Objective.-Collagenous spherulosis of the breast is an uncommon localized pattern of basement membrane material deposition that may be mistaken for atypical proliferations or carcinoma. This report describes 9 cases in which the predominant or exclusive appearance of the spherules was basophilic instead of eosinophilic.
Design.-The files of all cases of collagenous spherulosis diagnosed at the Armed Forces Institute of Pathology were reviewed to ascertain the frequency of diagnosis.
Results.-Spherulosis with a predominantly basophilic pattern had a histochemical and immunohistochemical
profile similar to collagenous spherulosis and was associated with more collagenous-appearing forms in 7 of 9 cases. Review of 81 cases showed that collagenous spherulosis was correctly diagnosed in 15% of referrals and was mistaken for intraductal or invasive carcinoma in 11% of cases.
Conclusions.-Mucinous and collagenous patterns appear to be related forms of spherulosis. They are underrecognized by pathologists and maybe mistaken for atypia or malignancy.
(Arch Pathol Lab Med. 1999;123:626-630)
Collagenous spherulosis (CS) of the breast was first described in 1987 1 and was noted to be associated with benign proliferative processes in that and subsequent reports.2 Subsequent publications have described CS in association with lobular neoplasia,3 and the differential diagnosis with adenoid cystic carcinoma has been reviewed in both the histology1 and cytology4,5 literature. The staining profile described in CS includes a variable positivity for acidic mucins. In examining breast biopsy specimens sent in consultation, we have noted foci of spherulosis predominantly composed of basophilic material and have designated this change as mucinous spherulosis (MS). This study was undertaken (1) to document the morphologic and immunohistochemical characteristics of MS, (2) to examine the association of CS and MS, and (3) to assess the degree of recognition of CS in the 10 years since its description.
MATERIALS AND METHODS
The files of the Armed Forces Institute of Pathology (AFIP) were searched for cases diagnosed as CS of the breast during the period 1987-1997. Eighty-one cases of CS were retrieved. The reports for 6 cases noted the presence of MS or were described as showing a "less mature" form of CS. Three more cases were identified during the study, bringing the total number of study cases to 9. Hematoxylin-eosin-stained slides were reviewed, and additional sections were stained with periodic acid-Schiff and Alcian blue at pH 2.5. Mucicarmine stain was performed on 5 cases. Immunoperoxidase stains using antibodies directed against laminin (Sigma Chemical Co, St Louis, Mo), collagen IV (Dako Corporation, Carpinteria, Calif), smooth muscle actin (Sigma), and alpha^sub 1^-antitrypsin (Dako) were performed.
To ascertain the frequency of diagnosis of CS, all 81 previously diagnosed cases of CS were reviewed. The diagnoses of the referring pathologists were extracted from the referral letter and compared with the diagnoses rendered by AFIP. Categories used were (1) the referring pathologist correctly diagnosed CS, (2) the pattern was noted but not diagnosed as CS, (3) a benign process combined with CS was diagnosed as atypical (4) CS was misinterpreted as intraductal or invasive carcinoma, (5) CS was diagnosed as adenoid cystic carcinoma, and (6) CS was completely overlooked.
RESULTS
The dimensions, focality, and associated pathologic changes in the immediate area of the spherulosis are listed in Table 1. Six lesions were multifocal and 3 were unifocal; the size of the largest focus ranged from 1 to 11 mm. A radial scar was present in 2 cases, 2 cases had sclerosing adenosis, 2 cases had an intraductal papilloma, and 4 cases had lobular neoplasia. One case had atypical intraductal hyperplasia, and 3 had ductal carcinoma in situ (DCIS). An invasive apocrine carcinoma was present in 1 case; the spherulosis in this case was within a duct containing a papilloma and was also present in ducts surrounded by carcinoma.
As in CS, a cribriform pattern was apparent on low-- power examination. In 8 of the 9 study cases, however, the "spherules" within the spaces of the cribriform structures were basophilic rather than eosinophilic and collagenous. One case had a strikingly clear appearance, with some spaces having only faint eosinophilic contents. The spherules did not contain stromal cells. The appearance of the basophilic material was variable in character and distribution (Table 2). The most frequent pattern was a central floccular aggregate with radiating spokes that merged with the scalloped projections at the periphery (Figure 1). Sometimes the material was finely granular and more evenly distributed (Figure 2). Distinct septa, composed of thin strands of eosinophilic material with myoepithelial and occasional epithelial cell nuclei, were present in a radial pattern. The periphery of the sphere was marked by an eosinophilic cuticle of variable thickness and staining intensity, associated with a myoepithelial cell nucleus.
Rather than having a perfect cribriform shape, the edge of the sphere was frequently scalloped with extensions of eosinophilic material from the cuticle toward the spherule. This shape was most apparent in larger lesions surrounded by a number of myoepithelial cells; smaller spherules surrounded by 1 or possibly 2 myoepithelial cells tended to have smoother contours. With increasing eosinophilia of the central material, myoepithelial cell nuclei were more readily recognizable on hematoxylin-eosin-stained sections. No strongly birefringent material was seen on viewing the foci with polarized light, suggesting that calcium oxalate was not present. Collagenous spherulosis as described by Clement et al1 was also present in the immediate vicinity and admixed with MS in 7 of 9 cases; the content of different spaces in a given cribriform structure varied from mucinous to collagenous in these 7 cases. Of the remaining 2 cases, one contained predominantly clear spherules (Figure 3), and the other contained mucinous spherules only.
Staining with periodic acid-Schiff preferentially enhanced the peripheral cuticle in all cases. More collagenous spherules also showed less intense central staining. Alcian blue stained the clear (Figure 4) and basophilic spheres moderately to intensely in all but 1 case, in which staining was weak. Mucicarmine stain was positive in all cases examined; the staining was distinct but less intense than that seen in true epithelial mucin on the same slide. The presence of myoepithelial cells surrounding the spheres was confirmed by positive staining for smooth muscle actin in all cases (Figure 5). The intensity of staining for collagen IV generally correlated with the degree of eosinophilia seen on the hematoxylin-eosin-stained slides, with clear and basophilic spheres showing weak staining for collagen IV. Laminin staining was successful in 5 of 9 cases and was of weak to moderate intensity. The pattern seen with the antibodies to the basement membrane components differed; collagen IV decorated the radiating fibrils and the peripheral cuticle, whereas laminin stained the cuticle, with little or no staining of the center of the sphere. al-Antitrypsin showed a variable pattern of staining and significant background uptake in most cases. It was felt to be weakly positive in the spherules in 3 cases and moderate to strong with respect to the background in 2 others.
The referring pathologists' diagnoses are shown in Table 3. Twelve (15%) of the 81 cases had been given a diagnosis of CS, and in another 7 cases (8.5%), a cribriform or hyaline pattern had been noted and was believed to be benign, but was not diagnosed as CS. In 14 cases (17%), CS had contributed to a diagnosis of atypia. The most common combination leading to diagnostic problems was the presence of CS associated with intraductal hyperplasia, which was diagnosed as atypical intraductal hyperplasia in 9 cases. Collagenous spherulosis with atypical epithelium (either ductal or lobular) adjacent to the spherules was diagnosed as DCIS, and sclerosing adenosis involved by CS was interpreted as atypical in 2 cases. Five cases (6%) were diagnosed as DCIS or invasive carcinoma of different types, including mucinous and secretory, and an additional 4 cases (5%) were diagnosed as adenoid cystic carcinoma. In 39 cases (48%), CS was apparently overlooked, as it was neither noted nor misinterpreted as anything else.
COMMENT
Collagenous spherulosis of the breast was first described as an entity by Clement et al1 in 1987. In the intervening decade, a small number of reports have described CS in the breast2,6 and in a number of salivary gland tumors.7 Similar structures have been described in chondroid syringomas of the skin,8 but these were not surrounded by actin-positive cells. Collagenous spherulosis has been detected in fine-needle aspirate material from the breast,4,5 and the differential diagnosis with adenoid cystic carcinoma has been stressed. Despite these reports, our study suggests that this entity is underrecognized by pathologists, with CS correctly diagnosed in only 15% of the cases we studied. Of more concern, however, is the frequency with which CS was diagnosed as atypical or malignant-28% of cases in this study. Failure to recognize CS as a benign lesion may result in overestimation of a patient's subsequent risk of invasive cancer, or at worst may cause inappropriate treatment on the basis of this diagnosis.
The coexistence of collagenous and mucinous types of spherulosis, the suggestion of forms intermediate between these patterns, and the similar immunohistochemical findings in MS and CS suggest that they reflect different stages of the same disease process (MS possibly representing an earlier form of CS). The spectrum of appearances adopted by the spherules in CS has been described in previous reports. In the original description, Clement et all noted some structures with hollow cores and a thin eosinophilic cuticle, others with central spherules and thin radiating arms, and a variable degree of eosinophilia in the central material. Wells et al,2 using an immunohistochemical and ultrastructural analysis, commented that what they referred to as "more mature forms" of spherules contained a more collagenous center and related this to the degree of myoepithelial differentiation of the surrounding cells. Sgroi and Koerner3 observed that the myoepithelial cells occasionally enclosed "faintly basophilic, fibrillar ground substances." A pattern similar to what we describe has been illustrated as an "unusual myxoid variant" of CS.9 Thus, in most reports, CS has been described with an overall eosinophilic quality, consonant with the tinctorial properties of collagen on routine hematoxylin-eosin staining. This may, however, hinder recognition of what we have shown to be a related process but with a predominantly or exclusively basophilic appearance to the spherules. We have used the term mucinous spherulosis in this context. To encompass the range of features seen, we suggest that the lesion be described as spherulosis, with the predominant type described as collagenous or mucinous. The term myxoid may more accurately describe an accumulation of connective tissue mucopolysaccharides. Our personal preference is for the term mucinous, as we feel that this word best describes a localized accumulation of basophilic material (which is also mucarmine positive) that has an apparent intraductal location. Many myxoid areas in connective tissue are poorly defined and contain admixed stromal cells, neither of these features is seen in MS.
Basophilic mucin may be present in a variety of intraductal proliferations. These may be separated from spherulosis by close attention to the pattern and distribution of the basophilic material, the shape and margin of the space in which this is present, and most importantly the nature of the adjacent cells. Compared with intraductal hyperplasia (Figure 6) or atypical intraductal hyperplasia, the basophilic material in spherulosis is arranged in a more structured pattern, and the radial spokes reach to the periphery on a more regular basis. The mucin in intraductal or atypical intraductal hyperplasia is frequently eosinophilic, but may be basophilic. It is coarser, more irregular in distribution both within and between spaces, and tends to retract to the center without a uniform radial spoke pattern peripherally. A characteristic of intraductal hyperplasia is the slitlike shape and peripheral distribution of the secondary lumens; an orderly distribution of rounded spaces along with a more rigid "Roman bridge" pattern is associated with atypical hyperplasia or carcinoma. The spaces of spherulosis are evenly distributed across a given duct space and have a rigid appearance, simulating cribriform DCIS. The potential for mistaking MS for the cribriform pattern of carcinoma is therefore enhanced, particularly when nearby ductal cells show even slightly atypical features. This can be avoided by searching for the thin cuticle that is variably present in cases of spherulosis. Identification of myoepithelial cells is easier in CS than in the mucinous or clear cells types, as the accumulation of basement membrane material and collagen makes their nuclei more apparent. In MS, these cells can only be appreciated in many cases by recognizing a change in the orientation of the cells adjacent to the spherule, a longitudinal orientation with respect to the spherule, betraying a myoepithelial phenotype. Confirmation of this is possible with immunoperoxidase stain for actin, although if suspected, a careful search will usually reveal more advanced lesions resembling CS in the same case.
Collagenous and muanous spherulosis are related lesions derived from a progressive accumulation of extracellular material, including mucopolysaccharides, collagen IV, and laminin, and transformation of the mucinous spherules of the early stage to the collagenous spherules of the end-stage lesion. The close proximity of the myoepithelial cell to the spherules suggests that it is the source of these products, and this suggestion is supported by the identification of CS in salivary gland tumors rich in myoepithelial cells.7 The association of such spherules with myoepithelial cells has been observed previously,lo although the authors of that article appeared to regard mucinous spherules as degenerate collagenous spherules, rather than as their precursors. The stimulus for spherule formation is unclear, but the association with radial scars1 suggests that localized patterns of fibrosis may serve as a promoting environment. Two of our cases had an associated radial scar, 2 contained papillomas, and an additional case showed sclerosing adenosis. These processes frequently result in sclerosis and myoepithelial cell hyperplasia. The myoepithelial cell contains a number of serine protease inhibitors, including aI-antitrypsin and maspin, and (alpha^sub 1^-antitrypsin has also been located in the extracellular matrix.11 Although localized overexpression of such substances is an attractive mechanism for the formation of spherules, we were unable to demonstrate a convincing pattern of selective staining of the spherules with this antibody in this study. An alternative hypothesis regarding the formation of spherules has recently been proposed.12 Based on light microscopic and ultrastructural observations in 2 cases, stromal invagination was suggested as the mechanism of formation of CS. We found spherules abutting on stroma in 5 out of 9 of our cases, but these appeared to be primarily in ducts or terminal ducts/lobular units and were intimately associated with epithelial and myoepithelial cells in all cases, rather than with stroma. We feel that the circumscription of the spherules, the presence of a peripheral cuticle, and the frequent identification of a compressed myoepithelial cell nucleus between the spherule and the adjacent breast stroma, in addition to the spectrum of mucinous and collagenous features seen in the spherules and not the stroma, favor the interpretation of spherulosis as the result of an intraductal proliferation.
References
1. Clement PB, Young RH, Azzopardi )G. Collagenous spherulosis of the breast. Am / Surg Pathol.1987:11:411-417.
2. Wells CA, Wells CW, Yeomans P, Vina M, Jordan S, d'Ardenne AJ. Spherical connective tissue inclusions in epithelial hyperplasia of the breast ("collagenous spherulosis"i. J Clin Pathol.1990:43:905-908.
3. Sgroi D, Koerner FC. Involvement of collagenous spherulosis bv lobular carcinoma in situ: potential confusion with cribriform ductal carcinoma in situ. Am J Surg Pathol.1995;19:1366-1370.
4. Highland KE, Finley IL, Neill JSA, Silverman JF. Collagenous spherulosis: report of a case with diagnosis by fine needle aspiration biopsy with immunocytochemical and ultrastructural observations. Acta Cytol. 1993;37:3-9.
5. Perez JS, Perez-Guillermo M, Bernal AB, Bermejo MR. Diagnosis of collagenous spherulosis of the breast by fine needle aspiration cytology: a report of two cases. Acta Cytol. 1993;37:725 728.
6. Grignon DJ, Ro JY, Mackay BN, Ordonez NG, Ayala AG. Collagenous spher
ulosis of the breast: immunohistochemical and ultrastructural studies. Am I Clin Pathol. 1989;91386-392.
7. Skalova A, Leivo I. Extracellular collagenous spherules in salivary gland tumors: immunohistochemical analysis of laminin and various types of collagen. Arch Pathol Lab Med. 1992;116:649-653.
8. Argenyi ZB, Balogh K. Collagenous spherulosis in chondroid syringomas. Am) Dermatopathol.1991;13:115-121.
9. Rosen PP. Rosen's Breast Pathology. Philadelphia, Pa: Lippincott-Raven; 1997:76
10. Hamperl H. The myoepithelia (myoepithelial cells). Curr Topics PathoL 1970;53:161-220.
11. Barsky SH, Sternlicht MD. Myoepithelial cell: a host defense against invasion. In: Silverstein MJ, ed. Ductal Carcinoma In Situ of the Breast. Baltimore, Md: Williams & Wilkins; 1997:65-76.
12. Maluf HM, Koerner FC, Dickersin GR. Collagenous spherulosis: an ultrastructural study. Ultrastruct Pathol. 1998:22:239-248.
Eoghan E. Mooney, MB, MRCPath; Naila Kayani, MD; Fattaneh A. Tavassoli, MD
Accepted for publication February 5, 1999.
From the Department of Gynecologic and Breast Pathology, Armed Forces Institute of Pathology, Washington, DC (Drs Mooney and Tavassoli), and Faculty of Health Sciences, Aga Khan University, Karachi, Pakistan (Dr Kayani).
Reprints: Fattaneh A. Tavassoli MD, Department of Gynecologic and Breast Pathology, Armed Forces Institute of Pathology, 14th St and Alaska Ave NW, Washington, DC 20306-6000.
Copyright College of American Pathologists Jul 1999
Provided by ProQuest Information and Learning Company. All rights Reserved