Varma, Sonal MD; Shin, Sandra J. MD
Mammary lesions composed of spindle cells can be diagnostically challenging for several reasons. First, a host of reactive, benign, and malignant entities comprise the differential diagnosis irrespective of their degree of cytomorphologic atypia and mitotic activity. Second, these entities have overlapping clinico-radiologic and immunohistochemical profiles that further muddy the diagnostic waters. Finally, even the most common entities comprising “spindle cell lesions” of the breast are relatively uncommon.
In this review, we outline a practical approach to these problematic lesions. We present a simple algorithm that can be used as a guide for evaluating key components of a spindle cell lesion. This, in turn, should lead to the formation of a reasonable morphologic differential diagnosis that can be further narrowed using appropriate immunohistochemical stains, ideally leaving 1 or 2 entities for serious consideration. In the second part of this review, we provide concise summaries of the most commonly encountered entities that comprise this morphologic subset of breast lesions.
KEY COMPONENTS IN THE EVALUATION OF A SPINDLE CELL LESION
An algorithm can provide an orderly and predictable means by which a spindle cell lesion can be initially evaluated. The algorithm’s structure can be further refined or revised over time as more information becomes available about entities that constitute the differential diagnosis. In any given spindle cell lesion, it is important to evaluate certain features: (1) composition of the cellular proliferation; (2) the presence of cytomorphologic atypia; (3) growth pattern including tumor border characteristics; (3) mitotic activity; (4) adjacent or admixed cells and/or tissue; and (5) clinico-radiologic features (Table 1).
The sample size of the specimen is often inversely related to the breadth of the differential diagnosis; the most limited samples (ie, needle core biopsies) produce the broadest differential diagnoses. This, of course, is due to the possibility that limited material may not contain all the key components of a lesion, thus, leading one to consider more entities than necessary or inadvertently leave out others. For instance, only the stromal component of a phyllodes tumor may be sampled in the initial needle core biopsy, and consequently, the case is worked-up as a “spindle cell lesion” without the possibility of a phyllodes tumor even being considered. It is only at the time of excisional biopsy that the diagnosis becomes clear (Fig. 1).
Composition of the Cellular Proliferation
Although tempting at the onset, it is better to refrain from scrutinizing the cytomorphologic features of the spindle cells themselves. A more productive approach is to first assess the overall composition of the cellular proliferation. Cellular proliferations described as “spindle cell lesions” are not always composed exclusively of cells with spindle morphology. For instance, spindle cell lesions can harbor epithelial or epithelioid-appearing cells and, in some cases, frankly glandular or squamous elements.
Mixed spindle cell and epithelial proliferations can be seen in 2 arrangements. Both cell types can be intimately admixed with one another or they can be coexisting but distinctly separate components. In the former scenario, spindle cell metaplastic carcinoma should be strongly considered. Other possibilities include displaced epithelial cells in biopsy site changes or metastasis from an extramammary site. The latter possibility where the 2 cells types are coexisting, but distinctly separate, mixed spindle cell metaplastic carcinoma and conventional invasive, poorly differentiated duct carcinoma should be considered. Fibroepithelial lesions (fibroadenomas, phyllodes tumors) and metastasis also fit this category.
If distinct glandular structures are present, entities such as low-grade adenosquamous carcinoma, syringomatous adenoma, adenomyoepithelial tumors, displaced glands in biopsy site changes, metastasis, or some radial sclerosing lesions (associated with spindle-appearing adjacent stroma) should be considered. Mixed spindle cell metaplastic carcinoma and conventional invasive (moderately differentiated) duct carcinoma can also be a possibility.
Pure Spindle Cell Proliferations
The distinction between cytologically bland and malignant spindle cells is simultaneously an important yet treacherous one. A cytologically bland-appearing lesion does not necessarily indicate that the lesion is benign (or reactive). Conversely, not all cytologically malignant-appearing spindle cell lesions are biologically aggressive. The exercise of dividing spindle cell lesions by this characteristic is solely done for the purpose of considering one of 2 distinctly different groups of entities, both of which contain lesions of benignity and malignancy.
Mixed Spindle Cell and Epithelial Proliferations
The differential diagnosis in this group of lesions is largely formed after assessing the arrangement (and relationship to each other) of these 2 components as discussed above (Table 2). The presence or absence of cytologic atypia of the epithelial component will further narrow the possibilities into benign or malignant groups with several exceptions. Cytologically bland spindle cells and minimally atypical epithelioid cells can constitute some examples of spindle cell metaplastic carcinomas. Also, the stromal component of phyllodes tumors tends to be heterogenous and only cytologically bland areas may be inadvertently sampled in needle core biopsies. Deceptively bland glandular elements are characteristic of low-grade adenosquamous carcinoma. In general, for mixed spindle cell and epithelial proliferative lesions, proper immunohistochemical work-up is a necessary next step (discussed below).
Displaced epithelium or glands can be benign or carcinomatous and, if the latter, one needs to exclude invasive carcinoma. Sometimes, this can be a very difficult distinction to make on morphologic grounds alone. If the carcinomatous epithelial cells/glands are embedded exclusively in biopsy site changes or a needle tract, they most likely represent epithelial displacement. In fact, the time interval between a needling procedure and excisional biopsy is inversely related to the likelihood of finding epithelial displacement. In 1 study,1 42% of patients with epithelial displacement had a needling procedure <15 days before excisional biopsy. Conversely, when the interval was >28 days, only 15% of patients were found to have epithelial displacement. Also, displaced epithelium loses viability within the inflammatory environment of the biopsy site and as such shows varying degrees of squamous metaplasia (likely an attempt to repair) and degenerative changes (pyknotic nuclei, shrunken eosinophilic cytoplasm). Invasive carcinoma, in contrast, usually infiltrates intact mammary stroma without evidence of focal squamous changes and is located away from the previous biopsy site. Further investigation by immunohistochemical studies can be useful only if the displaced epithelium demonstrates presence of myoepithelial cells. If no myoepithelial cells are detectable, it does not confirm the presence of invasive carcinoma as displaced epithelium can be denuded of these cells in some cases. If the question is whether the epithelium is benign or carcinomatous, immunohistochemical stains to show a common phenotype with the adjacent in situ carcinoma (ER, PR, Her-2) may provide supporting evidence in some cases.
Pure Spindle Cell Proliferations With Bland Cytomorphology
This category is diagnostically the most challenging as it encompasses a wide range of reactive, benign, and malignant entities with overlapping features (Table 1). Non-neoplastic (largely reactive) lesions include postoperative myofibroblastic proliferation (ie, spindle cell nodule), granulation tissue/scar formation after biopsy or trauma, traumatic neuroma, and inflammatory myofibroblastic tumor (inflammatory pseudotumor). Biphasic tumors such as fibroepithelial lesions should likewise be considered in needle core biopsies where it is possible that only the stromal component is sampled (as discussed previously). Locally aggressive tumors such as desmoid-type fibromatosis as well as more malignant ones such as spindle cell metaplastic carcinoma, particularly fibromatosis-like variant, low-grade angiosarcoma, and low-grade myofibroblastic sarcoma can be deceptively bland appearing. Myofibroblastoma, nodular pseudoangiomatous stromal hyperplasia (PASH), fascicular PASH, leiomyoma, spindle cell lipoma, angiolipoma, Rosai Dorfman disease (sinus histiocytosis with massive lymphadenopathy), and nerve sheath tumors (neurofibroma, schwannoma) are among the long list of benign tumors that fall under this category.
Because of the large number of possibilities, other characteristics can be used to narrow the differential diagnosis. Assessing growth pattern/tumor border, mitotic activity, and coexisting cells and/or tissue can be very helpful in this regard. These additional features are summarized in Table 3.
Growth Pattern/Tumor Border (of Pure Spindle Cell Proliferations With Bland Cytomorphology)
The growth pattern of lesions purely composed of spindle cells of bland cytomorphology can be “classic” for the particular entity, or in other cases misleading or not well represented especially in limited material (ie, needle core biopsy). Therefore, we caution that although helpful when available and representative of the lesion, sometimes it can represent a “red herring.”
Long, Collagenized Bands
This growth pattern is most commonly associated with fibromatosis or a well-healed scar. Less frequently but equally important to consider is spindle cell metaplastic carcinoma of the fibromatosis-like variant. All 3 proliferations can show invasive tumor borders, have associated chronic inflammation and lack mitoses (Fig. 2).
Capillary Vessel Like
This growth pattern can be characteristically seen in low-grade angiosarcoma, but is also seen in PASH, PASH tumor, angiolipoma, and granulation tissue (Fig. 3).
Short or Medium Fascicles, Woven
This growth pattern can be seen in examples of myofibroblastoma, nodular fasciitis, fasicular PASH, phyllodes tumor (stromal component), biopsy site changes, and spindle cell metaplastic carcinoma among others (Fig. 4).
This growth pattern can be seen in inflammatory myofibroblastic tumor (inflammatory pseudotumor), spindle cell nodule, Rosai-Dorfman disease, fat necrosis, and biopsy site with histiocytic infiltration (Fig. 5).
Assessing the tumor border can be helpful in some cases. Entities in the differential diagnosis that characteristically have circumscribed contours include angiolipoma, myofibroblastoma, PASH tumor, whereas others can have, at least partly, invasive or irregular tumor borders such as fibromatosis, spindle cell metaplastic carcinoma, scar, low-grade angiosarcoma, and granulation tissue. In most cases, any possibilities emerging from assessing growth pattern and tumor border need to be further studied by immunohistochemical stains (see below).
Mitoses (of Pure Spindle Cell Proliferations With Bland Cytomorphology)
Mitoses can be present in cytologically bland spindle cell lesions, which in most cases signify malignancy. Some exceptions include exuberant granulation tissue (status post biopsy) or a rare case of myofibroblastoma both of which can harbor rare mitotic figures. Unlike those arising in other anatomic sites, mammary fibromatosis characteristically does not contain lesional mitoses, and as such, the presence of considerable mitotic activity in a suspected case should alert one to consider an alternative diagnosis. Also, low-grade angiosarcomas contain rare if any lesional mitoses; therefore, the absence of mitotic activity does not preclude this diagnosis. This is also true for cytologically bland examples of spindle cell metaplastic carcinoma including the fibromatosis-like variant, and benign phyllodes tumors.
Coexisting Cells and Tissue (of Pure Spindle Cell Proliferations With Bland Cytomorphology)
Practicing pathologists are well aware of the fact that clinical history accompanying a specimen can be sketchy and even inaccurate at times. Consequently, gathering more clinical information is a common exercise during the work-up of a challenging case.
The location of the lesion is important to ascertain. In the vast majority of instances, spindle cell lesions arise in the breast proper. However, some tumors are found to arise from the subcutaneous tissue of the breast, the axilla, or the nipple, and this knowledge is critically important. Especially when evaluating a needle core biopsy specimen, the search for adjacent uninvolved tissue should be routine. The presence of uninvolved glandular breast tissue confirms the lesion’s location in the breast, whereas its absence may suggest another site such as the subcutis or axilla, especially if only adipose tissue is seen. Nodular fasciitis, for instance, more frequently arise in the subcutis and rarely in mammary parenchyma. Also, spindle cell tumors of dermatologic origin such as dermatofibroma, dermatofibrosarcoma pertuberans, and melanocytic lesions need to be additionally considered if the lesion is known to be arising from the skin or subcutis. Fibromatosis and biopsy site are also possible proliferations arising in this location. If a bland spindle cell lesion is arising in the nipple, leiomyoma should be considered.
Coexisting cells and their relationship with lesional spindle cells can provide additional clues. The presence of inflammatory cells, particularly lymphocytes, can be helpful as they are commonly associated with certain spindle cell proliferations including granulation tissue/scar, nodular fasciitis, inflammatory myofibroblastic tumor (primarily plasma cells), spindle cell metaplastic carcinoma, fibromatosis, low-grade adenosquamous carcinoma, and syringomatous adenoma. Furthermore, the arrangement of lymphocytes can provide insight as they tend to form peripheral aggregates in fibromatosis and low-grade adenosquamous carcinoma but are dispersed among spindle cells in other entities such as spindle cell metaplastic carcinoma, granulation tissue, inflammatory myofibroblastic tumor, and nodular fasciitis (Fig. 6). Red blood cells can be seen in spindle cell lesions that are fundamentally vascular in nature. Occasionally, microthrombi and, sometimes, hematopoietic cells such as megakaryocytes can be identified in examples of angiolipoma. Finally, the presence of multinucleated foreign body type giant cells can indicate a ruptured squamous cyst (in syringomatous adenoma) and foreign material (reactive lesions, biopsy site), whereas osteoclast-like giant cells can be seen in some examples of nodular fasciitis. Occasionally, benign multinucleated stromal giant cells can be seen in fibroepithelial lesions such as phyllodes tumors, which additionally should not be mistaken for stromal atypia.
Clinico-radiologic Features (of Pure Spindle Cell Proliferations With Bland Cytomorphology)
Additional clinical history or presentation can either help to further narrow the differential diagnosis or provide strong supportive evidence that the most likely pathologic diagnosis is correct. For instance, a history of prior biopsy or trauma fits well with a spindle cell proliferation that represents reactive changes including a spindle cell nodule, traumatic neuroma, fat necrosis, or a healing scar while excluding fibromatosis. In cases of myofibroblastoma, knowing the male gender of the patient is clinically compatible with the diagnosis. Palpable breast masses correlate well with mass-forming spindle cell lesions histologically.
The imaging characteristics can be contributory as a number of spindle cell lesions are detected by screening mammography±ultrasonography. The extent and contours of some lesions can be better delineated by imaging than physical examination. Low-grade angiosarcomas that can be confused with angiolipomas histologically are quite different in size, as examples of the former are rarely smaller than 3 cm, whereas this is commonly true for the latter. Some phyllodes tumors can appear lobulated, clefted, and/or cystic. Fibromatosis typically exhibits infiltrative tumor borders that can be appreciated radiographically, an appearance that would be highly unlikely in PASH tumor or myofibroblastoma among others with rounded contours.
Pure Spindle Cell Proliferations With Atypical/Malignant Cytomorphology
Pure spindle cell proliferations with atypical or frankly malignant cytomorphology include both reactive as well as malignant processes (Table 1). Reactive proliferations such as exuberant granulation tissue can not only show cytologic atypia but also mitotic figures. Nodular fasciitis can also mimic a malignant proliferation for the same reasons. However, the majority of lesions that fall under this category are malignant and that significantly simplifies this subgroup. Spindle cell metaplastic carcinoma, borderline or malignant phyllodes tumor, adenomyoepithelioma with carcinoma, high-grade sarcoma including angiosarcoma, lymphoma, and metastases (spindle cell malignant melanoma, carcinoma, sarcoma) are included in this category.
Growth Pattern/Tumor Border, Mitoses, Coexisting Cells/Tissues, and Clinicoradiologic Features (of Pure Spindle Cell Proliferations With Atypical/Malignant Cytomorphology)
The growth pattern/tumor border are features that are less helpful when confronted with pure spindle cell proliferations with atypical/malignant cytomorphology, with the exception of granulation tissue associated with a recent history of biopsy or trauma. Mitoses are typically present in all of these entities; however, the degree of activity may not be accurately assessed in small samples. For possible cases of metastases, a known history of an extramammary malignancy and/or radiologic evidence of metastatic disease or extramammary malignancy are critical to confirm. Angiosarcomas may clinically present with skin alterations such as a bruise or similar discoloration. Malignant phyllodes tumor may appear partially cystic by imaging. In general, the greater work to be done with these lesions is to determine the phenotype (and in some cases, the extramammary site of origin) of the proliferation, which is largely achieved through immunohistochemistry (Fig. 7).
Once a reasonable morphologic differential diagnosis is constructed, a panel of selected immunohistochemical stains can be performed to further narrow the list of possibilities (Table 4). In most cases, this stage of the work-up is straightforward; however, immunohistochemical stains are wrought with their own pitfalls, some of which are highlighted here.
Positive Staining Can Be Sporadic, Focal, and/or Weak
Spindle cell metaplastic carcinoma is a diagnosis that should be excluded in virtually all cases of spindle cell proliferations arising in the breast. Demonstrating immunohistochemical evidence of its epithelial phenotype can be challenging and sometimes, in small samples, unachievable. The reason for this difficulty is that immunoreativity of lesional spindle cells with cytokeratins and/or p63 can be focal in extent and weak in staining intensity. Furthermore, which cytokeratin markers will be immunoreactive in a given case is unpredictable, and in rare cases all cytokeratins used will be negative. In such cases, p63 positivity will confirm the diagnosis. Assessing p63 immunoreactivity should also be performed with caution, as this stain can be positive in neighboring lymphocytes and epithelial cells. To exclude spindle cell metaplastic carcinoma in any case, we recommend performing a panel of cytokeratins (CKAE1/3, 34βE12, CK7, Cam 5.2) and p63. Additional cytokeratin stains such as CK5, CK14, and MNF-116 can be used if initial markers are negative. More recently, focal and patchy cytoplasmic immunoreactivity in stromal cells has been reported in a subset of phyllodes tumors, however, all were negative for p63 which would be helpful in distinguishing this entity from spindle cell metaplastic carcinoma.3 In needle core biopsy samples, a negative immunohistochemical work-up does not exclude the possibility of spindle cell metaplastic carcinoma, and therefore, these stains should be repeated on the excisional biopsy specimen in such cases.
An Immunohistochemical Stain Can Be Positive in Several Spindle Cell Lesions of the Breast and Not Specific for a Particular Entity
Nuclear localization of β-catenin can be seen in almost all cases of fibromatosis; however, this is not specific to this entity as immunoreactivity can be seen with other morphologically similar spindle cell proliferations such as scar, spindle cell metaplastic carcinomas, and phyllodes tumors. Lacroix-Triki et al2 found that among the spindle cell metaplastic carcinomas studied, 22% showed nuclear, 100% showed cytoplasmic, and 27% showed membranous β-catenin expression, respectively. Nuclear expression in the stromal component of phyllodes tumors was observed in 94% of cases studied. In contrast, all 8 cases of mammary fibromatosis showed nuclear expression in at least a third of the tumor.
CD34 is a key immunostain used to confirm the diagnosis of myofibroblastoma; however, positive staining can also be seen in PASH tumor, fascicular PASH, spindle cell lipoma, angiosarcoma, and the stromal cells of phyllodes tumors. Therefore, it is important to use additional immunostains such as CD31 and S-100 protein to exclude angiosarcoma and spindle cell lipoma, respectively. Myofibroblastomas are also variably positive for smooth muscle actin and desmin while negative for S-100 protein.
Staining Characteristics Are Not Well Characterized
Contrary to results of limited studies4,5 available in the current literature, we have observed that consistent staining patterns are not seen in adenomyoepithelial tumors; and only in a subset of cases do the epithelial and myoepithelial components of such tumors stain with appropriate epithelial-specific and myoepithelial-specific immunostains. Phyllodes tumors constitute another subset of breast lesions that are not characterized by specific or a panel of immunostains. The immunohistochemical staining characteristics of epithelial and myoepithelial markers in low-grade adenosquamous carcinoma has only been recently described; where inconsistent staining with myoepithelial markers are commonly observed and when present, “core” staining by cytokeratin immunostains is characteristic of this entity.6
MORE COMMONLY ENCOUNTERED PURE SPINDLE CELL LESIONS OF THE BREAST
We conducted an informal inventory of mammary spindle cell lesions reported at our institution, in our general surgical pathology files and our consultation material. In our general surgical pathology files, among 42 spindle cell lesions identified, we found the relative frequencies of diagnoses rendered in this subset to be: spindle cell metaplastic carcinoma (46%), fibromatosis (13%), myofibroblastoma (10%), primary mammary sarcomas (6%), malignant phyllodes tumor (4%), benign phyllodes (3%), and biopsy site/scar (3%). The remainder consisted of various entities, each constituting 2% (or less) of the group. Interestingly, these relative frequencies were similar when separately evaluating cases received through our breast pathology consultation service. Among 248 spindle cell lesions submitted to our consultation service, the frequencies of diagnoses were: spindle cell metaplastic carcinoma (45%), fibromatosis (12%), myofibroblastoma (12%), primary mammary sarcomas (12%), benign phyllodes (5%), and biopsy site/scar (5%). The remainder consisted of various entities, each constituting 2% (or less) of the group.
These analyses bring forth 2 important observations. First, they confirm that the morphologic subset of spindle cell lesions of the breast is comprised entirely of uncommon entities and their relative frequencies within this category do not reflect their true prevalence. Second, they unequivocally demonstrate that spindle cell metaplastic carcinoma is the most common diagnosis rendered of spindle cell lesions of the breast and therefore, as mentioned before, this diagnosis should be excluded in virtually every case.
Spindle Cell Metaplastic Carcinoma
Metaplastic carcinoma is uncommon and accounts for <0.5% of invasive breast cancers.7 This diagnosis includes lesions that are predominantly of spindle cell, squamous or heterologous metaplasia including matrix-producing examples.8–10 A given tumor may show various combinations of these metaplastic elements and be admixed with conventional types (usually ductal) of invasive mammary carcinoma. Currently, it is believed that all metaplastic components are derived from the same clone and that there is sarcomatous metaplasia or “transdifferentiation” in a proportion of malignant cells. Metaplastic carcinomas show upregulation of genes involved in epithelial mesenchymal transition and downregulation of epithelial genes by transcriptional profiling.11,12 Metaplastic carcinomas display myoepithelial differentiation very similar to basal-like breast carcinomas and are typically triple negative (ER, PR, HER-2 negative).13
Metaplastic carcinomas occur in an age group of women similar to that of invasive ductal carcinoma not otherwise specified. The tumor size ranges from 1.5 up to 15 cm. By imaging, metaplastic carcinomas usually do not show the typical features of invasive ductal carcinoma and resemble benign lesions radiologically.14 The tumors can be circumscribed to only mildly infiltrative on gross examination. Cystic and necrotic areas may be evident in some tumors. It usually appears as a firm, white mass well demarcated from the adjacent breast parenchyma. The patterns of squamous differentiation can range from mature keratinizing to undifferentiated spindle cell types.
Spindle cell carcinomas can be deceptively benign appearing and low grade or may be present as overtly malignant high grade tumors. Cytologically bland metaplastic tumors include low-grade adenosquamous carcinoma, and fibromatosis-like (see below) and fasciitis-like variants of spindle cell carcinoma. High-grade metaplastic carcinomas are highly resistant to conventional chemotherapy for breast cancer.15 They usually do not show axillary node involvement, but extranodal metastases are common.
Fibromatosis-like Variant of Spindle Cell Metaplastic Carcinoma
The fibromatosis-like variant of spindle cell metaplastic carcinoma is important to include in the differential diagnosis of cytologically bland spindle cell lesions.
First described in 1999,16 this variant is typically seen in 40- to 80-year-old women as a single palpable mass. On gross examination, these tumors are typically firm white and predominantly solid masses with circumscribed to partially infiltrative edges. The tumors show infiltrative finger-like projections at low power and are predominantly composed of spindle cells. Squamous or glandular component only constitutes 5% or less of the cells. There is an admixture of plump epithelioid spindle cells and more slender fibromatosis-like spindle cells at the periphery of the tumors. The spindle cells show minimal atypia and are variably cellular. Mitotic rate can vary but usually low (0 to 3 per 10 high-power microscopic fields). Inflammatory lymphoplasmacytic aggregates can be seen at the periphery or scattered within the tumor, which can further complicate the distinction from fibromatosis (Fig. 8). This variant along with low-grade adenosquamous carcinoma can be found to arise in association with complex sclerosing or papillary lesions.17,18 Adjacent ductal carcinoma in situ can be identified in a small number of cases.16
Patients are most at risk for local recurrence. No one to date has developed axillary disease and wide local excision is the treatment of choice. Pulmonary metastases, however, were documented in 2 of 24 patients in 1 study.19
Mammary fibromatosis is an uncommon low-grade proliferation of fibroblasts believed to be derived from the pectoral muscle fascia or Cooper ligaments. It is rarely associated with Gardner syndrome or familial multicenteric fibromatosis.20 Approximately 10% of mammary fibromatosis cases reported in the literature are associated with breast implants.21 They are thought to arise from the fibrous implant capsule and are seen in both silicone and saline implants.22 Trauma or prior breast surgery may be associated with fibromatosis.
Clinically, fibromatosis is seen predominantly in young women with a median age of 45 years.23 Rare cases have been reported in males. The patients present with a palpable mass that is clinically and radiologically suspicious for malignancy.24 Bilaterality has been seen in a small proportion of cases.25 Grossly, the tumors are firm, have ill-defined stellate edges, and measure 2 to 3 cm typically, although very large tumors have been reported. Microscopically, the tumors are composed of bland spindle cells and collagenous stroma. The spindle cells can show fascicular, storiform or diffuse architecture with infiltration into adipose tissue, skeletal muscle, or blend with the adjacent breast parenchyma. The tumors can be variably cellular and collagen may be dense, keloidal like, or wispy. The tumor invades around normal lobules and may even appear to wrap them. Peripheral lymphoid aggregates are seen, sometimes with germinal centers. Mitoses and necrosis are rare in mammary fibromatoses26 (Fig. 9).
Fibromatosis shows nuclear β-catenin expression and recently shown to have CTNNB1 (β-catenin) mutation in 9 of 10 sporadic cases of mammary fibromatoses. Even though benign phyllodes tumor and metaplastic carcinoma can show β-catenin expression, they do not show mutation of CTNNB1 gene.27
Local recurrence is quite common in fibromatosis.23 It has been shown that tamoxifen with or without nonsteroidal anti-inflammatory drugs such as sulindac may be useful in treatment of desmoid fibromatosis and that regression in size may be seen in a quarter of such cases. Mammary fibromatosis has also been shown to respond to tamoxifen.24 Wide local excision with clear margins and clinical follow-up is the current first line of treatment.
Biopsy Site Changes/Scar
The histologic findings change with the age of the biopsy site. Within the first week after a biopsy procedure, the site largely consists of granulation tissue and if exuberant, can morphologically resemble a vascular lesion. However, after about 2 weeks, this area has undergone collagenization and wound contraction. The histologic findings at this point consist mainly of fibrosis and scar formation, the appearance of which can be mistaken for fibromatosis or spindle cell metaplastic carcinoma (fibromatosis-like variant). Especially when excluding the possibility of fibromatosis, it is helpful to ascertain whether there is a history of prior biopsy or needling procedure in the area of concern and if so, the time interval since the procedure.
Myofibroblastoma was originally described in males, but since then has been reported to occur in females as well.28 The average age at presentation is 63 years, but it ranges from 41 to 85 years. It presents as a slow growing, usually subareolar palpable nodule. Bilaterality is rare. It appears as well-circumscribed oval or round dense solid mass on mammography. By ultrasonography, these tumors are seen as circumscribed, slightly hypoechoic masses. Magnetic resonance imaging shows homogeneous enhancement.29 Grossly, these appear as well-circumscribed, round to oval, rubbery, pale solid masses without necrosis or hemorrhage.30 Histologically, myofibroblastoma is seen as a circumscribed mass pushing the normal breast parenchyma aside. Occasionally, there may be entrapment of normal lobules at the periphery of the lesion. In its classic form, there are short intersecting fascicles of bland spindle cells within a collagenized background. Collagen may be present as thick bands, chicken-wire pattern, or keloid-like forms (Fig. 10). There may be numerous mast cells sprinkled throughout the tumor. Mitoses are typically 2 or fewer per 10 hpf. Benign heterologous elements such as cartilage, smooth muscle, or bone may be seen as can incorporation of adipose tissue within the tumor. Several variants of myofibroblastoma have been described including an infiltrative variant that can mimic spindle cell metaplastic carcinoma.31
Myofibroblastoma does not recur after complete excision.
Phyllodes Tumors (Stromal Component)
Phyllodes tumors are fibroepithelial neoplasms that can be seen in a wide age range of women including young adolescents and elderly. Bilateral tumors have also been reported.32,33 Rare case reports have also shown that phyllodes tumor can occur in men, sometimes in a state of endocrine imbalance such as after hormonal therapy for prostate cancer.34 Some phyllodes tumor may arise from fibroadenomas, but how often this happens is unknown. A rapid increase in size in a previously stable fibroadenoma should raise the suspicion for phyllodes tumor. Calcifications may be rarely seen.
It is difficult to differentiate between various benign and malignant fibroepithelial tumors by imaging alone. Phyllodes tumor appear grossly as unencapsulated, circumscribed, often multinodular, firm, solid masses with cleft-like spaces. Infarction, cystic degeneration, necrosis, and hemorrhage may be seen, depending upon the grade of the tumor. Microscopically, they are comprised of proliferating, elongated ducts within hypercellular stroma. They are devoid of lobules. There is stromal heterogeneity within the same tumor and some areas may show hyalinized stroma, whereas others may be more cellular. PASH-like areas are quite common in phyllodes tumor. The edges may be infiltrative and are frequently multinodular.
World Health Organization grades phyllodes tumors into 3 categories: benign, borderline (low malignant potential), and malignant. Benign phyllodes tumors have mild stromal atypia and pleomorphism, mitotic rate of <2 per 10 hpf, relatively circumscribed edges with only focal infiltration and multinodularity may be evident. Malignant phyllodes tumors show brisk mitotic activity over 5 per 10 hpf, marked stromal atypia, areas of stromal overgrowth, infiltrative borders and may have malignant heterologous elements such as chondrosarcoma, liposarcoma, or osteosarcoma.35–37 Stromal overgrowth, defined as the presence of stroma without epithelial elements in at least 1 low-power (×4) field has been reported as a reliable indicator of malignancy in phyllodes tumor.38
It is important to grade phyllodes tumors as benign tumors are prone to local recurrence only, but borderline and malignant tumors show both early local recurrence and distant metastases followed by death from the disease in some cases.
Spindle cell lesions of the breast represent a diagnostically challenging subset of entities that requires scrutiny from all fronts including morphologic evaluation, clinicoradiologic correlation, and immunohistochemical characterization. The differential diagnosis consists of uncommon entities that makes this exercise even more daunting to the practicing pathologist. Utilizing an algorithmic approach to this group of morphologically similar but clinically different entities may help to lessen this well-known diagnostic burden.
1. Diaz LK, Wiley EL, Venta LA. Are malignant cells displaced by large-gauge needle core biopsy of the breast? Am J Roentgenol. 1999;173:1303–1313
2. Lacroix-Triki M, Geyer FC, Lambros MB, et al. Beta-catenin/Wnt signalling pathway in fibromatosis, metaplastic carcinomas and phyllodes tumours of the breast. Mod Pathol. 2010;23:1438–1448
3. Chia Y, Thike AA, Cheok PY, et al. Stromal keratin expression in phyllodes tumours of the breast: a comparison with other spindle cell breast lesions. J Clin Pathol. 2012;65:339–347
4. Nga ME, Lim KH, Tan EY, et al. Malignant adenomyoepithelial tumor of the breast: multi-immunolabeling technique and detailed immunophenotypic study. Appl Immunohistochem Mol Morphol. 2008;16:100–104
5. Hungermann D, Buerger H, Oehlschlegel C, et al. Adenomyoepithelial tumours and myoepithelial carcinomas of the breast--a spectrum of monophasic and biphasic tumours dominated by immature myoepithelial cells. BMC Cancer. 2005;5:92
6. Kawaguchi K, Shin SJ. Immunohistochemical staining characteristics of low-grade adenosquamous carcinoma of the breast. Am J Surg Pathol. 2012;36:1009–1020
7. Khan HN, Wyld L, Dunne B, et al. Spindle cell carcinoma of the breast: a case series of a rare histological subtype. Eur J Surg Oncol. 2003;29:600–603
8. Wargotz ES, Norris HJ. Metaplastic carcinomas of the breast. I. Matrix-producing carcinoma. Hum Pathol. 1989;20:628–635
9. Wargotz ES, Deos PH, Norris HJ. Metaplastic carcinomas of the breast. II. Spindle cell carcinoma. Hum Pathol. 1989;20:732–740
10. Wargotz ES, Norris HJ. Metaplastic carcinomas of the breast. IV. Squamous cell carcinoma of ductal origin. Cancer. 1990;65:272–276
11. Weigelt B, Kreike B, Reis-Filho JS. Metaplastic breast carcinomas are basal-like breast cancers: a genomic profiling analysis. Breast Cancer Res Treat. 2009;117:273–280
12. Lien HC, Hsiao YH, Lin YS, et al. Molecular signatures of metaplastic carcinoma of the breast by large-scale transcriptional profiling: identification of genes potentially related to epithelial-mesenchymal transition. Oncogene. 2007;26:7859–7871
13. Pezzi CM, Patel-Parekh L, Cole K, et al. Characteristics and treatment of metaplastic breast cancer: analysis of 892 cases from the National Cancer Data Base. Ann Surg Oncol. 2007;14:166–173
14. Yang WT, Hennessy B, Broglio K, et al. Imaging differences in metaplastic and invasive ductal carcinomas of the breast. Am J Roentgenol. 2007;189:1288–1293
15. Chen IC, Lin CH, Huang CS, et al. Lack of efficacy to systemic chemotherapy for treatment of metaplastic carcinoma of the breast in the modern era. Breast Cancer Res Treat. 2011;130:345–351
16. Gobbi H, Simpson JF, Borowsky A, et al. Metaplastic breast tumors with a dominant fibromatosis-like phenotype have a high risk of local recurrence. Cancer. 1999;85:2170–2182
17. Gobbi H, Simpson JF, Jensen RA, et al. Metaplastic spindle cell breast tumors arising within papillomas, complex sclerosing lesions, and nipple adenomas. Mod Pathol. 2003;16:893–901
18. Denley H, Pinder SE, Tan PH, et al. Metaplastic carcinoma of the breast arising within complex sclerosing lesion: a report of five cases. Histopathology. 2000;36:203–209
19. Sneige N, Yaziji H, Mandavilli SR, et al. Low-grade (fibromatosis-like) spindle cell carcinoma of the breast. Am J Surg Pathol. 2001;25:1009–1016
20. Haggitt RC, Booth JL. Bilateral fibromatosis of the breast in Gardner’s syndrome. Cancer. 1970;25:161–166
21. Hammoudeh ZS, Darian VB. Desmoid tumor (fibromatosis) of the breast after augmentation with saline implants. Plast Reconstr Surg. 2012;129:753e–754e
22. Jamshed S, Farhan MI, Marshall MB, et al. Fibromatosis of the breast after mammary prosthesis implantation. Clin Adv Hematol Oncol. 2008;6:687–694
23. Neuman HB, Brogi E, Ebrahim A, et al. Desmoid tumors (fibromatoses) of the breast: a 25-year experience. Ann Surg Oncol. 2008;15:274–280
24. Plaza MJ, Yepes M. Breast fibromatosis response to tamoxifen: dynamic MRI findings and review of the current treatment options. J Radiol Case Rep. 2012;6:16–23
25. Taylor TV, Sosa J. Bilateral breast fibromatosis: case report and review of the literature. J Surg Educ. 2011;68:320–325
26. Wargotz ES, Norris HJ, Austin RM, et al. Fibromatosis of the breast. A clinical and pathological study of 28 cases. Am J Surg Pathol. 1987;11:38–45
27. Kim T, Jung EA, Song JY, et al. Prevalence of the CTNNB1 mutation genotype in surgically resected fibromatosis of the breast. Histopathology. 2012;60:347–356
28. Wargotz ES, Weiss SW, Norris HJ. Myofibroblastoma of the breast. Sixteen cases of a distinctive benign mesenchymal tumor. Am J Surg Pathol. 1987;11:493–502
29. Greenberg JS, Kaplan SS, Grady C. Myofibroblastoma of the breast in women: imaging appearances. Am J Roentgenol. 1998;171:71–72
30. Magro G. Mammary myofibroblastoma: a tumor with a wide morphologic spectrum. Arch Pathol Lab Med. 2008;132:1813–1820
31. Rosen PP. Benign mesenchymal neoplasms. Rosen’s Breast Pathology. 2008 Philadelphia Lippincott Williams & Wilkins:852–855
32. Ball AB, Fisher C. Bilateral malignant phyllodes tumours. J Surg Oncol. 1991;46:67–69
33. Saleh HA, Klein LH. Cystosarcoma phyllodes arising synchronously in right breast and bilateral axillary ectopic breast tissue. Arch Pathol Lab Med. 1990;114:624–626
34. Johansson L, Balldin G. Malignant cystosarcoma phyllodes in a man treated with polyestradiolphosphate. Case report. Acta Chir Scand. 1986;152:781–785
35. Vera-Sempere F, Garcia-Martinez A. Malignant phyllodes tumor of the breast with predominant chondrosarcomatous differentiation. Pathol Res Pract. 2003;199:841–845
36. Kracht J, Sapino A, Bussolati G. Malignant phyllodes tumor of breast with lung metastases mimicking the primary. Am J Surg Pathol. 1998;22:1284–1290
37. Silver SA, Tavassoli FA. Osteosarcomatous differentiation in phyllodes tumors. Am J Surg Pathol. 1999;23:815–821
38. Ward RM, Evans HL. Cystosarcoma phyllodes. A clinicopathologic study of 26 cases. Cancer. 1986;58:2282–2289
© 2013 Lippincott Williams & Wilkins, Inc.