| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Breast Imaging |
1 From the Department of Radiology, Johns Hopkins University School of Medicine, Baltimore, Md (H.K.G.), and Departments of Pathology (O.B.I.) and Radiology (W.A.B.) and Greenebaum Cancer Center (W.A.B.), University of Maryland, 419 W Redwood St, Suite 110, Baltimore, MD 21201. From the 2000 RSNA scientific assembly. Received April 18, 2002; revision requested June 19; final revision received November 6; accepted November 27. Address correspondence to W.A.B. (e-mail: waberg@umaryland.edu).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
MATERIALS AND METHODS: From a database of 1,166 percutaneous biopsies in which sclerosing adenosis was reported, 88 (7.5%) lesions were identified, and imaging and histopathologic findings were reviewed for concordance. Sclerosing adenosis proved to be a minor component at core biopsy for 44 lesions, including one invasive ductal carcinoma, one ductal carcinoma in situ (DCIS), one focus of atypical ductal hyperplasia (ADH), and one atypical lobular hyperplasia. Sclerosing adenosis was a major (≥50%) component for 44 lesions, including four malignancies, all DCIS manifested as clustered calcifications (pleomorphic [n = 2] or amorphous [n = 2]), and seven foci of ADH manifested as amorphous calcifications. In 30 patients with 33 lesions without atypia or malignancy, sclerosing adenosis was the major finding at core biopsy (21 lesions at 14-gauge core biopsy and 12 at 11-gauge vacuum-assisted biopsy); these patients formed the study population. Mammographic (33 lesions) and sonographic (18 lesions) features were recorded. Twenty-seven lesions had at least 20-month follow-up (n = 25) or excision (n = 2).
RESULTS: One spiculated mass was considered discordant and was excised, showing a prospectively unrecognized radial sclerosing lesion with several 2–5-mm foci of invasive tubular and lobular carcinoma. Seventeen (53%) of 32 lesions manifested as masses; 10 (59%) were circumscribed, five (29%) were indistinctly marginated (one with punctate calcifications), and two (12%) were partially circumscribed and partially obscured (one with amorphous calcifications). Fifteen (47%) lesions manifested as clustered calcifications; nine (60%) were amorphous and indistinct, four (27%) were pleomorphic, and two (13%) were punctate. Of 27 lesions with acceptable follow-up, 26 (96%) were believed to have been accurately sampled at core biopsy. Of six radial sclerosing lesions associated with the original 88 lesions, only three (50%) were prospectively recognized.
CONCLUSION: Sclerosing adenosis is an acceptable result at core biopsy of circumscribed masses and nonpalpable indistinctly marginated masses and for clustered amorphous, pleomorphic, and punctate calcifications. Recognition and reporting of coexistent radial sclerosing lesions is encouraged and may prompt excision. Malignancy can be seen with sclerosing adenosis; core biopsy was accurate in six (86%) of seven coexistent malignancies in this series.
© RSNA, 2003
Index terms: Breast, biopsy, 00.114, 00.1261, 00.1267, 00.12985 • Breast, diseases, 00.31, 00.32, 00.72, 00.815
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
Adenosis and sclerosing adenosis retain the lobular architecture, but it becomes exaggerated and distorted. The involved lobules show an increased number of acini, which become compressed and obliterated by stromal fibrosis in sclerosing adenosis (Fig 1). There is atrophy of the epithelial cells and increase in the number of myoepithelial cells that become very prominent. The process is more pronounced in the center of former lobules, whereas the periphery is characterized by cystically dilated epithelial profiles. Although usual forms of sclerosing adenosis retain lobular architecture, there are instances in which sclerosing adenosis extends into fat or even shows perineural invasion (4). By definition, individual lobular units have at least 50% of their acini involved with the sclerosing process.
|
We sought to determine concordance of imaging findings and a diagnosis of sclerosing adenosis at histopathologic core biopsy and to establish the accuracy of core biopsy when cancer was coexistent.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
|
In 33 of 44 lesions in 30 patients, sclerosing adenosis was the major and clinically most important finding (with no atypia or malignancy) at core biopsy. These patients formed the study population (Fig 2). Of the 33 lesions, 15 were masses that were sampled with a 14-gauge automated biopsy gun (Bard Monopty, Covington, Ga) with sonographic guidance by using a linear-array broadband transducer with center frequency of 10 MHz (Acoustic Imaging, Tempe, Ariz). A median of five specimens were obtained (range, 2–7). Five lesions that manifested as clustered calcifications were subjected to biopsy with stereotactic guidance (LoRad [now Hologic]; DSM, Danbury, Conn) by using a 14-gauge automated biopsy gun (Manan Medical Products, Northbrook, Ill), with a median of six passes (range, five to six). Another 13 lesions (11 clustered calcifications, one spiculated mass, and one circumscribed mass with calcifications) were subjected to biopsy by using stereotactic guidance and an 11-gauge vacuum-assisted directional probe (Mammotome; Biopsys, division of Ethicon Endo-Surgery, Cincinnati, Ohio), with a median of 12 specimens (range, 6–18) obtained circumferentially. Specimen radiography was performed after all biopsies and specimens containing calcifications were inked according to the method of Berg et al (7). For all lesions manifested as calcifications, calcifications were retrieved. Clip placement (Micromark I or II; Biopsys, division of Ethicon Endo-Surgery) was performed coaxially with the probe for all 11-gauge biopsies.
Imaging and Review
The age of patients and any risk factors for breast cancer that are routinely asked in our screening program (family history of breast cancer, personal history of breast cancer, or atypia) were reviewed. Mammograms were obtained for all 33 lesions and were reviewed again by two authors (H.K.G. and W.A.B.) in consensus for size and the Breast Imaging Reporting and Data System, or BI-RADS, features of the lesions subjected to biopsy. Mammographically circumscribed masses recommended for biopsy were either new (n = 2), enlarging (n = 3), palpable (n = 2), or ipsilateral (n = 1) to cancer. Spot magnification views were available for all lesions with calcifications. If both pleomorphic and amorphous calcifications (8) were present, the lesion was classified as pleomorphic. If both amorphous and punctate calcifications were present, the lesion was classified as amorphous. Sonography was performed for 18 mass lesions, and static images were reviewed again by the same two authors in consensus for size, margin analysis, echogenicity, and posterior features according to the draft of the BI-RADS ultrasonographic lexicon (9).
We routinely recommend excision for all malignant and atypical results and when imaging findings are considered highly suspicious for malignancy. We also recommend excision after needle biopsy yields a radial scar or a radial sclerosing lesion due to a potential risk of associated malignancy (10,11). Six-, 12-, and 24-month follow-up is recommended for all other results. Of the 33 lesions, one was excised due to a highly suspicious mammographic finding and one was excised as part of a mastectomy for the ipsilateral invasive ductal cancer with associated sclerosing adenosis. After biopsy, imaging follow-up of at least 20 months (mean, 36 months; range, 20–66 months) was available for 25 (81%) of the 31 remaining lesions. The six lesions with no follow-up were four nonpalpable circumscribed masses, one indistinctly marginated mass, and one cluster of amorphous calcifications believed to have been removed at initial biopsy.
Concordance
A lesion was considered well sampled at core biopsy and concordant with the diagnosis of sclerosing adenosis in the following circumstances: We required the absence of atypia or malignancy at histopathologic examination; sclerosing adenosis was the most important finding. We deliberately sampled to include the edge of masses; for masses, the interface between a lesion and fat had to be seen at histopathologic examination. The sclerosing adenosis had to appear as a discrete process, representing 50% or more of the material present, with the lesion diameter at histopathologic examination within 2 mm of the diameter at mammography and/or sonography (or occupying the entire specimen for masses >2 cm in diameter). For calcifications, several situations applied. For foci less than 10 mm in diameter, the entire lesion was believed to have been removed or nearly removed; for foci 10 mm or greater, the equivalent of at least two specimens had to contain at least five calcifications each at specimen radiography. In each scenario, equivalent calcifications had to be identified at histopathologic examination in areas of sclerosing adenosis. For all lesions with calcifications, sclerosing adenosis had to represent at least 50% of the lesion present at histopathologic examination. Finally, if the imaging findings were highly suggestive of malignancy, and core biopsy findings showed only sclerosing adenosis, discordance was suspected and excision was recommended.
Malignancies
We separately analyzed all final diagnoses of malignancy with coexisting sclerosing adenosis in the context of other imaging and histopathologic findings at review of the original 88 lesions. For lesions manifested as calcifications, we examined whether calcifications were in malignancy, sclerosing adenosis, neither, or both.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
Twenty-seven lesions had follow-up after initial biopsy. Of 14 foci of calcifications with follow-up, 11 (79%) were gone and three (21%) had stable residual calcifications. Thirteen masses had follow-up. One spiculated mass was excised because of a highly suspicious imaging appearance and yielded several small foci (≤2 mm in diameter) of invasive carcinoma with tubular and lobular features adjacent to and involving a radial sclerosing lesion (evident retrospectively at the core biopsy, Fig 3). This represents a false-negative core biopsy finding. Another indistinctly marginated mass ipsilateral to cancer was excised at mastectomy and was confirmed to be sclerosing adenosis. Eight masses (four circumscribed, three indistinctly marginated, and one partially obscured) were stable, and two circumscribed masses and one partially obscured mass had increased in diameter. The patients chose continued follow-up despite recommendation for repeat biopsy. One mass doubled in largest diameter in the first 3 years of follow-up but then remained stable for 2 additional years. Two masses showed 20% increase in the largest diameter in the first 2 years of follow-up and then remained stable for 2 and 3 additional years.
|
|
|
|
|
|
|
|
|
|
|
Accuracy
Of the 27 lesions with acceptable follow-up, 26 (96%) were believed to have been adequately sampled at core biopsy.
Malignancies
Overall, of 88 lesions with sclerosing adenosis, seven (8%) proved to be malignant. Of seven malignancies, five (71%) were clustered calcifications in both DCIS and sclerosing adenosis. Two foci of DCIS manifested as pleomorphic calcifications; two, as amorphous; and one, as punctate calcifications. The two invasive cancers were masses; one was spiculated (Fig 3), and one was palpable and indistinctly marginated and seen only at sonography. Of seven coexistent malignancies, six (86%) were accurately diagnosed at core biopsy.
Of the 88 lesions, six (7%) had an associated radial sclerosing lesion, and only three (50%) of these six were prospectively recognized. As described, the one spiculated mass had an associated prospectively unrecognized radial sclerosing lesion at core biopsy and was excised, yielding malignancy.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
Several observations may help explain the increased risk of breast cancer with sclerosing adenosis. Findings of one study (15) demonstrated increased estrogen receptor expression, as well as Ki-67 in sclerosing adenosis, radial scars, papillomas, fibroadenomas, and phyllodes tumors but not in apocrine cysts. Shoker et al (15) suggest that the degree of dysregulation of estrogen receptor correlates with the degree of risk of developing breast cancer. Increased estrogen receptor expression in proliferative lesions may also help explain the observation of Cahn et al (16): Predisposing lesions (sclerosing adenosis, papilloma, or ADH) were seen in 36 (63%) of 57 women receiving hormonal replacement therapy, whereas only 30 (30%) of 99 women with nonproliferative lesions were treated with hormonal replacement therapy (odds ratio, 3.9).
Sclerosing adenosis can coexist with cancer, as was seen in 8% of lesions in this series. For six (86%) of the seven lesions, the coexistent cancer was successfully sampled, and, to our knowledge, for the other 80 benign lesions the diagnosis was accurate, with an overall accuracy of 99% (87 of 88 lesions). For the 27 lesions with sclerosing adenosis as the major finding at core biopsy and an acceptable follow-up, 26 (96%) were accurately sampled. One mammographically spiculated mass had in retrospect a radial sclerosing lesion at core biopsy that was associated with sclerosing adenosis and was found to have several 2-mm or smaller foci of infiltrating carcinoma with tubular and lobular features in and around the radial sclerosing lesion at excision (false-negative core biopsy finding). In the series of Nielsen (1), five (19%) of 27 adenosis tumors were associated with DCIS. Carcinoma can be a minor component of sclerosing lesions, as was the case in five (71%) of seven cancers in this series. Westenend and Liem (17) recently reported one false-negative core biopsy diagnosis with missed associated DCIS and invasive carcinoma in a 6-cm mass, which showed only sclerosing adenosis at core biopsy.
Sclerosing adenosis can be difficult to distinguish from infiltrating carcinoma both mammographically and microscopically. In the series of Tinnemans et al (18), two patients underwent unnecessary mastectomy on the basis of interpretation of sclerosing adenosis on a frozen section as cancer. In the series of 27 adenosis tumors reported by Nielsen (1), three patients underwent unnecessary mastectomy. Distinction between adenosis and infiltrating carcinoma is based on the presence (and hyperplasia) of an outer myoepithelial cell layer in adenosis and its absence in invasive cancer. A useful adjunct in such a differential diagnosis is the use of immunostaining for smooth muscle actin in myoepithelial cells and thereby in sclerosing adenosis but not in invasive cancers (19).
Involvement of sclerosing adenosis by in situ carcinoma can be extremely difficult to distinguish from invasive carcinoma (20,21). This has been reported with both ductal (22,23) and lobular carcinomas in situ (24–26). Staining for smooth muscle actin to demonstrate myoepithelial cells or for collagen type IV or laminin to demonstrate preservation of basement membrane can help, as both are lost in invasive lesions (22,23).
Sclerosing adenosis can be part of a papilloma, a radial scar, or the larger complex sclerosing lesion. Indeed, associated complex or radial sclerosing lesions were seen in six (7%) of 88 lesions in this series. While radial sclerosing lesions can be incidental findings, when depicted mammographically they are usually seen as spiculated masses or areas of isolated architectural distortion (27). Across several series (10,11,27), small associated carcinomas (usually DCIS or tubular cancer) were seen in 48 (22%) of 221 excised radial sclerosing lesions and another 39 (18%) lesions had associated ADH or lobular carcinoma in situ.
The management of core biopsies yielding radial sclerosing lesions remains somewhat controversial. Liberman’s review (28) revealed that three (13%) of 23 lesions yielding a radial sclerosing lesion at core biopsy proved to be malignant at excision. Reynolds (29) found that 43% of radial sclerosing lesions that were visible mammographically as spiculated lesions or masses with architectural distortion proved to be malignant, compared with a 1.2% malignancy rate when a radial sclerosing lesion was an incidental adjacent finding at histopathologic examination when core biopsy was performed for another lesion. While sclerosing adenosis can cause a spiculated mass both in this setting and without an associated radial sclerosing lesion (30), the risk of missing an associated carcinoma would seem to justify excision when a spiculated mass yields only sclerosing adenosis at core biopsy, as in the one missed cancer in our series. It is not clear at this time that excision is required when calcifications have been well sampled percutaneously with a benign concordant result and an incidental radial scar is identified at histopathologic examination. In this series, in the one case of pleomorphic calcifications yielding sclerosing adenosis with a radial scar, in retrospect, the lesion was believed to have been completely removed, with no residual lesion at 24-month follow-up. Philpotts et al (31) found no malignancies among eight radial scars diagnosed at stereotactic biopsy of calcifications. Brenner et al (32) reviewed findings of 157 core biopsies yielding a radial scar with 24 months of follow-up. Of 29 lesions with associated ADH, eight (28%) proved to be malignant (32). Of 128 lesions without associated atypia, five proved to be malignant, including two (3%) of 60 masses and three (8%) of 40 architectural distortions, with none of 28 calcifications without atypia proving to be malignant (32). No associated malignancies were missed when sampling was performed with a directional vacuum-assisted device and when at least 12 specimens were obtained (32).
If the central elastosis of a radial sclerosing lesion is not adequately demonstrated in core specimens, only sclerosing adenosis may be seen. Indeed, the one cancer missed at core biopsy in this series did have an associated radial sclerosing lesion evident at second review (Fig 3). Recognition of the radial sclerosing lesion at core biopsy can be problematic because the overall architecture is not always evident. Indeed, in our study, three (9%) of 33 core biopsies with sclerosing adenosis as the major finding proved to have an unreported radial sclerosing lesion in retrospect, and only three (50%) of six associated radial sclerosing lesions were prospectively recognized. When present at core biopsy, it appears important that the pathologist attempt to recognize and report a radial sclerosing lesion, as it may affect the decision to excise the lesion.
Fibroadenomas can show a spectrum of associated fibrocystic lesions, including sclerosing adenosis. These fibroadenomas are often referred to as complex and carry an increased risk of malignancy, as much as threefold in either breast (33). Indeed, when we reviewed lesions initially interpreted as sclerosing adenosis, we found 10 (11%) of 88 were seen within fibroadenomas and six (7%) had small adjacent papillomas not reported initially. Shabtai et al (34) found 23% of fibroadenomas with associated sclerosing adenosis. We are unable to specifically address the frequency of sclerosing adenosis in all fibroadenomas in our database.
Isolated sclerosing adenosis and simple adenosis have been reported in association with amorphous indistinct calcifications (35,36), ADH (35,36), and typically low-grade DCIS (36,37). Further, all of these entities can coexist. In one large series (12), atypical lobular hyperplasia was found to be 2.7 times more common in sclerosing adenosis lesions than in other biopsy findings. Clustered amorphous calcifications were the most common type associated with sclerosing adenosis in this series, though pleomorphic and punctate calcifications were also seen. Provided the calcifications have been well sampled at core biopsy, sclerosing adenosis is an acceptable result for such lesions. As described by Preece (38), sclerosing adenosis can also be evident mammographically as scattered bilateral punctate and amorphous calcifications, though such findings are usually considered benign or probably benign and do not generally result in biopsy (39,40). We do not accept a diagnosis of sclerosing adenosis for core biopsy of calcifications of any morphology in a segmental or linear distribution nor for individually linear or fine-branching calcifications typical of high-grade DCIS (37,41–43).
Both in the series of Liberman et al (44) and that of Philpotts et al (45), no underestimation of disease occurred when all calcifications were believed to have been removed at 11-gauge biopsy. In the series of Berg et al (36), three of 113 stereotactic biopsies yielding amorphous calcifications were underestimates (97% accuracy). Since completion of this analysis, we found one 4-mm focus of amorphous calcifications in a linear distribution, which had been only minimally sampled at stereotactic vacuum-assisted biopsy (with two calcifications in one core specimen) that yielded sclerosing adenosis; this proved to be DCIS at excision.
Masses caused by nodular sclerosing adenosis have been previously reported (1,6,46,47). In the series of DiPiro et al (6), seven (70%) of 10 such masses were circumscribed at mammography and five (71%) of seven were oval at sonography. None were spiculated. Nielsen and Nielsen (46) reported that sclerosing adenosis could present as a circumscribed or stellate mass or irregular density with or without calcifications. As mentioned, Franquet et al (30) also describe sclerosing adenosis as a potential cause of a spiculated mass. Again, we would suggest that the risk of an associated and unsampled cancer at core biopsy in the setting of a spiculated mass justifies excision. We do accept sclerosing adenosis as concordant for circumscribed or indistinctly marginated masses. Given the potential risk of associated DCIS and even invasive cancer, it would also seem prudent to recommend excision or at least rebiopsy of subsequently enlarging masses that yield only sclerosing adenosis at core biopsy, though this requires further study. The three patients in this series with an enlarging mass that was previously sampled, yielding sclerosing adenosis, have declined surgical intervention or rebiopsy, and the masses did subsequently stabilize at longer term follow-up.
In summary, careful review of the imaging and histopathologic findings is required after a core biopsy diagnosis of sclerosing adenosis. Such a result is not acceptable for lesions with high suspicion of malignancy, including spiculated masses, branching or fine linear calcifications, or calcifications in a segmental or linear distribution. Recognition of associated radial sclerosing lesions may be problematic; if present, these should be reported, and excision may be appropriate for masses or architectural distortion in association with radial sclerosing lesions. Excision or rebiopsy may also be appropriate for subsequently enlarging masses or increasing calcifications that yield only sclerosing adenosis at core biopsy because of the potential association with high-risk lesions, DCIS, and invasive cancer. Sclerosing adenosis is an acceptable, concordant result at core biopsy of circumscribed masses and nonpalpable indistinctly marginated masses and at vacuum-assisted biopsy of well-sampled clustered punctate, amorphous, and/or pleomorphic calcifications.
|
FOOTNOTES |
|---|
Author contributions: Guarantor of integrity of entire study, W.A.B.; study concepts and design, W.A.B.; literature research, W.A.B., H.K.G., O.B.I.; clinical studies, W.A.B., H.K.G., O.B.I.; data acquisition, W.A.B., H.K.G., O.B.I.; data analysis/interpretation, W.A.B., H.K.G.; manuscript preparation and editing, W.A.B.; manuscript definition of intellectual content, W.A.B., O.B.I.; manuscript revision/review and final version approval, W.A.B., H.K.G., O.B.I.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  M. Guray and A. A. Sahin Benign breast diseases: classification, diagnosis, and management. Oncologist, May 1, 2006; 11(5): 435 - 449. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
