Original papers
DOI: 10.11152/mu.2013.2066.172.cclAbstract
Objective: This study aimed at investigating the characteristic imaging findings of the columnar cell lesions (CCLs) of the breast via mammography (MG), ultrasonography (US), and magnetic resonance imaging (MRI). Materials and meth- ods: The MG, US and MRI findings of 72 patients with histopathological diagnosis of CCLs were retrospectively evaluated.
Histopathologically, the CCLs were divided into those with and without atypia; the radiological findings of these two groups were compared with a Chi-square test. Results: Sixty-nine patients underwent stereotaxic biopsy (MG-guided in 50 patients and US-guided in 19 patients) and 3 patients underwent US-guided core needle biopsy; all of these patients were diagnosed with CCLs based on a histological examination. The evaluation of the CCLs in patients that underwent MG-guided stereotaxic biopsy revealed that the most common type of microcalcifications were amorphous-indistinct (52%, n= 26/50) and the most common microcalcification distribution pattern was clustered type (76%, n= 38/50). The ratio of CCLs with atypia was similar in patients with high-risk microcalcifications and in those with benign or intermediate-risk microcalcifications (OR: 1.13, 95%
CI: 0.573-2.227, p: 0.475). On the other hand, those patients who underwent US-guided biopsies for the evaluation of CCLs had similar proportions of cystic or solid lesions, posterior acoustic shadowing and contour irregularities whether or not they had atypia (p: 0.584, 0.075, 0.187, respectively). Patients with atypia had a higher number of lesions greater than 1 cm via US as compared to those without atypia, but this difference was not statistically significant (p: 0.06). MRI findings were also similar in patients with and without atypia. Conclusions: MG revealed that clustered distribution patterns and amorphous- in- distinct type microcalcifications were more commonly seen in patients with CCLs; however, there was no significant relation- ship between US or MRI findings and CCLs. In addition, the MG, US and MRI findings were similar in patients with CCLs that did or did not have histopathological characteristics of atypia.
Keywords: breast, columnar cell lesion, microcalcification, mammography, magnetic resonance imaging, ultrasonography
Columnar cell lesions of the breast: radiological features and histological correlation
Aktas Elif, Sahin Burcu, Ciledag Nazan, Cosar Zehra Sumru, Arda Niyazi Kemal
Oncology Education and Research Hospital, Department of Radiology, Ankara, Turkey
Received 14.01.2015 Accepted 20.03.2015 Med Ultrason
2015, Vol. 17, No 2, 147-154 Corresponding author: Elif Aktaş
Mehmet Akif Ersoy Mah. 13. Cad. No 56.
06200 Yenimahalle, Ankara, Turkey Phone :90-05063431724 / 90-312-3360909 Fax: 90312 334 03 52
Email: [email protected]
Introduction
Columnar cell lesions (CCLs) of the breast represent changes in epithelial cells characterized by enlarged ter- minal ductal lobular units that contain snout cells [1].
CCLs have been described by a variety of terms, includ- ing columnar cell change or columnar cell hyperplasia with or without atypia, columnar metaplasia, columnar
alteration of lobules, blunt duct adenosis, columnar al- teration with prominent snouts and secretions, enlarged lobular units with columnar alterations, hyperplastic un- folded lobules and hyperplastic enlarged lobular units [2,3]. The World Health Organization Working Group on the Pathology and Genetics of Tumors of the Breast rec- ommended the term flat epithelial atypia (FEA) for CCLs containing atypia. It is important to distinguish CCLs with atypia from those without atypia. The lesions with atypia are often accompanied by ductal carcinoma in situ (DCIS), atypical ductal hyperplasia (ADH) and tubular carcinoma (TC). Therefore, all lesions with atypia must be surgically removed, and every part of the removed specimen should be examined. Recently, advanced digi- tal mammography (MG) techniques have allowed for better detection of microcalcifications. Accordingly, the
number of biopsies has increased, which has led to the increased detection of CCLs [3]. If specific radiological characteristics of CCLs with atypia can be detected, there may be little need for biopsy in the future.
CCLs are frequently observed via MG as amorphous and fine pleomorphic microcalcifications; this type of mi- crocalcification is also observed in DCIS, ADH, ordinary epithelial hyperplasia, papillomatosis, and sclerosing adenosis [3]. To our knowledge, there is little published regarding the ultrasonography (US) characteristics of CCLs. The two published series we found have a limited number of patients and identified CCLs in lesions with suspicious morphological characteristics via US [4,5].
Further, there are few reports in the literature re- garding the imaging characteristics of CCLs, and to our knowledge, no magnetic resonance imaging (MRI) find- ings of CCLs have been reported in the English litera- ture. Therefore, in the current study, we aimed to identify radiological findings that could indicate CCLs with and without atypia via MG, US and MRI in patients with his- topathological proved CCLs.
Materials and methods
The Ethics Committee of the Ankara Oncology Edu- cation and Research Hospital approved this retrospective analysis.
Study population and inclusion criteria
Upon review of our hospital database, it was deter- mined that 1570 consecutive needle biopsies under ste- reotactic MG or US guidance were performed between April 2009 and August 2011.
Patients were included in the study if they had histo- pathological diagnosis of CCLs and if they underwent MG and US examinations prior to biopsy. Patients were exclud- ed from the study if their histological examination revealed CCLs accompanied by lobular neoplasia (LN), DCIS or TC.
CCLs of the breast were identified in 89 cases (89/1570 [5.6%]). Sixteen lesions were excluded due to their association with concomitant LN (10 patients), DCIS (4 patients) or TC (2 patients). One patient was excluded from the study due to the lack of MG. The re- maining 72 patients constituted the study group.
All of the clinical and patient data used in this study were gathered from the electronic medical record system:
age, disease symptoms, personal history of breast cancer, number of lesions, lesion size, histopathology results, and imaging results from follow-up. All cases were in- terpreted by a pathologist with ten years of experience in breast pathology.
Fifty of the 72 patients with histopathologically de- tected CCLs had suspicious microcalcifications by MG;
in these cases, MG-guided wire localization was applied and surgical excision was performed. The other 22 pa- tients underwent US-guided biopsy. In 14 patients from this group US examination showed cystic lesions and fine needle aspiration cytology (FNAC) found suspicious atypia; therefore, US-guided wire localization and exci- sional biopsy was performed in the second stage and in this way the diagnosis of CCLs was established. In other 8 patients solid lesions were found by US and core needle biopsy (CNB) was performed (5 patients with CCLs with atypia and US-guided excisional biopsy performed during the second phase and 3 patients without atypia in (fig 1).
Fig 1. Algorithm showing patient selection and stages during the diagnosis in patients diagnosed with CCLs (CCLs: columnar cell le- sions, DCIS: ductal carcinoma in situ, LN: lobuler neoplasia, TN: tubuler neoplasia, MG: mammography, US: ultrasonography, CNB:
core needle biopsy, FNAC: fine needle aspiration cytology)
The US findings of 50 patients who had indications for excisional biopsy based on the MG findings were retrospectively evaluated. Further, the MG findings of 19 patients with indication for excisional biopsy based on their US findings and 3 patients who underwent only CNB were retrospectively evaluated.
Breast MRI was used for the additional evaluation of inconclusive imaging findings in 10 patients.
Imaging technique
MGs were performed with a dedicated digital MG system (Lorad/Hologic, Danbury, USA). Standard cranio-caudal and mediolateral oblique views were rou- tinely obtained, and further mammographic views were obtained as necessary.
An experienced radiologist performed US with a GE Vivid 3 pro (USA) with 10 MHz linear array transducers.
All of the MRI examinations were performed on a 1.5 MRI System (Signa Horizon Echo Speed; GE Health- care, Milwaukee, USA) with a bilateral 8-channel breast coil. Standard sequences were obtained as follows: axial T1-weighted spin-echo sequence, sagittal fat suppressed T2-weighted fast spin-echo sequence, axial STIR se- quence, and axial T1-weighted 3D fat-suppressed fast spoiled gradient-echo sequence. These sequences were taken prior to and four times after the rapid injection of a bolus of 0.1 mmol/L Gadodiamid (Omniscan, Opakim) per kilogram of body weight followed by 20 ml saline solution.
Two radiologists, one with 10 and one with 5 years of experience in breast imaging, retrospectively reviewed the MG, US and MRI findings; they were in consensus on all findings. The radiologists classified the lesions ac- cording to the American College of Radiology Breast Imaging Reporting and Data System (BI-RADS) lexicon [6].
At MG, microcalcifications were classified into four different categories as follows: amorphous or indistinct, fine pleomorphic, coarse and heterogeneous, or round.
Their distribution was classified as regional, clustered or segmental. Accompanying findings, such as asymmetric density and architectural distortion, were also noted. If the patient had undergone previous mammographic ex- aminations, the progression of microcalcifications was also observed.
At US, the lesions were classified by their contour, shape, orientation, echo pattern, and internal and poste- rior acoustic features.
At MRI the shape, margin, enhancement, kinetic curve pattern and T2-weighted appearance of the masses were evaluated.
The histopathology findings were classified as CCLs without atypia or CCLs with atypia.
Statistical analysis
Normally distributed continuous data were expressed as mean ± SD and categorical data were expressed as n (%). A normal distribution was defined by a histogram and the One Sample Kolmogorov Smirnov Test. The Statistical Package for Social Sciences (SPSS) 13.0 for Windows was used to analyze the data. For continuous variables, a Student’s t-test was used to compare differ- ences between groups, and all continuous variables were defined as mean ± standard deviation (SD). The Chi- square test was used to compare categorical variables, and all categorical variables were defined as numbers and percentages. Linear regression analysis was used to evaluate the association between CCLs with atypia and microcalcification patterns and US findings. A value of p<0.05 was considered statistically significant.
Results
The mean age of the 72 patients was 47.9 ± 6.37 years (range: 34 to 67 years), and 6 of the patients were young- er than 40 years. Fifteen patients were admitted with complaints of mastalgia. Palpable painful swelling was observed in 5 patients, while 49 patients were asympto- matic. Two patients had a history of contralateral breast carcinoma.
Histopathologically, 42 lesions were determined as CCLs without atypia (58.3%) and the remaining 30 le- sions were associated with atypia (41.6%). The CCLs with atypia were detected in 21 patients that underwent MG-guided biopsy and in 9 patients that underwent US- guided biopsy (Table I).
MG revealed that amorphous-indistinct microcalci- fications (52%, 26/50) and fine-pleomorphic microcal- cifications (34%, 17/50) were more frequently seen in patients with CCLs (fig 2, fig 3). The most common type of microcalcification distribution pattern was clustered (76%, 38/50). MG also revealed that CCLs with atypia were detected in similar ratios in patients with high-risk microcalcifications and in patients with benign and in- termediate-risk microcalcifications (OR: 1.13, 95% CI:
0.573-2.227, p=0.45). Moreover, there was no signifi- cant difference in the incidence of CCLs with atypia in patients whose MG showed cluster-forming microcalci- fications compared with patients that had regional mi- crocalcifications (OR: 1.004 and 95% CI: 0.733-1.376, p=0.627).
Further, there was no significant relationship between the CCLs and the US findings. The US findings of the pa- tients who underwent US-guided biopsy are listed in Table II. The greatest dimension of the masses on US ranged from 6 to 20 mm (mean=10.5±3.8mm). Comparing the
US findings of patients with CCLs with and without atyp- ia revealed that cystic or solid lesions, posterior acoustic shadowing and contour irregularities were detected in similar proportions in both groups (p=0.584, 0.075, 0.187, respectively) (fig 4, fig 5). Moreover, the US indication of lesions larger than 1 cm was more common in patients with CCLs with atypia as compared to those without atyp- ia, but this difference was not significant (p=0.06).
In 11 patients, MG revealed suspicious microcalcifica- tions; focal ductal ectasia was detected in the localizations of Table I. Ultrasonography vs mammography findings and histopathological comparison of patients with CCLs.
Total CCLs with atypia CCLs without atypia P* value Mammography
Shape (n, %)
Round microcalcification Amorphous–indistinct Heterogeneous Fine-pleomorphic
426 317
3 (75%) 9 (34.6%) 1 (33.3)%
8 (47.1%)
1 (25%) 17 (65.4%) 2 (66.7%) 9 (57.9%)
0.45
Distribution (n, %) Regional
Clustered 12
38 5 (41.7%)
16 (42.1%) 7 (58.3%)
22 (57.9%)
0.62
Accompanying Findings (n, %) Asymmetric density
Architectural distortion 2
1 0
0 2 (100%)
1 (100%)
0.31
Change (n, %) Progression Stable Not known
124 34
3 (25%) 1 (25%) 17 (50%)
9 (75%) 3 (75%) 17 (50%)
0.32
Ultrasonography Solid lesions (n, %) Cystic lesions (n, %)
Simple cyst Complicated cyst Complex cyst
8 13 10
3 (37.5%) 02 (66.7%) 3 (30%)
5 (62.5%) 1 (100%) 1 (33.3%) 7 (70%)
0.74
Posterior acoustic shadowing (n, %) Present
Absent 6
16 1 (16.7%)
7 (43.8%) 5 (83.3%)
9 (56.2%)
0.35
Contour of lesions (n, %) Regular
Irregular Indistinct
133 6
5 (38.5%) 0 (0%) 3 (50%)
8 (61.5%) 3 (100%) 3 (50%)
0.48
Size of the lesion
≤1 cm
>1cm 10
12 1 (10%)
7 (58.3%) 9 (90%)
5 (41.7%)
0.06
P* < 0.05 was statistically significant; CCLs: columnar cell lesions
Fig 2. a. Clustered fine pleomorphic microcalcifications in the upper outer quadrant of the right breast are seen on craniocau- dal-mammography, b. Specimen radiography shows that micro- calcifications were removed successfully. Pathology detected columnar cell hyperplasia with atypia.
Fig 3. a. Left craniocaudal mammogram shows clustered coarse-heterogeneous microcalcifications in the inner quadrant, b. Specimen radiography demonstrates that microcalcifications were removed successfully. Pathology detected columnar cell alterations.
Table II. The ultrasonographic findings of patients with colum- nar cell lesions that underwent ultrasonographic-guided biopsy
Solid 8\22 (36.3%)
Cystic
• Simple cyst
• Complicated cyst (with low level homo- geneous echo)
• Complex cyst
▪With hyperechogenic debris and thick septations
▪With irregular thick wall ▪With solid components
14\22 (63.6%) 1 (4.5%) 3 (13.6%) 10 (45.4%) 4 (18.1%) 4 (18.1%) 2 (9.09%)
these suspicious microcalcifications by US. Seven patients had cysts containing thin septa, while 32 patients had no detectable pathology in the localizations of the suspicious microcalcifications. The MG findings of 12 patients who underwent US-guided biopsy revealed dense breast paren- chymal structure. Moreover, MG showed nodular densities superimposed with the parenchyma in 8 patients who were determined to have solid lesions based on US. Furthermore, US revealed complex cysts in 2 patients, while MG indi- cated scattered round and amorphous microcalcifications.
Ten patients underwent MRI after US (6 patients with complex cyst, 2 patients with cyst with irregular thick walls, and 2 patients with regular hypoechoic solid masses on US and family history of breast cancer). There was no significant relationship between the CCLs and the MRI findings, which are described in Table III.
Fig 4. a. The sonogram shows a mass with regular margin, complex echo texture and parallel orienta- tion, b. The mass is seen as hyperintense on the T2 weighted-image (arrow), c. The mass enhancement in the outer quadrant of the left breast on the contrast-enhanced T1-weighted 3D spoiled gradient-echo axial image, d. Kinetic curve assessment of the mass shows a fast initial increase and then a plateau in the post-initial phase. Surgical excision pathology revealed columnar cell changes.
Fig 5. The sonogram depicts a mass with oval shape, lobular contour, complex echo texture, internal hyperecogenic component and parallel orientation. Due to the increase in size, surgi- cal excision was performed. Pathology showed columnar cell hyperplasia with atypia.
BI-RADS 4b category was the most frequently de- tected in patients with CCLs (43.05%, 31/72). The fre- quencies of BI-RADS 4a, 4b and 4c were not signifi- cantly different in patients with CCLs with and without atypia (p: 0.867) (Table IV).
Discussions
Results of this study show that amorphous-indistinct and cluster-forming microcalcifications were the most frequently seen in the mammograms of patients with CCLs. The radiological findings of patients with CCLs were most frequently associated with the BI-RADS 4b category. No significant difference at US and MG ex- aminations was obtained between the findings of patients with or without CCLs atypia.
Recently, pathological diagnoses of CCLs have in- creased due to a greater number of stereotactic biopsies performed on suspicious microcalcifications [4]. There- fore the number of studies focusing on CCLs is increas- ing. Schnitt et al reported that advanced histological im- aging and surgical treatment are not necessary in CCLs
without atypia [2]. However, excisional specimens of CCLs with atypia must be examined by removing the sections from all incisions to investigate whether there is accompanying ADH and DCIS [7-9]. Goldstein et al reported that 44% of TCs exist with CCLs with atypia [7], while Leibl et al found that 86% of LN cases were accompanied by CCLs [10]. In the current series, there were 10 (10/89 [11.2%]) CCLs with LN, 4 (4/89, [4.4%]) with DCIS, and 2 (2/89, [%2.2]) with TC. These patients were excluded from the study due to their radiological findings. None of the patients in the current study had concomitant neoplasia.
The most significant problem in the follow-up of be- nign breast lesions is the risk of developing malignancy.
In a series reported by Eusibe et al none of the 25 patients with a confirmed FEA developed invasive carcinoma in 19 years of follow-up [11]. Similarly, Bijker et al report- ed that none of the 59 patients with benign breast lesions developed invasive carcinoma in 5 years follow-up [12].
These results indicate that there is a very small risk of FEA progressing to invasive cancer. The median follow- up period of the current series was 18 months; none of the 72 patients with CCLs developed local recurrence or invasive breast cancer during that period.
Recently, pathologists have been reporting CCLs af- ter biopsies of suspicious microcalcifications detected by MG. However, the imaging findings and the nomencla- ture of the CCLs were not investigated in these studies.
Kim et al. reported that of 9 lesions from the MG of 9 patients with CCLs, 5 appeared as clustered amorphous- indistinct, 3 appeared as fine pleomorphic and 1 appeared as round microcalcifications [4]. Solorzona et al reported that in the MG of 32 patients with FEA, the most com- mon morphological feature was amorphousness (13/20 [65%]), and the most common distribution was clus- tered (14/20 [70%]). Other types of microcalcifications reported in Solorzona’s study were fine pleomorphic (2/20[10%]) and coarse heterogeneous (5/20[25%]) [5].
In the current study, the most common microcalcification in CCLs was the amorphous-indistinct and clustered type, which is in agreement with the findings of other studies.
Additionally, in our study, there were a greater number of fine pleomorphic microcalcifications as compared to the results reported by Solorzona et al. We performed MG- guided biopsies because we observed an increase in the number of round microcalcifications in 4 patients during the follow-up period as compared to the previous year.
Patients that had round microcalcifications did not have biopsies during their clinical follow-ups, which might be the reason why less round microcalcifications were de- tected upon follow-up. In our current study, there was no significant difference in terms of the incidence of high- Table III. MRI findings of patients diagnosed with columnar
cell lesions
Findings n (%)
Shape Round
Oval 9 (90%)
1 (10%) Margin
Regular
Lobular 6 (60%)
4 (40%) Enhancement
Peripheric
Homogeneous 1 (10%)
9 (90%) Kinetic curve
Plateau
Wash-out 6 (60%)
4 (40%) T2 weighted
Hyperintense
Hypointense 9 (90%)
1 (10%) CCLs: columnar cell lesions
Table IV. The comparison of histopathological findings in pa- tients with columnar cell lesions according to BI-RADs catego- rization.
BI-RADs CCLs with atypia CCLs without atypia
4A 9 (12.5%) 12 (16.6%)
4B 13 (18.05%) 18 (25%)
4C 8 (11.1%) 12 (16.6%)
Total 30 (41,6%) 42 (58.3%)
p* 0.867
P* < 0,05 was statistically significant, CCLs- columnar cell lesions
risk microcalcifications between patients with CCLs with and without atypia. To our knowledge, there are no stud- ies in the literature comparing the radiologic findings of patients with CCLs with and without atypia.
Upon a review of the literature, we found only two reports of US appearances of CCLs [4,5]. Solorzona et al reported that 9 masses were found in 11 patients (82%), and architectural distortion in 2 patients (18%). Most of the masses displayed an irregular shape (6/9 [67%]) with microlobulated margins (5/9 [56%]) and a hypoechoic or complex echo texture (7/9 [78%]) [5]. Kim et al report- ed in 12 lesions found in 9 patients non-circumscribed masses in 6 lesions, microcalcifications 4 lesions (3 of microcalcifications were simultaneous with masses) and no US focal lesions in another 5 patients [4]. In the cur- rent study, irregular margin and irregular shape were seen in 3% of the lesions, while posterior acoustic shadow- ing was observed in 6 of the patients who underwent US-guided biopsy (27.6%). However, these malignancy criteria were not reported in the studies by Solorzona et al and Kim et al studies [5,6]. Unlike the literature, we observed a simple cyst at US in one patient with CCLs.
FNAC was performed in this patient because the cyst was painful and it increased in size; surgical excisional biopsy was performed because the cytology results in- dicated the presence of suspicious atypia. In our current study, the US findings of patients with CCLs with and without atypia were similar. However, we might not have seen a significant difference due to the small number of patients in our study.
MRI findings of CCLs have not previously been re- ported in the English published literature, only a few re- ports regarding the MRI findings of benign lesions. Sohns et al reported that a plateau of the kinetic curve was seen in 44.2% of cases (42/95 patients), washout was docu- mented in 5.2% of cases (5 patients) and a further increase was seen in 9.4% (9 patients) of cases. Thirty-seven of the benign tumors (38.9%) had a hyperintense appear- ance on T2-weighted images. Eleven of the 14 (78.6%) oval-shaped tumors had a benign histology, and 14 of the 25 (56%) lobulated tumors were categorized as benign [13]. In the current study, no further increase in the ki- netic curve was observed. Only one case had a hypoin- tense appearance on T2-weighted images and most cases had regular and regularly lobular contours. These findings suggest that CCLs show benign features with respect to margin, shape characteristics and their appearance on T2- weighted images. Nevertheless, kinetic curve features can be suspicious and misleading. In addition, MRI can be a confusing method for assessing CCLs since kinetic curve patterns can show plateau or washout. If the radiologist finds a regular mass that is hyperintense on T2-weighted
images and there is a Type 2 or Type 3 kinetic curve, they should consider CCLs in the differential diagnosis.
There were some limitations to this study. Most impor- tantly, the sample size was small and the data was based on the reviews of a single institution. In addition, we did not compare different radiological modalities in terms of the CCLs findings. Further, we might not have been able to see microcalcifications observed in MG with US due to the low density of microcalcifications and the lack of high-definition in devices that we used for investigation.
In addition, a dense breast appearance on MG might have masked the lesions, and therefore, we might not have been able to see the same lesions that we saw in the US. Solor- zona et al reported that US did not detect any evidence of lesions in 23.6% of the patients that were determined to have lesions by MG [5]. Moreover, Kim et al did not detect any lesions by US in 55.5% of patients that were determined to have microcalcifications by MG [4].
Due to the limited number of patients in our current study, there were no significant differences regarding the US and MR findings of CCLs. While MG showed that the clustered distribution pattern and amorphous- indistinct and fine-pleomorphic microcalcifications were more common in patients with CCLs, there was no sig- nificant difference between CCLs with and without atyp- ia in terms of MG findings. The majority of the patients with CCLs were consistent with the BI-RADS 4b catego- ry, but the ratios of CCLs with and without atypia were similar with regards to the BI-RADS categories. Our data indicate that there was no difference in the radiological findings of patients with CLLs with and without atypia;
in such cases, the differential diagnosis should be made by histopathological examination.
Conclusion
Amorphous-indistinct type and clustered distribution pattern microcalcifications are more commonly seen in CCLs. However, CCLs with and without atypia exhibited similar radiological findings. Further studies with more patients are required.
Conflict of interest: none References
1. Jara-Lazaro AR, Tse GM, Tan PH. Columnar cell lesions of the breast: an update and significance on core biopsy.
Pathology 2009; 41: 18-27.
2. Schnitt SJ, Vincent-Salomon A. Columnar cell lesions of the breast. Adv Anat Pathol 2003; 10: 113-124.
3. Pandey S, Kornstein MJ, Shank W, de Paredes ES. Colum- nar cell lesions of the breast: mammographic findings with
histopathologic correlation. Radiographics 2007; 27 Suppl 1: S79-S89.
4. Kim MJ, Kim EK, Oh KK, Park BW, Kim H. Columnar cell lesions of the breast: mammographic and US features. Eur J Radiol 2006; 60: 264-269.
5. Solorzano S, Mesurolle B, Omeroglu A, et al. Flat epithelial atypia of the breast: pathological-radiological correlation.
AJR Am J Roentgenol 2011; 197: 740-746.
6. American College of Radiology. Breast Imaging Report- ing and Data System (BI-RADS) Ultrasound. Reston, VA:
American College of Radiology; 2003.
7. Goldstein NS, O’Malley BA. Cancerization of small ectatic ducts of the breast by ductal carcinoma in situ cells with apocrine snouts: a lesion associated with tubular carcino- ma. Am J Clin Pathol 1997; 107: 561-566.
8. Page DK, Jensen R. Hypersecretory hyperplasia with atypia in breast biopsies. What is the proper level of clinical con- cern? Pathol Case Rev 1996; 1: 36-40.
9. Weidner N. Malignant breast lesions that may mimic be- nign tumors. Semin Diagn Pathol 1995; 12: 2-13.
10. 10. Leibl S, Regitnig P, Moinfar F. Flat epithelial atypia (DIN 1a, atypical columnar change): an underdiagnosed entity very frequently coexisting with lobular neoplasia.
Histopathology 2007; 50: 859-865.
11. Eusebi V, Feudale E, Foschini MP, et al. Long-term follow- up of in situ carcinoma of the breast. Semin Diagn Pathol 1994; 11: 223-235.
12. Bijker N, Peterse JL, Duchateau L, et al. Risk factors for recurrence and metastasis after breast-conserving therapy for ductal carcinoma-in-situ: analysis of European Organi- zation for Research and Treatment of Cancer Trial 10853. J Clin Oncol 2001; 19: 2263-2271.
13. Sohns C, Scherrer M, Staab W, Obenauer S. Value of the BI-RADS classification in MR-Mammography for diagno- sis of benign and malignant breast tumors. Eur Radiol 2011;
21: 2475-2483.
