University Hospital, Seoul, Korea 2 Department of Radiology, Research Institute of Radiology, Asan Medical Center, University of Ulsan College of Medicine, Seoul, Korea Purpose: The purpose of our study was to investigate the morphology and distribution of calcifications initially interpreted as benign or probably benign, but proven to be malignant by subsequent stereotactic biopsy, and to identify the reason for misinterpretation or underestimation at the initial diagnosis. Materials and Methods: Out of 567 women who underwent stereotactic biopsy for calcifications at our hospital between January 2012 and December 2014, 167 women were diagnosed with malignancy. Forty-six of these 167 women had previous mammography assessed as benign or probably benign which was changed to suspicious malignancy on follow-up mammography. Of these 46 women, three women with biopsy-proven benign calcifications at the site of subsequent cancer were excluded, and 43 patients were finally included. The calcifications (morphology, distribution, extent, associated findings) in the previous and follow-up mammography examinations were analyzed according to the Breast Imaging Reporting and Data System (BI-RADS) lexicon and assessment category. We classified the patients into two groups: 1) group A patients who were still retrospectively re-categorized as less than or equal to BI-RADS 3 and 2) group B patients who were re-categorized as equal to or higher than BI- RADS 4a and whose results should have prompted previous diagnostic assessment. Results: In the follow-up mammography examinations, change in calcification morphology (n = 27, 63%) was the most frequent cause of assessment change. The most frequent previous mammographic findings of malignant calcification were amorphous morphology (n = 26, 60%) and grouped distribution (n = 36, 84%). The most frequent calcification findings at reassessment were amorphous morphology (n = 4, 9%), fine pleomorphic calcification (n = 30, 70%), grouped distribution (n = 23, 53%), and segmental calcification (n = 12, 28%). There were 33 (77%) patients in group A, and 10 patients (23%) in group B. Conclusion: Amorphous morphology and grouped distribution were the most frequent mammographic findings of calcifications that were misinterpreted or underestimated by the initial radiologist and confirmed as malignancy at follow-up. Index terms Breast Calcification Neoplasms Received March 31, 2016 Revised May 25, 2016 Accepted July 23, 2016 *Corresponding author: Joo Hee Cha, MD Department of Radiology and Research Institute of Radiology, University of Ulsan College of Medicine, Asan Medical Center, 88 Olympic-ro 43-gil, Songpa-gu, Seoul 05505, Korea. Tel Fax jhcha@amc.seoul.kr This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License ( which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited. INTRODUCTION When calcifications are detected by mammography, their morphology and distribution require further analysis. The Breast Imaging Reporting and Data System (BI-RADS) has defined morphologic descriptors that predict benignity or malignancy, and according to this system, distribution is at least as important as morphology. In a previous study assessing the interobserver Copyrights 2017 The Korean Society of Radiology 251

2 Misinterpreted or Missed Calcifications on variability of BI-RADS usage, calcification descriptors were inconsistently used among the readers (2). Moreover, a large retrospective review has indicated that two-thirds of all calcifications referred for biopsy are described as pleomorphic (3), which is a rather ambiguous term that relies on the intuition of the radiologist. Malignant breast lesions that initially have a probably benign mammographic appearance should be promptly identified with a short-term mammographic follow-up (4), but perception errors or interpretation failures by the radiologists are still unavoidable. It is important to reduce the frequency of missed or misinterpreted cancers with calcifications, especially on mammography. In a previous study, calcifications with or without density were found in one in every four missed or minimal sign interval cancers (5). Vitak et al. (6) also reported calcifications in 13% of missed cancers in women aged years. In our present study, we carefully investigated the mammographic findings of calcifications that were missed or misinterpreted at initial mammography and subsequently diagnosed as malignant after stereotactic biopsy on follow up mammography. MATERIALS AND METHODS Patient Selection Our Institutional Review Board approved this retrospective study and waived the requirement for informed consent. We used the examination code for stereotactic core biopsy to search through our Picture Archiving and Communication System for the records of patients who underwent this biopsy procedure from January 2012 to December Stereotactic biopsy was performed with 11-gauge vacuum probes (Mammotome; Ethicon Endo-Surgery, Cincinnati, OH, USA) on an upright stereotactic digital unit using the Senographe Essential Stereotaxy machine (General Electric Medical Systems, Milwaukee, WI, USA). After a mean number of 12 specimens per biopsy had been retrieved using this procedure, specimen radiography was performed to confirm the inclusion of representative calcifications. Of the 567 patients who had undergone a stereotactic core biopsy for calcifications, 167 women were diagnosed with malignancy in the final pathologic records. The inclusion criteria for our study were as follows: patients whose calcifications on mammography were initially diagnosed as benign or probably benign and subsequently changed to suspicious malignancy during follow-up mammography and whose final biopsy or surgical pathologic results revealed a malignancy. Forty-six patients who met our inclusion criteria were selected, although three patients with an existing biopsy-proven benign pathology were excluded later. Finally, 43 patients, all asymptomatic, ranging in age from 31 to 72 years (mean, 51.1 years), were included in our analyses. Mammographic Evaluation All mammography examinations were performed using the Senographe Essential machine (General Electric Medical Systems). Most patients underwent routine two-view mammography (craniocaudal and mediolateral oblique views) and additional diagnostic magnification views were obtained, if needed. If mammography from an outside hospital was available and a good image quality had been obtained, only an additional magnification view was obtained at our institution. Follow-up mammography was carried out after 6 months if the initial mammography showed the diagnosis of BI-RADS 3 or after 1 year if a diagnosis of BI-RADS 1 or 2 was obtained. Additional diagnostic magnification views were obtained if the radiologist had observed any increase in the extent or change in morphology or distribution. Twenty-seven (62.7%) patients underwent twoview routine mammography only (without an additional magnification view), while 39 (90.7%) patients underwent routine mammography and an additional magnification view simultaneously or magnification view only in the follow-up study. Two radiologists with 4 and 15 years experience in breast imaging retrospectively reviewed the calcifications to assess their morphology, distribution, extent, and associated findings according to the BI-RADS lexicon. The final assessment was determined by consensus between the two readers on both the initial and follow-up mammography. The readers were blinded to the follow-up mammography and the final pathologic result while interpreting the initial mammography findings. The morphology and distribution descriptors of calcifications in the BI-RADS fifth edition (1) were used; the radiologist could select terms from the morphologic descriptors typically benign (punctate) or suspicious (amorphous, coarse heterogeneous, fine pleomorphic, and fine linear branching) and the distribution descriptors diffuse (randomly throughout the breast), regional 252 jksronline.org

3 Su Min Ha, et al [scattered in a large volume (> 2 cm) of the breast not conforming to a ductal distribution], grouped [at least five calcifications occupying a small volume (< 1 cm) of breast tissue], segmental (covering less than a quadrant and extending in a triangular distribution with the apex pointing at the nipple), or linear (arrayed in a line). The interpreting radiologist was allowed to use more than one morphologic and distribution descriptor. If more than one descriptor was assigned, we used the most suspicious descriptor in the analysis since treatment decisions should be based on this descriptor. Readers also evaluated whether there were combined associated findings such as focal asymmetry, architectural distortion, and a suspicious mass. The extent of calcification was also measured by using the distance between both ends of calcifications with any distribution. Two radiologists classified calcifications into one of the following categories: category 2, benign; category 3, a probably benign lesion with 2% or lower probability of malignancy; category 4a, an indeterminate lesion with low suspicion of malignancy (risk of malignancy of 3 10%); category 4b, an intermediate lesion with moderate suspicion of malignancy (risk of malignancy of 11 49%); category 4c, an intermediate lesion with moderate-to-high suspicion of malignancy (risk of malignancy of 50 94%); and category 5, a malignant lesion with 95% or higher likelihood of malignancy. The readers recorded the mammographic findings (Table 1) and assessment categories of the initial mammography described previously by the initial radiologists, and then they re-analyzed the calcification and re-categorized the initial mammography Table 1. Morphology and Distribution of Initial and Follow-Up as Interpreted by the Initial Radiologist (n = 43) Initial Follow-Up Morphology Amorphous 32 (74%) 0 Coarse heterogeneous 6 (14%) 9 (21%) Punctate 5 (12%) 0 Fine pleomorphic 0 25 (58%) Linear branching 0 9 (21%) Distribution Grouped 32 (75%) 19 (44%) Regional 7 (16%) 2 (5%) Segmental 4 (9%) 18 (42%) Linear 0 4 (9%) Diffuse 0 0 according to the BI-RADS classification to identify the most likely reason for the misinterpretation or underestimation of calcifications. We also analyzed how the mammographic findings of calcifications later changed on the follow-up diagnostic examination to being suggestive of suspicious malignancy. Three additional radiologists with 5 years experience in breast imaging, and who were blinded to patient s information and the followup mammography, were involved in BI-RADS re-categorization of the initial mammography. Finally, we classified the patients into two groups: 1) group A: patients who were still retrospectively re-categorized as less than or equal to BI-RADS 3, and 2) group B: patients who were re-categorized as equal to or higher than BI-RADS 4a and whose results should have previously prompted subsequent diagnostic assessment. If more than three radiologists agreed for recall of the initial mammographic finding, the patients were assigned to group B. The time between the initial mammographic detection and tissue diagnosis was determined by calculating the time from the initial follow-up recommendation to the time of the biopsy recommendation. Statistical Analysis A two-sample t-test was used to assess the relationship between malignancy and the difference in extent of calcification between the previous and follow-up mammography examinations and between the surgical method (breast conserving operation or total mastectomy) and the difference in calcification. A p value less than 0.05 was considered to indicate a statistically significant difference. RESULTS The mean interval between the initial follow-up recommendation and the subsequent biopsy recommendation for 43 malignancies was 13.6 months (range, 4 69 months). At least a 6-month mammographic follow-up was performed for 4 of the 43 lesions (9%), follow-up within 12 months was performed for 17 lesions (40%), follow-up within 2 years was performed for 18 lesions (42%) and follow-up for more than 2 years was performed for 4 lesions (9%). Of these 43 lesions, according to the pathologic examination, 35 (82%) were ductal carcinomas in situ or ductal carcinomas in situ with microinvasion, 5 (12%) jksronline.org 253

4 Misinterpreted or Missed Calcifications on were invasive ductal carcinomas, 1 (2%) was mixed tubular and lobular carcinoma, 1 (2%) was tubular carcinoma, and 1 (2%) A B Fig. 1. Mammographic images of a 59-year-old woman diagnosed with ductal carcinoma in situ. The initial mammographic image (A) is interpreted as showing scattered, punctate, and amorphous calcifications (arrows), which are categorized as BI-RADS category 3 and later re-categorized as BI-RADS category 3, without a magnification view (group A). A follow-up mammographic image with a magnification view (B) shows fine pleomorphic calcifications (arrows), which have increased in their extent, with segmental distribution. The calcifications are thus categorized as BI-RADS category 4a. BI-RADS = Breast Imaging Reporting and Data System Table 2. Morphology and Distribution Changes on Follow-Up in the Minimal Sign Group (n = 33) Initial Follow-Up Morphology Amorphous 20 (61%) 2 (6%) Coarse heterogeneous 4 (12%) 0 Punctate 9 (27%) 0 Fine pleomorphic 0 25 (76%) Linear branching 0 6 (18%) Distribution Grouped 29 (88%) 19 (58%) Regional 3 (9%) 4 (12%) Segmental 1 (3%) 9 (27%) Linear 0 1 (3%) Diffuse 0 0 was invasive lobular carcinoma. Increased extent of calcification was associated with malignancy in 30 of the 43 patients (70%); the average extent of calcification was 1.22 cm (range, cm) on the previous mammography and 2.30 cm (range, ) in the follow-up study (p < 0.001). Thirty-two (74%) patients underwent breast conserving operation and 11 (26%) patients were treated with total mastectomy. The surgical method was not related to an increase in the extent of calcification (p = 0.56). Eight (19%) patients were confirmed as showing lymph node metastasis in the surgical specimen, while 35 (81%) patients were pathologically negative. Out of the 43 patients, there were 33 (77%) patients with calcifications that were evident in retrospect but they were not thought to be suspicious and were re-categorized as less than or equal to BI-RADS 3 (group A). Ten patients (23%) had clearly abnormal findings that were atypical and suggestive of cancer and they were re-categorized as equal to or higher than BI-RADS 4a (group B, which represented misinterpreted or overlooked ). In group A, out of the 33 patients assessed as benign or probably benign by the initial interpreting radiologist, 20 (61%) showed amorphous morphology and 29 (88%) had grouped distribution of calcifications (Fig. 1) on the initial mammography. Among these 33 patients, fine pleomorphic (25/33, 76%) morphology and grouped (19/33, 58%) distribution were assessed as suspicious malignancy on follow up mammography prior to biopsy (Table 2). In group B, out of the 10 patients who were misinterpreted or overlooked by the initial radiologist, 6 cases (60%) showed amorphous morphology and 7 cases (70%) showed grouped calcifications on the initial mammography (Fig. 2). Among these 10 patients, fine pleomorphic (5/10, 50%) morphology and grouped (4/10, 40%) distribution were assessed as suspicious malignancy on follow up mammography prior to biopsy (Table 3). Analysis of previous and follow-up mammography for cancers in both groups showed that a change in calcification morphology (n = 27, 63%), distribution (n = 3, 7%), or both (n = 13, 30%) were the most frequent reasons for the assessment change. Most calcifications did not show any associated findings (34/43, 79%); out of the remaining patients, eight patients (19%) presented with asymmetry and one patient (2%) presented with architectural distortion on the previous and follow-up mammography. Finally, retrospective reassessment of the BI-RADS category showed that the previous mammography identified 254 jksronline.org

5 Su Min Ha, et al A Fig. 2. Mammographic images of a 40-year-old woman diagnosed with ductal carcinoma in situ. The initial mammographic image (A) shows grouped, amorphous, and punctate calcifications, which are considered probably benign and are categorized as BI-RADS category 3 by the initial radiologist, but are retrospectively re-categorized as BI-RADS category 4a (group B). A subsequent mammographic image with a magnified view (B) acquired after 13 months shows an increase in the extent of fine pleomorphic and linear branching calcifications, which are re-categorized as BI-RADS category 4b. BI-RADS = Breast Imaging Reporting and Data System B Table 3. Morphology and Distribution Changes on Follow-Up in the Missed Cancer Group (n = 10) Initial Follow-Up Morphology Amorphous 6 (60%) 2 (20%) Coarse heterogeneous 2 (20%) 0 Punctate 2 (20%) 0 Fine pleomorphic 0 5 (50%) Linear branching 0 3 (30%) Distribution Grouped 7 (70%) 4 (40%) Regional 1 (10%) 0 Segmental 1 (10%) 3 (30%) Linear 0 3 (30%) Diffuse 1 (10%) 0 Group A Initial BI-RADS II: 5 III: 28 Group B Initial BI-RADS Re-categorized BI-RADS II: 5 23 (82%) 5 (100%) Follow up mammography 3 (11%) III: 28 IVb: 3 Re-categorized BI-RADS IVa: 8 2 (7%) 6 (75%) IVa: 28 IVc: 2 Follow up mammography IVa: 6 five patients with BI-RADS category 2; all with punctate morphology with grouped distribution, and twenty-eight patients with BI-RADS category 3 in group A, and two patients with BI- RADS 4b category and eight patients with BI-RADS 4a category in group B. In the follow-up mammography prior to stereotactic biopsy, all patients were categorized as equal to or higher than BI-RADS category 4a (Fig. 3). In group A, 26 (79%) ductal carcinomas in situ or ductal carcinomas in situ with microinva- III: 10 IVb: 2 2 (25%) 2 (100%) IVb: 2 Fig. 3. Flow chart of change in the BI-RADS assessment category in groups A and B. BI-RADS = Breast Imaging Reporting and Data System sion and 7 (21%) invasive cancers were diagnosed pathologically, and in group B, 9 (90%) ductal carcinomas in situ or ductal carcinomas in situ with microinvasion and 1 (10%) invasive cancer were diagnosed pathologically. In group A, all five pa- V: 2 jksronline.org 255

6 Misinterpreted or Missed Calcifications on tients assessed as BI-RADS 2 were diagnosed as having ductal carcinoma in situ. DISCUSSION Various types of calcifications have different probabilities of malignancy (7, 8), and diagnostic interpretation of these calcifications is difficult and subjective and therefore it varies among readers. Indeed, although calcification descriptors and categories in the BI-RADS lexicon predict the risk of malignancy for suspicious calcifications and facilitate communication between radiologists, there is still considerable variability in interpretation (3, 9, 10). While interpreting calcifications on mammography, we have to consider both morphology and distribution. The calcification morphologic descriptors coarse heterogeneous, amorphous, fine pleomorphic, and fine linear branching calcifications have a progressively increasing risk of malignancy (11). With distribution descriptors, calcifications in a ductal distribution are much more likely to be located in contiguous terminal ductal lobular units and represent malignancies such as ductal carcinoma in situ, whereas grouped calcifications represent an intermediate risk of malignancy. Diffuse or regional calcifications show a low likelihood of malignancy due to calcifications in scattered glands, lobules, or stromal elements of the breast and are thus more likely to be benign (11). More than half of the cancers in our study were amorphous (60%) or grouped (84%) calcifications in the initial mammography and were assessed as suspicious malignancy in the follow-up mammography as 30 fine pleomorphic (70%) and 23 grouped (53%) calcifications. Mammographically amorphous calcifications have been considered indeterminate with variable recommendations for follow-up or biopsy. Amorphous calcifications are considered punctate and followed up by one radiologist or pleomorphic and subject to biopsy by another (10). They have also been described as sufficiently small or hazy that a more specific morphologic classification cannot be made (1). In a prior study by Liberman et al. (3), 74% of all segmentally distributed calcifications and 68% of those in a linear distribution proved malignant compared with 36% of grouped calcifications. In the current study, we found a significant percentage of malignancies with grouped distribution of calcifications: 58% (19/33) in group A were grouped calcifications and 40% (4/10) in group B were grouped calcifications. We realized that radiologists often consider a higher than expected percentage of grouped calcifications to be probably benign or less than a BI- RADS category 3. A BI-RADS category 3, the probably benign finding, is suggested for lesions with a low likelihood of malignancy (risk of malignancy < 2%) that do not require immediate biopsy. It has been reported that a small percentage of lesions assigned to the probably benign group that are actually malignant should be rapidly identified within 6 12 months by a change in appearance at subsequent imaging (12). The mean time to biopsy of malignant lesions in our study was 13.6 months (range, 4 69 months). In the Sickles study (4), almost all cancers were identified by a change in the mammographic appearance at a 6- or 12-month follow-up. According to Sickles (4), a comparison with prior mammography or additional views such as a magnification view will result in a different management recommendation than a shortinterval follow-up. In our present study, 63% (27/43) were imaged with two-view routine mammography only (without an additional magnification view) initially. In contrast, 90.7% (39/43) were evaluated with either routine mammography and an additional magnification view or magnification view only at the follow-up, where a suspicious malignancy was identified by the interpreting radiologist who subsequently recommended stereotactic biopsy. Therefore, for correct characterization of the morphology, it is important to acquire an additional magnification view. In real practice, every radiologist tends to analyze mammography images based on their individual training and instincts. Thus, assigning a category based on BI-RADS descriptors involves subjective interpretation by the radiologist (11). One previous study evaluated mammography performance and found that discordant recommendations were evident for mammograms classified as probably benign and that 46% of interval cancers had been assessed as probably benign (13). In the current study, the previous radiologist described suspicious calcifications but finally regarded them to be BI-RADS 3 and recommended short-term follow-up rather than stereotactic biopsy. Again, we contend that the use of the term benign or probably benign should be based on descriptors and categories in the 256 jksronline.org

7 Su Min Ha, et al BI-RADS lexicon rather than on the individual interpretation of the radiologist, and our study underscores the importance of identifying malignancies among benign-appearing lesions on mammography. There were several limitations to our present study. First, our analyses were retrospective and although two radiologists reviewed the images in consensus, we did not evaluate interobserver variability. Second, at the initial mammography, the interpreting radiologist did not examine the magnification view and a 6-month or 1-year follow up was recommended and this omission may have caused an interpretation error. Third, there is a possibility of a sample-selection bias since we did not analyze patients initially assigned as benign or probably benign on mammography and diagnosed as pathologically benign. Fourth, we had a limited number of patients, which potentially limits the general applicability of our findings. In conclusion, we observed that on mammography more than half of breast cancers manifest as amorphous or grouped calcifications. It is unavoidable that radiologists occasionally miss or misinterpret subtle mammographic findings. However, it is worth emphasizing that the benign or probably benign category should be assessed based on the BI-RADS lexicon and assessment category and an effort should be made to follow up patients more closely, thus, decreasing the incidence of delayed diagnosis of breast cancer while at the same time decreasing unnecessary breast biopsies. REFERENCES 1. American College of Radiology. ACR BI-RADS atlas: breast imaging reporting and data system. 5th ed. Reston, VA: American College of Radiology, Baker JA, Kornguth PJ, Floyd CE Jr. Breast imaging reporting and data system standardized mammography lexicon: observer variability in lesion description. AJR Am J Roentgenol 1996;166: Liberman L, Abramson AF, Squires FB, Glassman JR, Morris EA, Dershaw DD. The breast imaging reporting and data system: positive predictive value of mammographic features and final assessment categories. AJR Am J Roentgenol 1998;171: Sickles EA. Periodic mammographic follow-up of probably benign lesions: results in 3,184 consecutive cases. Radiology 1991;179: Hofvind S, Geller B, Skaane P. Mammographic features and histopathological findings of interval breast cancers. Acta Radiol 2008;49: Vitak B, Olsen KE, Månson JC, Arnesson LG, Stål O. Tumour characteristics and survival in patients with invasive interval breast cancer classified according to mammographic findings at the latest screening: a comparison of true interval and missed interval cancers. Eur Radiol 1999;9: Sickles EA. Breast calcifications: mammographic evaluation. Radiology 1986;160: Sickles EA. Mammographic features of early breast cancer. AJR Am J Roentgenol 1984;143: Berg WA, Arnoldus CL, Teferra E, Bhargavan M. Biopsy of amorphous breast calcifications: pathologic outcome and yield at stereotactic biopsy. Radiology 2001;221: Berg WA, Campassi C, Langenberg P, Sexton MJ. Breast Imaging Reporting and Data System: inter- and intraobserver variability in feature analysis and final assessment. AJR Am J Roentgenol 2000;174: Burnside ES, Ochsner JE, Fowler KJ, Fine JP, Salkowski LR, Rubin DL, et al. Use of microcalcification descriptors in BI- RADS 4th edition to stratify risk of malignancy. Radiology 2007;242: Rosen EL, Baker JA, Soo MS. Malignant lesions initially subjected to short-term mammographic follow-up. Radiology 2002;223: Poplack SP, Tosteson AN, Grove MR, Wells WA, Carney PA. in 53,803 women from the New Hampshire mammography network. Radiology 2000;217: jksronline.org 257

8 Misinterpreted or Missed Calcifications on 스크리닝유방촬영술에서양성으로진단되었던악성석회화에대한후향적분석 하수민 1 차주희 2 * 김학희 2 신희정 2 채은영 2 최우정 2 목적 : 유방촬영술에서의미세석회화를양성또는양성추정소견이라고분석하였으나추적검사후입체정위생검술을시행 하여유방암으로진단된경우를대상으로후향적으로석회화의모양및분포에대해재분석하여영상의학적해석또는인 지의오류의원인을알아보고자하였다. 대상과방법 : 유방입체정위생검술을 2012 월 1 월부터 2014 년 12 월까지시행한환자 567 명중 167 명이유방암으로진단되 었다. 이중 46 명의환자가이전유방촬영술에서양성 / 양성추정소견을진단받고추적검사중악성의심소견이유방촬영술 에서보였다. 이중 3 명이조직검사를통해양성으로진단된군으로제외하였고, 총 43 명이연구에포함되었다. 석회화 ( 모 양, 분포, 범위, 동반된소견 ) 의분석은 Breast Imaging Reporting and Data System ( 이하 BI-RADS) lexicon 에따라기술 하였다. 그룹 A 군은후향적으로보았을때에도 BI-RADS 3 보다낮거나같은경우, 그룹 B 군은후향적으로보았을때 BI- RADS 4a 보다높거나같은경우로, 2 개의그룹으로환자군을나누어정리하였다. 결과 : 유방촬영술추적검사결과, 석회화의모양변화 (n = 27, 63%) 가 BI-RADS category 의변화를가장많이초래하였 다. 가장흔한이전유방촬영술소견은무정형 (n = 26, 60%), 군집성 (n = 36, 84%) 석회화의모양및분포가해석및인 지오류를일으켰고, 추적검사시시행한유방촬영술에서는무정형 (n = 4, 9%), 미세불균질 (n = 30, 70%), 군집성 (n = 23, 53%), 분절성 (n = 12, 28%) 의미세석회화의모양및분포가가장흔하였고유방암으로진단되었다. 33 명 (77%) 이그 룹 A 군으로, 10 명 (23%) 이그룹 B 군으로분류되었다. 결론 : 무정형및군집성미세석회화를유방촬영술에서판독의사가가장흔히놓치거나양성으로진단하였으며추적검사 를통해악성소견이의심되어궁극적으로유방암으로진단되었다. 1 중앙대학교병원영상의학과, 2 울산대학교의과대학서울아산병원영상의학과 258 jksronline.org