- Email: [email protected]
Pseudoangiomatous hyperplasia of mammary stroma
Pseudoangiomatous Hyperplasia of Mammary Stroma MILAN F. VUITCH, MD,* PAUL PETERROSEN, MD, AND ROBERTA ERLANDSON, PHD Grossly c i r c u m s c r i b e d , n o n h e m o r r h a g i e b r e a s t masses consisting of m a m m a r y stromal proliferations that simulated vasc u l a r l e s i o n s were studied in n i n e w o m e n . Histologically, a striking p a t t e r n , w h i c h a p p e a r e d to consist o f complex intera n a s t o m o s i n g c h a n n e l s l i n e d b y s l e n d e r s p i n d l e cells, was present in the m a m m a r y parenchyma. T h e importance of this benign lesion, referred to as pseudoangiomatous hyperplasia of m a m m a r y stroma, is its distinction from angiosarcoma. The patients ranged in age from 22 to 52 years; al! were premenopausal. Each presented with a palpable unilateral mass, measuring up to 7 cm in diameter. The patients were treated by excisional biopsy and remained well for up to 2.5 years after excision. One patient had two local recurrences within one year of the original excision, and a second patient had a local recurrence at 14 months. No patient had another concurrent or metachronous malignant tumor of the breast or other organ, a n d no abnormal hormonal status was found. Complete local excision appears to be adequate treatment. It remains to be determined whether this is a neoplastic process. However, there is no evidence that it is a precursor of angiosarcoma, and ultrastructural observations demonstrate that the spaces found in the lesion are not true vascular channels. Rather, they appear to arise by a process that involves disruption and separation of stromal collagen fibers. Since small loci of this c h a n g e are c o m m o n in h y p e r p l a s t i c breast tissue from premenopausal women, it is likely that the development of a discrete t u m o r with this p a t t e r n r e p r e s e n t s an exaggerated form of stromal hyperplasia. HuM PATnOL 17:185--191, 1986.
Vascular tumors of the breast were tile subject of a series of previous reports. 1-3 Among the lesions included in our studies of vascular tumor were seven breast masses characterized by m a m m a r y stromal proliferation with the f o r m a t i o n o f interanastomosing slits that simulated vascular spaces. These cases were most often submitted with the provisional diagnosis of low-grade angiosarcoma. Tile recent opportunity to examine two such tumors from patients treated at Memorial Sloan-Kettering Cancer Center (MSKCC) made possible more detailed studies of this uncommon lesion, which we call pseudoangiomatous hyperplasia of mammary stroma. MATERIALSAND METHODS In addition to the two patients treated at MSKCC (case 1 and 2), seven others were seen as consultations from other hospitals. Submitted slides stained with hematoxylin-eosin were reviewed in each case. Clin-
ical infornmtion was obtained from the referring pathologists. Wlten sufficient material was available, trichrome, Alcian blue (with and without hyaluronidase), reticulin, and elastic stains were used. Submitted slides were stained by the immunoperoxidse techniques for Factor VIII, Ulex europaeus agglutinin1, S-100 protein, vimentin, myosin, myoglobin, and actin antigens. For electron microscopy, formalinfixed tissue (case 1) was cut into small cubes, washed in normal saline for two days, and postfixed in 1 per cent o s m i u m tetroxide b u f f e r e d with s-collidine. Specimens were then washed in distilled water and left overnight in 0.25 per cent aqueous uranyl acetate to enlmnce contrast. Following dehydration in graded ethanols, tissues were embedded in Maraglas-DER 732 epoxy resin. 4 For orientation, l-tim sections were stained with sodium borate-buffered 1 per cent toluidine blue. Thin sections were stained sequentially with uranyl acetate and lead citrate and examined in a transmission electron microscope.
RESULTS
Clinical Data Details of individual cases are presented in table 1. T h e patients were premenopausal women (age range, 22 to 52 )'ears; mean, 40 )'ears). Tile lesion generally p r e s e n t e d as a discrete, painless breast htmp that was firm or rubber)' in consistency. Aspiration biopsies in three cases failed to yield cells. In one case a mammogram was reported as questionable. All tumors were treated by excisional biopsy, and seven patients had no subsequent difficulties, with follow-up periods extending to 30 months. A 22year-old woman underwent two subsequent excisions in an 11-month period for persistent lesions, and a 49-year-old woman underwent excision of a local recurrence at 14 months. One patient underwent bilateral subcutaneous mastectomies when the lesion was initially diagnosed as an "atypical vasoformative process."
Pathologic Data Grossly, tile tumors were well circumscribed with the smooth external surface resembling a capsnle in some cases. T h e texture was usually described as rubbery. Tile greatest diameter as measured on the excised specimen, ranged from 2 to 7 cm. No tumor had extensive hemorrhage oi" necrosis. Microscopicall)', the lesions were composed o f intermixed
Received from the Department of Pathology, Memorial tlos3ital, Memorial Sloan-Kettering Cancer Center, New York, New York. Revision accepted for publication August 6, 1985. *Recipient of an American Cancer Society Fellowship. Address correspondence and reprint requests to Dr. Rosen: ) e p a r t m e n t of Pathology, Memorial ttnspital, 1275 York Avenue, qew York, NY 10021.
185
HUMAN PATHOLOGY TABLE t.
Case
Patient's Age ()'r)
1
38
2
52
3
44
4
35
5
22
6
44
7
37
8
35
9
49
Clinical Features
Volume 17, No. 2 (February1986) Summary of Cases
Size (cm)
Gross Description
"Fibroadenoma," 1 too; aspiration unsuccessful Lump, 3 wk; mammogram questionable; 2 aspirations unsuccessflfl Lump; bilateral subcutaneous mastectomies at 30 mo Lump
7 • 5 • 4
Fibroadenonm
5 x 4 x 3
Encapsulated, firm, lobulated Fibrous
3 • 2
Rubbery
Fibrocystic disease with atypical vasoformative process
7 x 5 • 4
Rubber)', nodular
Lump; recurrences at 6 and 13 mo Painless lump
2
Giant adenofibroma, cystosarcoma phylloides ruled out Fibroadenoma variant
"Thickening," aspiration unsuccessfnl Lump, 3 wk
3
Ovoid, rubbery, wellcircu inscribed Well-circumscribed nodule I11 defined
6 • 4 • 3
Lobulated, firm
Painless lump; 5 too; recnrrence at 15 ino
3
Ovoid, circumscribed
3
stromal and epithelial elements. In general, tile epithelium within the lesions was essentially normal, with only mild patchy hyperplasia or duct papillomatosis. Areas of epithelium with apocrine secretory snouts were observed in three cases, and in two cases myoepithelial cells with rotmded-up clear cytoplasm were present in many lobules. T h e ducts and lobules were composed of cells with relatively uniform nuclei, occasionally several cells thick. Epithelial atypia was not present. Occasionally, attenuated ducts similar to those in fibroadenomas were present. The characteristic and most prominent stromal feature was a complex pattern of interanastomosing empty spaces that was easily seen at low power (fig. I). These channels were present within as well as between breast lobules. In most cases they tended to be a r r a n g e d concentrically immediately a r o u n d the breast lobules (fig. 2). Often, spindly stromal cells were present at tile margins of these spaces, simulating endothelial cells (fig. 3). The impression on low-power examination was that of a complex vascular proliferation, although only very rarely were isolated red blood cells present in these pseudoangiomatous spaces. The stroma also contained abundant collagen, which was focally very dense, and elongated wavy spindle cells with poorly defined cytoplasmic borders. T h e nuclei o f these cells were attenuated and usually had pointed ends. Mitoses were not seen, nor was atypia or p l e o m o r p h i s m present. Abundant, distinct small blood vessels with readily apparent endothelial cells were also seen in the stroma. Frozen section analysis was performed in three cases. It is noteworthy that the stromal pseudoangiomatous spaces were discernible in fi'ozen sections, indicating that they were not an artifact of formalin186
Original Diagnosis
Pseudoangiomatous hyperplasia of mammary stroma
Grade I angiosarcoma Probably benign reactive vascular proliferation Grade I angiosarcoma versus benign angiomatous proliferation Well-differentiated angiosarcoma versus fibroadenomatoid tumor
fixed, p a r a f f i n - e m b e d d e d tissue. Periodic a c i d Schiff stain showed no evidence of basement membrane material near the clear channels. Alcian blue staining within the stromal collagen as well as in the clear pseudoangiomatous spaces was removed by hyaluronidase treatment. T r i c h r o m e staining of collagen was strong, while reticulin and elastic stains were negative except in definite blood vessels. Immunoperoxidase stains for Factor VIII protein and Ulex europaeus agglutinin- 1 stained endothelial cells of definite blood vessels, but not tile stromal cells adjacent to the clefts (fig. 4). Tissue from case 1 was examined by electron microscopy. The specimen consisted of epithelial ducts in a dense fibrovascular stroma. The ducts were lined by three to four tiers of ductular epithelial cells separated from the stroma by a single discontinuous layer o f myoepithelial cells resting on a basement membrane. T h e epithelial-stromal junction was intact. Only tile fine structure of tile stromal elements is detailed in this report. This abundant fibrovascular tissue consisted of capillaries (small arterioles and venules were also observed), vascular pericytes, modified stronml pericytes, fibroblasts, many o f which lined pseudoangiomatous spaces, and dense bands of collagen fibrils (figs. 5 and 6). The closed lumina of true blood vessels were lined by endothelial cells joined by tight junctions and surrounded by an adluminal basement membrane (figs. 5 to 7). The basement membrane was occasionally multilayered (fig. 6). The cytoplasm of the endothelial cells contained prominent surface pinocytotic vesicles, a meshwork of 10-nm intermediate filaments, and Weibel-Palade bodies (figs. 6 and 7). All of the vessels were intimately surrounded by a discontinuous layer of pericytes, one to two cells thick, enveloped by a basement
'
.
.
dr~
S
.
~
q~
w
;A ~" T,! " ~ k
f~'qF
t
, ~ "
I
.'~
."
"
" 9
. . . . . . . . . . . . . . .
FIGURE 4 (top left). Case 1. Discrete, seemingly encapsulated 24-g tumor excised from the right breast of a 38-year-old woman. The lowpower view shows pseudoangiomatous stromal proliferation between duct and Iobular units. [ x 40.] FIGURE 2 [top right]. Enlarged view of the complex arrangement of spaces resembling vascular structures that constitute pseudoangiomatous hyperplasia of mammary stroma. ( x ~00.] FIGURE 3 [bottom left). Case 3. Rubbery mass, excised from the left breast of a 44-year-old woman. The high-power view reveals pseudoangiomatous slits, adjacent spindle cells, and intervening dense collagen. Nuclear a~/pia and pleomorphism are absent. [ x 250.] FIGURE 4 [bottom right). Immunoperoxidase staining for Factor VIII antigen; hematoxylin counterstain. Staining is evident in the endothelium of a true blood vessel (top of field] but not in tissue exhibiting pseudoangiomatous stromal hyperplasia. ( x 250.]
187
HUMAN PATHOLOGY
Volume 17, No. 2 (February 1986]
~j:-~
...;' FIGURE 5 (top]. Survey electron micrograph illustrating [from left to right] a portion of a small arteriole, modified pericytes in a dense collagenous stroma, and a pseudoangiomatous space devoid of collagen fibrils. [ x 6,240.] FIGURE 6 [bo#om). Portion of a capillary intimately Surrounded by a discontinuous layer of pericytes. Cytoplasmic processes c o a t e d by remnants of basement membrane, probably from attenuated modified pericytes, are visible in the adjacent collagenous 9stroma. [x13,600.)
cells were distinguished by a fairly well developed rough endoplasmic reticulum and, often, a prominent Golgi apparatus. Pinocytotic vesicles, Weibel-Palade bodies, and basement membranes were not identified (fig. 8, left). The slender cytoplasmic processes o f the fibroblasts t e r m i n a t e d in tile collagenous stroma and were occasionally joined by small rudimentary cell junctions (fig. 8, right). The stroma between the small blood vessels (pre-
membrane. Tile pericytes displayed pinocytotic vesicles and dense attachment plaques on tile outer plasmalemma, distal to the endothelial cells, and contained variable numbers of longitudinally oriented 6nm microfilaments associated with widely scattered fusiform dense bodies (figs. 5 to 7). Pseudoangiomatous spaces, devoid of collagen fibrils and lined by a discontinuous sheath of fibroblastic cells were also evident (fig. 8). The fibroblastic 188
PSEUDOANGIOMATOUS HYPERPLASIAOF MAMMARY STROMA (Vuitch et al.)
FIGURE 7. Detail of the wall of a small capillary (top of field]. The endothelial cell cytoplasm contains numerous Weibel-Palade bodies [arrows], subsurface pinocytotic vesicles, and intermediate filaments. The endothelial ceil is surrounded by portions of three pericyte processes. The distal plasmalemma of the endothelial cell and pericytes is covered by a distinct basement membrane. I x 28,000.]
strnation. A stud)' by Vogel et al. 7 in 1981 confirmed the distinct histologic patterns of breast epithelium and perilobular stroma at different phases of the menstrual cycle. T h o s e investigators r e p o r t e d a "loose, broken" stroma in tile luteal phase (days 15 to 20) and a "loose, edematous" stroma in the secretory phase (days 21 to 27). T h e stroma ill our cases has some similarity to the stroma in tile physiologic luteal and secretory phases of the menstrual cycle. T h e open clefts containing hyaluronidase-sensitive mucopolysaccharide resemble the histologic features shown in 4 and 5 in the report of Vogel et al. 7 We have also seen similar perilobular clefts focally in hyperplastic breast tissue fl'om premenopausal women, sonletimes associated with secretory changes in the acini. Ttle cases of pseudoangionmtous hyperplasia of the mammary stroma presented here differ from the physiologic states in important respects. In only three cases was secretory activity present in some acini, demonstrating a definite stromal-epithelial dissociatlon m most cases. We believe that the stromal changes may originate as a response to progesterone ill estrogen-primed tissue. It appears that tile pseudovascular pattern persists due to unknown factors, suggesting that the stroma becomes autonomons. The stromal pseudoangiomatous cleft formation was
d o m i n a n t l y capillaries) and the pseudoanigiomatous spaces lined by fibroblasts contained longitudinally a r r a n g e d , extremely a t t e n u a t e d long cytoplasmic processes joined by occasional tight or rudimentary cell junctions (figs. 5 and 6). The cytoplasmic processes of these cells, presumed to be modified pericytes, retained fragments of basement membrane on both surfaces, pinocytotic vesicles, and occasional small clusters of intermediate filaments (fig. 9). DISCUSSION Relatively few morphologic descriptions of tile mammary stroma in physiologic and pathologic conditions are available. Mammary stroma is generally accepted as lmving two elements, a specialized hormonally responsive intralobular and periductal component, and a supporting interlobular stroma, which is believed to be largely unresponsive to hormones. In 1945 Foote and Stewart 5 described the histologic changes in the intralobular stroma during the menstrual cycle. Ozzello and Speer 6 correlated those histologic" changes with fluctuations in the hyaluronidase-susceptible mucopolysaccharide content of the stroma, which increased dramatically in the premenstrual phase after day 12 to 15, along with secretory activity of the acini, and decreased following men189
HUMAN PATHOLOGY
Volume 17, No. 2 (February 1986]
FIGURE 8. Left,, pseudoangiomatous space lined by slender fibroblastic cells. The cells contain a well-developed rough endoplasmic reticulum and Golgi apparatus. No basement membrane is evident. (• 10,000.] Right blunt termination of fibroblastic processes lining pseudoangiomatous space. Cell processes are joined by rare rudimentary cell junctions [arrows]. Numerous collagenous fibrils seen in cross and longitudinal sections surround the pseudovascular space. (x 15,000.)
plasmic processes. T h e findings in our case were similar to those of Ahmed. I-~ T h e importance of recognizing pseudoangiomatous stromal hyperplasia of the breast as an entity is to prevent misdiagnosis of the condition as a l o w grade angiosarcoma. T h e features o f these lesions that suggest angiosarcoma are the size of the mass (usually 3 cm or greater) and the low-power microscopic appearance suggesting anastomosing channels or clefts. However, pseudoangiomatous stromal hyperplasia produces a discrete, rounded, often rubbery mass that grossly resembles a fibroadenoma and does not have the henmrrhagic appearance that typifies angiosarcomas. N u m e r o u s histologic features permit differentiation of this lesion from angiosarcoma. In pseudoangiomatous hyperplasia of mammary stroma, pleomorphism an& hyperchromasia of the stromal cell nuclei are minimal to absent, papillary endothelial ingrowths into the clefts are not present, and the channels only very rarely contain red blood cells. Mitoses and necrosis are not seen, nor is destruction o f m a m m a r y epithelial structures evident, even when the process extends into lobtfles. Intermingling of the stromal channels within lobules is
so extensive in our cases that it extended between lobules. We cannot be certain whether interlobular stroma, normally hormone-insensitive, is recruited into the lesion, or whether marked expansion of the intralobular stroma merely displaces interlobtdar stroma. An interesting but unexplored feature o f these masses is whether they contain estrogen and/or progesterone receptors. T h e ultrastructural features of breast stroma have not been investigated intensively. Fanger and Ree, s in their study of the uhrastructural changes in the breast during the menstrual cycle, confined their observations to the epithelium. Ozzello 9 described the normal markedly attenuated "delimiting fibroblasts" around lobules. Most ultrastructural studies of breast stroma have concentrated o n fibroadenomas, with some controversy as to the nature of the stromal cells. According to Archer and Omar, 1~ the stromal cells are fibrocytes, and Ozzello 9 agreed with that conclusion. Murad et al. 11 considered the stromal cells pericytes. A h m e d 12 stated that he found fibroblasts as well as some cells with features of pericytes, i.e., basal lamina, hemidesmosomes, and numerous pinocytotic vesicles, along the plasma membrane of long cyto190
PSEUDOANG1OMATOUS HYPERPLASIAOF MAMMARY STROMA [Vuitch et al.]
.
.
.
.
.
- '..
.7:/
:
-2.-
:
-
.~
.-
.!if
.:~
~z.
~
,.'
.
.~
.~
~-~r
FIGURE 9. Cytoplasmic processes, probably from modified pericytes surrounded by numerous collagen fibrils in the stroma. Pinocytotic vesicles [arrow] and remnants of basement membrane are visible. [ x 56,000.}
t.ypical, frequently with concomitant mild proliferanon of the epithelial structttres. I m m u n o p e r o x i d o s e reagents such as Factor VIII, are not consistently positive in angiosarcoma and are therefore not helpfitl. O n e patient had two local recurrences, a n d a second patient had one recurrence. However, microscopic examination disclosed no greater cellularity or pleomorphism in the tumors in these patients than in those in other patients. T h e r e was no change in cellularity or pleomorphism from the initially excised masses to the recurrences in those two patien.ts. Since the initial masses in the patients with recurrences were the smallest in o u r series, it is likely that excision was incomplete. Altei'natively, the breast stronm adjacent to the excised lesion may become susceptible later to the same u n k n o w n stimulus that causes pseudoangiomatous hyperplasia of m a m m a r y stroma. We r e c o m m e n d c o m p l e t e local excision with clinical t-ollow-up evaluation that probably does not need to extend beyond three )'ears.
REFERENCES 1. Donnell RM, Rosen PP, Lieberman PH, et al: Angiosarcoma and other vascular tumors of the breast: pathologic analysis as a guide to prognosis. Am J Snrg Pathol 5:629, 1981 2. Jocefczyk M, Rosen PP: Vascular tumors of the breast: I1. Perilobular angiomas and angiomas. Am J Surg l'athol 9:491, 1985 3. Rosen PP, Jocefczyk M: Vascular tumor of the breast: III. The venous hemangioma. Am J Surg Pathol 9:652, 1985 4. Erlandson RA: A new Maraglas, DER 732, embedment for electron microscopy. J Cell Biol 22:704, 1964 5. Foote FW, Stewart FW: Comparative studies of cancerous versus noncancerous breasts. Ann Surg 121:6; 197, 1945 6. Ozzello L, Speer FD: The mucopol)saccharides in the normal and diseased breast. Their distribution and significance. Am J Pathol 34:993, 1958 7. Vogel PM, Georgiade NG, Fetter BF, et al: The correlation of histologic changes in the human breast with the menstrual cycle. Am J Pathol 104:23, 1981 8. Fanger H, Ree HJ: Cyclic changes of human mammary gland epithelium in relation to the menstrual cycle--an uhrastructural study. Cancer 34:574, 1974 9. Ozzello L: Ultrastructnre o f the human mammary gland. Pathol Annu 6:1, 1971 10. Archer F, Omar M: The fine structure of fibro-adenoma of the human breast. J Pathol 99:113, 1969 I 1. Murad TM, Greider MH, Scarpelli DG: The uhrastructure of human mammary fibroadenoma. Am J Pathol 51:663, 1967 12. A h m e d A: T h e myoepithelium in human breast fibroadenoma. J Pathol 114:135, 1974
Acknowledgments. T h e a u t h o r s a c k n o w l e d g e the con:ribution o f cases a n d clinical i n f o r m a t i o n flora the folo w i n g p a t h o l o g i s t s : J . E. Boline, R. Chapwick, E. L. Coffey, [r., S. Dias, M. P. H y m a n , J. A. R u d d , and A. S e y m o u r . ? h o t o g r a p h s w e r e p r e p a r e d by Roy K e p p i e , Kin K o n g , m d M a r g a r e t Ryon.
191
Vascular tumors of the breast were tile subject of a series of previous reports. 1-3 Among the lesions included in our studies of vascular tumor were seven breast masses characterized by m a m m a r y stromal proliferation with the f o r m a t i o n o f interanastomosing slits that simulated vascular spaces. These cases were most often submitted with the provisional diagnosis of low-grade angiosarcoma. Tile recent opportunity to examine two such tumors from patients treated at Memorial Sloan-Kettering Cancer Center (MSKCC) made possible more detailed studies of this uncommon lesion, which we call pseudoangiomatous hyperplasia of mammary stroma. MATERIALSAND METHODS In addition to the two patients treated at MSKCC (case 1 and 2), seven others were seen as consultations from other hospitals. Submitted slides stained with hematoxylin-eosin were reviewed in each case. Clin-
ical infornmtion was obtained from the referring pathologists. Wlten sufficient material was available, trichrome, Alcian blue (with and without hyaluronidase), reticulin, and elastic stains were used. Submitted slides were stained by the immunoperoxidse techniques for Factor VIII, Ulex europaeus agglutinin1, S-100 protein, vimentin, myosin, myoglobin, and actin antigens. For electron microscopy, formalinfixed tissue (case 1) was cut into small cubes, washed in normal saline for two days, and postfixed in 1 per cent o s m i u m tetroxide b u f f e r e d with s-collidine. Specimens were then washed in distilled water and left overnight in 0.25 per cent aqueous uranyl acetate to enlmnce contrast. Following dehydration in graded ethanols, tissues were embedded in Maraglas-DER 732 epoxy resin. 4 For orientation, l-tim sections were stained with sodium borate-buffered 1 per cent toluidine blue. Thin sections were stained sequentially with uranyl acetate and lead citrate and examined in a transmission electron microscope.
RESULTS
Clinical Data Details of individual cases are presented in table 1. T h e patients were premenopausal women (age range, 22 to 52 )'ears; mean, 40 )'ears). Tile lesion generally p r e s e n t e d as a discrete, painless breast htmp that was firm or rubber)' in consistency. Aspiration biopsies in three cases failed to yield cells. In one case a mammogram was reported as questionable. All tumors were treated by excisional biopsy, and seven patients had no subsequent difficulties, with follow-up periods extending to 30 months. A 22year-old woman underwent two subsequent excisions in an 11-month period for persistent lesions, and a 49-year-old woman underwent excision of a local recurrence at 14 months. One patient underwent bilateral subcutaneous mastectomies when the lesion was initially diagnosed as an "atypical vasoformative process."
Pathologic Data Grossly, tile tumors were well circumscribed with the smooth external surface resembling a capsnle in some cases. T h e texture was usually described as rubbery. Tile greatest diameter as measured on the excised specimen, ranged from 2 to 7 cm. No tumor had extensive hemorrhage oi" necrosis. Microscopicall)', the lesions were composed o f intermixed
Received from the Department of Pathology, Memorial tlos3ital, Memorial Sloan-Kettering Cancer Center, New York, New York. Revision accepted for publication August 6, 1985. *Recipient of an American Cancer Society Fellowship. Address correspondence and reprint requests to Dr. Rosen: ) e p a r t m e n t of Pathology, Memorial ttnspital, 1275 York Avenue, qew York, NY 10021.
185
HUMAN PATHOLOGY TABLE t.
Case
Patient's Age ()'r)
1
38
2
52
3
44
4
35
5
22
6
44
7
37
8
35
9
49
Clinical Features
Volume 17, No. 2 (February1986) Summary of Cases
Size (cm)
Gross Description
"Fibroadenoma," 1 too; aspiration unsuccessful Lump, 3 wk; mammogram questionable; 2 aspirations unsuccessflfl Lump; bilateral subcutaneous mastectomies at 30 mo Lump
7 • 5 • 4
Fibroadenonm
5 x 4 x 3
Encapsulated, firm, lobulated Fibrous
3 • 2
Rubbery
Fibrocystic disease with atypical vasoformative process
7 x 5 • 4
Rubber)', nodular
Lump; recurrences at 6 and 13 mo Painless lump
2
Giant adenofibroma, cystosarcoma phylloides ruled out Fibroadenoma variant
"Thickening," aspiration unsuccessfnl Lump, 3 wk
3
Ovoid, rubbery, wellcircu inscribed Well-circumscribed nodule I11 defined
6 • 4 • 3
Lobulated, firm
Painless lump; 5 too; recnrrence at 15 ino
3
Ovoid, circumscribed
3
stromal and epithelial elements. In general, tile epithelium within the lesions was essentially normal, with only mild patchy hyperplasia or duct papillomatosis. Areas of epithelium with apocrine secretory snouts were observed in three cases, and in two cases myoepithelial cells with rotmded-up clear cytoplasm were present in many lobules. T h e ducts and lobules were composed of cells with relatively uniform nuclei, occasionally several cells thick. Epithelial atypia was not present. Occasionally, attenuated ducts similar to those in fibroadenomas were present. The characteristic and most prominent stromal feature was a complex pattern of interanastomosing empty spaces that was easily seen at low power (fig. I). These channels were present within as well as between breast lobules. In most cases they tended to be a r r a n g e d concentrically immediately a r o u n d the breast lobules (fig. 2). Often, spindly stromal cells were present at tile margins of these spaces, simulating endothelial cells (fig. 3). The impression on low-power examination was that of a complex vascular proliferation, although only very rarely were isolated red blood cells present in these pseudoangiomatous spaces. The stroma also contained abundant collagen, which was focally very dense, and elongated wavy spindle cells with poorly defined cytoplasmic borders. T h e nuclei o f these cells were attenuated and usually had pointed ends. Mitoses were not seen, nor was atypia or p l e o m o r p h i s m present. Abundant, distinct small blood vessels with readily apparent endothelial cells were also seen in the stroma. Frozen section analysis was performed in three cases. It is noteworthy that the stromal pseudoangiomatous spaces were discernible in fi'ozen sections, indicating that they were not an artifact of formalin186
Original Diagnosis
Pseudoangiomatous hyperplasia of mammary stroma
Grade I angiosarcoma Probably benign reactive vascular proliferation Grade I angiosarcoma versus benign angiomatous proliferation Well-differentiated angiosarcoma versus fibroadenomatoid tumor
fixed, p a r a f f i n - e m b e d d e d tissue. Periodic a c i d Schiff stain showed no evidence of basement membrane material near the clear channels. Alcian blue staining within the stromal collagen as well as in the clear pseudoangiomatous spaces was removed by hyaluronidase treatment. T r i c h r o m e staining of collagen was strong, while reticulin and elastic stains were negative except in definite blood vessels. Immunoperoxidase stains for Factor VIII protein and Ulex europaeus agglutinin- 1 stained endothelial cells of definite blood vessels, but not tile stromal cells adjacent to the clefts (fig. 4). Tissue from case 1 was examined by electron microscopy. The specimen consisted of epithelial ducts in a dense fibrovascular stroma. The ducts were lined by three to four tiers of ductular epithelial cells separated from the stroma by a single discontinuous layer o f myoepithelial cells resting on a basement membrane. T h e epithelial-stromal junction was intact. Only tile fine structure of tile stromal elements is detailed in this report. This abundant fibrovascular tissue consisted of capillaries (small arterioles and venules were also observed), vascular pericytes, modified stronml pericytes, fibroblasts, many o f which lined pseudoangiomatous spaces, and dense bands of collagen fibrils (figs. 5 and 6). The closed lumina of true blood vessels were lined by endothelial cells joined by tight junctions and surrounded by an adluminal basement membrane (figs. 5 to 7). The basement membrane was occasionally multilayered (fig. 6). The cytoplasm of the endothelial cells contained prominent surface pinocytotic vesicles, a meshwork of 10-nm intermediate filaments, and Weibel-Palade bodies (figs. 6 and 7). All of the vessels were intimately surrounded by a discontinuous layer of pericytes, one to two cells thick, enveloped by a basement
'
.
.
dr~
S
.
~
q~
w
;A ~" T,! " ~ k
f~'qF
t
, ~ "
I
.'~
."
"
" 9
. . . . . . . . . . . . . . .
FIGURE 4 (top left). Case 1. Discrete, seemingly encapsulated 24-g tumor excised from the right breast of a 38-year-old woman. The lowpower view shows pseudoangiomatous stromal proliferation between duct and Iobular units. [ x 40.] FIGURE 2 [top right]. Enlarged view of the complex arrangement of spaces resembling vascular structures that constitute pseudoangiomatous hyperplasia of mammary stroma. ( x ~00.] FIGURE 3 [bottom left). Case 3. Rubbery mass, excised from the left breast of a 44-year-old woman. The high-power view reveals pseudoangiomatous slits, adjacent spindle cells, and intervening dense collagen. Nuclear a~/pia and pleomorphism are absent. [ x 250.] FIGURE 4 [bottom right). Immunoperoxidase staining for Factor VIII antigen; hematoxylin counterstain. Staining is evident in the endothelium of a true blood vessel (top of field] but not in tissue exhibiting pseudoangiomatous stromal hyperplasia. ( x 250.]
187
HUMAN PATHOLOGY
Volume 17, No. 2 (February 1986]
~j:-~
...;' FIGURE 5 (top]. Survey electron micrograph illustrating [from left to right] a portion of a small arteriole, modified pericytes in a dense collagenous stroma, and a pseudoangiomatous space devoid of collagen fibrils. [ x 6,240.] FIGURE 6 [bo#om). Portion of a capillary intimately Surrounded by a discontinuous layer of pericytes. Cytoplasmic processes c o a t e d by remnants of basement membrane, probably from attenuated modified pericytes, are visible in the adjacent collagenous 9stroma. [x13,600.)
cells were distinguished by a fairly well developed rough endoplasmic reticulum and, often, a prominent Golgi apparatus. Pinocytotic vesicles, Weibel-Palade bodies, and basement membranes were not identified (fig. 8, left). The slender cytoplasmic processes o f the fibroblasts t e r m i n a t e d in tile collagenous stroma and were occasionally joined by small rudimentary cell junctions (fig. 8, right). The stroma between the small blood vessels (pre-
membrane. Tile pericytes displayed pinocytotic vesicles and dense attachment plaques on tile outer plasmalemma, distal to the endothelial cells, and contained variable numbers of longitudinally oriented 6nm microfilaments associated with widely scattered fusiform dense bodies (figs. 5 to 7). Pseudoangiomatous spaces, devoid of collagen fibrils and lined by a discontinuous sheath of fibroblastic cells were also evident (fig. 8). The fibroblastic 188
PSEUDOANGIOMATOUS HYPERPLASIAOF MAMMARY STROMA (Vuitch et al.)
FIGURE 7. Detail of the wall of a small capillary (top of field]. The endothelial cell cytoplasm contains numerous Weibel-Palade bodies [arrows], subsurface pinocytotic vesicles, and intermediate filaments. The endothelial ceil is surrounded by portions of three pericyte processes. The distal plasmalemma of the endothelial cell and pericytes is covered by a distinct basement membrane. I x 28,000.]
strnation. A stud)' by Vogel et al. 7 in 1981 confirmed the distinct histologic patterns of breast epithelium and perilobular stroma at different phases of the menstrual cycle. T h o s e investigators r e p o r t e d a "loose, broken" stroma in tile luteal phase (days 15 to 20) and a "loose, edematous" stroma in the secretory phase (days 21 to 27). T h e stroma ill our cases has some similarity to the stroma in tile physiologic luteal and secretory phases of the menstrual cycle. T h e open clefts containing hyaluronidase-sensitive mucopolysaccharide resemble the histologic features shown in 4 and 5 in the report of Vogel et al. 7 We have also seen similar perilobular clefts focally in hyperplastic breast tissue fl'om premenopausal women, sonletimes associated with secretory changes in the acini. Ttle cases of pseudoangionmtous hyperplasia of the mammary stroma presented here differ from the physiologic states in important respects. In only three cases was secretory activity present in some acini, demonstrating a definite stromal-epithelial dissociatlon m most cases. We believe that the stromal changes may originate as a response to progesterone ill estrogen-primed tissue. It appears that tile pseudovascular pattern persists due to unknown factors, suggesting that the stroma becomes autonomons. The stromal pseudoangiomatous cleft formation was
d o m i n a n t l y capillaries) and the pseudoanigiomatous spaces lined by fibroblasts contained longitudinally a r r a n g e d , extremely a t t e n u a t e d long cytoplasmic processes joined by occasional tight or rudimentary cell junctions (figs. 5 and 6). The cytoplasmic processes of these cells, presumed to be modified pericytes, retained fragments of basement membrane on both surfaces, pinocytotic vesicles, and occasional small clusters of intermediate filaments (fig. 9). DISCUSSION Relatively few morphologic descriptions of tile mammary stroma in physiologic and pathologic conditions are available. Mammary stroma is generally accepted as lmving two elements, a specialized hormonally responsive intralobular and periductal component, and a supporting interlobular stroma, which is believed to be largely unresponsive to hormones. In 1945 Foote and Stewart 5 described the histologic changes in the intralobular stroma during the menstrual cycle. Ozzello and Speer 6 correlated those histologic" changes with fluctuations in the hyaluronidase-susceptible mucopolysaccharide content of the stroma, which increased dramatically in the premenstrual phase after day 12 to 15, along with secretory activity of the acini, and decreased following men189
HUMAN PATHOLOGY
Volume 17, No. 2 (February 1986]
FIGURE 8. Left,, pseudoangiomatous space lined by slender fibroblastic cells. The cells contain a well-developed rough endoplasmic reticulum and Golgi apparatus. No basement membrane is evident. (• 10,000.] Right blunt termination of fibroblastic processes lining pseudoangiomatous space. Cell processes are joined by rare rudimentary cell junctions [arrows]. Numerous collagenous fibrils seen in cross and longitudinal sections surround the pseudovascular space. (x 15,000.)
plasmic processes. T h e findings in our case were similar to those of Ahmed. I-~ T h e importance of recognizing pseudoangiomatous stromal hyperplasia of the breast as an entity is to prevent misdiagnosis of the condition as a l o w grade angiosarcoma. T h e features o f these lesions that suggest angiosarcoma are the size of the mass (usually 3 cm or greater) and the low-power microscopic appearance suggesting anastomosing channels or clefts. However, pseudoangiomatous stromal hyperplasia produces a discrete, rounded, often rubbery mass that grossly resembles a fibroadenoma and does not have the henmrrhagic appearance that typifies angiosarcomas. N u m e r o u s histologic features permit differentiation of this lesion from angiosarcoma. In pseudoangiomatous hyperplasia of mammary stroma, pleomorphism an& hyperchromasia of the stromal cell nuclei are minimal to absent, papillary endothelial ingrowths into the clefts are not present, and the channels only very rarely contain red blood cells. Mitoses and necrosis are not seen, nor is destruction o f m a m m a r y epithelial structures evident, even when the process extends into lobtfles. Intermingling of the stromal channels within lobules is
so extensive in our cases that it extended between lobules. We cannot be certain whether interlobular stroma, normally hormone-insensitive, is recruited into the lesion, or whether marked expansion of the intralobular stroma merely displaces interlobtdar stroma. An interesting but unexplored feature o f these masses is whether they contain estrogen and/or progesterone receptors. T h e ultrastructural features of breast stroma have not been investigated intensively. Fanger and Ree, s in their study of the uhrastructural changes in the breast during the menstrual cycle, confined their observations to the epithelium. Ozzello 9 described the normal markedly attenuated "delimiting fibroblasts" around lobules. Most ultrastructural studies of breast stroma have concentrated o n fibroadenomas, with some controversy as to the nature of the stromal cells. According to Archer and Omar, 1~ the stromal cells are fibrocytes, and Ozzello 9 agreed with that conclusion. Murad et al. 11 considered the stromal cells pericytes. A h m e d 12 stated that he found fibroblasts as well as some cells with features of pericytes, i.e., basal lamina, hemidesmosomes, and numerous pinocytotic vesicles, along the plasma membrane of long cyto190
PSEUDOANG1OMATOUS HYPERPLASIAOF MAMMARY STROMA [Vuitch et al.]
.
.
.
.
.
- '..
.7:/
:
-2.-
:
-
.~
.-
.!if
.:~
~z.
~
,.'
.
.~
.~
~-~r
FIGURE 9. Cytoplasmic processes, probably from modified pericytes surrounded by numerous collagen fibrils in the stroma. Pinocytotic vesicles [arrow] and remnants of basement membrane are visible. [ x 56,000.}
t.ypical, frequently with concomitant mild proliferanon of the epithelial structttres. I m m u n o p e r o x i d o s e reagents such as Factor VIII, are not consistently positive in angiosarcoma and are therefore not helpfitl. O n e patient had two local recurrences, a n d a second patient had one recurrence. However, microscopic examination disclosed no greater cellularity or pleomorphism in the tumors in these patients than in those in other patients. T h e r e was no change in cellularity or pleomorphism from the initially excised masses to the recurrences in those two patien.ts. Since the initial masses in the patients with recurrences were the smallest in o u r series, it is likely that excision was incomplete. Altei'natively, the breast stronm adjacent to the excised lesion may become susceptible later to the same u n k n o w n stimulus that causes pseudoangiomatous hyperplasia of m a m m a r y stroma. We r e c o m m e n d c o m p l e t e local excision with clinical t-ollow-up evaluation that probably does not need to extend beyond three )'ears.
REFERENCES 1. Donnell RM, Rosen PP, Lieberman PH, et al: Angiosarcoma and other vascular tumors of the breast: pathologic analysis as a guide to prognosis. Am J Snrg Pathol 5:629, 1981 2. Jocefczyk M, Rosen PP: Vascular tumors of the breast: I1. Perilobular angiomas and angiomas. Am J Surg l'athol 9:491, 1985 3. Rosen PP, Jocefczyk M: Vascular tumor of the breast: III. The venous hemangioma. Am J Surg Pathol 9:652, 1985 4. Erlandson RA: A new Maraglas, DER 732, embedment for electron microscopy. J Cell Biol 22:704, 1964 5. Foote FW, Stewart FW: Comparative studies of cancerous versus noncancerous breasts. Ann Surg 121:6; 197, 1945 6. Ozzello L, Speer FD: The mucopol)saccharides in the normal and diseased breast. Their distribution and significance. Am J Pathol 34:993, 1958 7. Vogel PM, Georgiade NG, Fetter BF, et al: The correlation of histologic changes in the human breast with the menstrual cycle. Am J Pathol 104:23, 1981 8. Fanger H, Ree HJ: Cyclic changes of human mammary gland epithelium in relation to the menstrual cycle--an uhrastructural study. Cancer 34:574, 1974 9. Ozzello L: Ultrastructnre o f the human mammary gland. Pathol Annu 6:1, 1971 10. Archer F, Omar M: The fine structure of fibro-adenoma of the human breast. J Pathol 99:113, 1969 I 1. Murad TM, Greider MH, Scarpelli DG: The uhrastructure of human mammary fibroadenoma. Am J Pathol 51:663, 1967 12. A h m e d A: T h e myoepithelium in human breast fibroadenoma. J Pathol 114:135, 1974
Acknowledgments. T h e a u t h o r s a c k n o w l e d g e the con:ribution o f cases a n d clinical i n f o r m a t i o n flora the folo w i n g p a t h o l o g i s t s : J . E. Boline, R. Chapwick, E. L. Coffey, [r., S. Dias, M. P. H y m a n , J. A. R u d d , and A. S e y m o u r . ? h o t o g r a p h s w e r e p r e p a r e d by Roy K e p p i e , Kin K o n g , m d M a r g a r e t Ryon.
191