Search

BACKGROUND Using fine-needle aspiration (FNA) smears, it is difficult to distinguish low grade phyllodes tumor (PT) from fibroadenoma (FA) due to overlapping cytologic features between the two lesions. The authors retrospectively studied 45 histologically proven fibroepithelial breast tumors of which 33 were FA and 12 were PT (1 malignant, 8 borderline, and 3 benign) to define cytologic features that can help in the accurate categorization of these lesions by using FNA samples. METHODS The cytologic features analyzed included: 1) epithelial component for number ( 5), architecture, apocrine metaplasia, squamous metaplasia, nuclear pleomorphism, and mitosis; 2) stromal fragments for number ( 5), cellularity (on a scale of 1+ to 3+), borders, cell characteristics, nuclear pleomorphism, and mitosis; 3) individual dispersed stromal cells in the background for cellularity (on a scale of 1+ to 3+), and cellular shape (short/round/oval or long spindle) based on whether they were smaller or larger than 2 times the size of a small round lymphocyte. RESULTS The mean age of patients with FA was 34 years and of those with PT 44 years. The average size of FA was 2.0 cm, and the average size of PT was 4.0 cm. The characteristics of the epithelial fragments of PT and FA were not significantly different. Stromal fragments were noted in 60% of FA and 83% of PT samples examined. Fifty-six percent of PT and 30% of FA exhibited hypercellular stromal fragments (3+ cellularity), and the difference was not statistically significant. Large club-shaped hypercellular stromal fragments were present only in FA (in 21% of the samples). There was no difference in the overall cellularity of the background stromal nuclei in the two types of lesions. Long spindle nuclei averaging greater than 30% of the dispersed stromal cell population in the background were found only in cases of PT (in 57% of the samples; P < 0.001). Short/round/oval nuclei characterized most FAs. Long spindle nuclei constituting 10–30% of the dispersed stromal cells, however, occurred in both PT and FA to the extent of 43% and 21%, respectively. CONCLUSIONS Hypercellular stromal fragments occur not only in PT, but also in FA, and hence they cannot be used as the sole criterion for making a diagnosis of PT on FNA. The proportion of individual long spindle nuclei (>30%) amid the dispersed stromal cells in the background is the most reliable discriminator between the two lesions. Lesions in which long spindle nuclei constitute between 10% and 30% may represent either PT or FA, and therefore such lesions should be categorized as indeterminate on FNA. Cancer (Cancer Cytopathol) 2000;90:342–349. © 2000 American Cancer Society.

BACKGROUND Cystosarcoma phyllodes is a rare sarcoma of the breast. Although surgical removal is the mainstay of treatment, the extent of surgery required (excision vs. mastectomy) and the need for additional local therapy, such as radiotherapy, are unclear. The current study evaluated the rate of local and distant failure, as well as potential prognostic factors, to better define appropriate treatment strategies. METHODS One hundred one patients treated primarily for cystosarcoma phyllodes of the breast were evaluated. These tumors were classified histologically into benign (58%), indeterminate (12%), and malignant (30%) based on well defined criteria. Stromal overgrowth (29%) was considered separately. Surgery was comprised of local excision with breast conservation (47%) or mastectomy (53%). Microscopic surgical margins were negative in 99% of cases. Six patients received adjuvant radiotherapy. RESULTS Overall survival for the 101 patients was 88%, 79%, and 62% at 5, 10, and 15 years, respectively. For patients with nonmalignant (benign or indeterminate) and malignant cystosarcoma phyllodes, the overall survival was 91% and 82%, respectively, at 5 years, and 79% and 42%, respectively, at 10 years. Similar rates were observed based on the presence or absence of stromal overgrowth. Local recurrence occurred in 4 patients, with an actuarial 10-year rate of 8%. Eight patients developed distant metastases, with an actuarial 10-year rate of 13%. Multivariate analysis using Cox proportional hazards regression revealed stromal overgrowth to be the only independent predictor of distant failure. CONCLUSIONS Local failure in this group of largely margin negative patients with cystosarcoma phyllodes of the breast was low, showing that breast-conserving surgery with appropriate margins is the preferred primary therapy. The current study data do not support the use of adjuvant radiotherapy for patients with adequately resected disease. Patients with stromal overgrowth, particularly when the tumor size was > 5 cm, were found to have a high rate of distant failure; such patients merit consideration of a trial that examines the efficacy of systemic therapy. Cancer 2000;89:1502–11. © 2000 American Cancer Society.

BACKGROUND. Fine-needle aspiration biopsy (FNA) has been successful in diagnosing epithelial lesions of the breast. Its role in the evaluation of spindle cell and mesenchymal lesions of the breast, which include a variety of benign and malignant conditions, is less clear. This article discusses the cytologic features and differential diagnosis of these lesions, as well as the potential diagnostic pitfalls associated with them. METHODS. FNAs of the breast, in which a spindle cell or mesenchymal component was a key or dominant feature, were retrieved. Fibroadenomas without cellular stroma and typical lipomas were excluded. RESULTS. Forty-six aspirates (0.87%) in a series of 5306 breast FNAs contained a significant spindle cell or mesenchymal component. The aspirates were classified into 4 categories: 1) reactive conditions, including 2 diabetic mastopathies, 3 granulation tissue specimens, and 7 granulomatous lesions; 2) benign neoplastic conditions, including I mammary hamartoma, 1 dermatofibroma, 1 fibromatosis, 2 granular cell tumors, 2 angiolipomas, and 7 cellular fibroadenomas; 3) low grade malignant neoplastic lesions, including 10 low grade phyllodes tumors; and 4) high grade malignant neoplastic lesions, including 1 metaplastic carcinoma with chondroid stroma, 1 pleomorphic liposarcoma, 2 malignant fibrous histiocytomas, 2 osteosarcomas, and 4 metastatic melanomas. A specific diagnosis was rendered in 38 cases (82.6%). The mammary hamartoma was diagnosed as fibrocystic changes; the dermatofibroma as benign spindle cell lesion, not otherwise specified (NOS); and the primary osteosarcoma as an atypical spindle cell proliferation, NOS. The reactive ductal epithelial cells in one of the granulomatous mastitis specimens, as well as the hyperplastic ductal epithelial cells in one of the phyllodes tumors, were interpreted as atypical ductal proliferation. The marked cytologic atypia displayed by one granular cell tumor was interpreted as low grade adenocarcinoma and the primary liposarcoma as poorly differentiated carcinoma. CONCLUSIONS. Breast lesions with a significant spindle cell or mesenchymal component are rarely encountered in FNA and constitute a heterogeneous group that may pose a diagnostic dilemma. FNA should be the initial diagnostic procedure for investigating these lesions, as a specific diagnosis was rendered in the majority of cases.

BACKGROUND Fine-needle aspiration (FNA) biopsy is an established, highly accurate method for diagnosing palpable breast lesions. However, in some cases a definitive diagnosis cannot be made by FNA alone, either due to the inherent limitations of cytology itself or the ability to obtain adequate material for diagnosis. This study evaluated the usefulness of a supplemental core needle biopsy performed by a cytopathologist in conjunction with an FNA biopsy in select patients. METHODS Twelve patients with palpable breast lesions underwent the combined FNA/core needle biopsy procedure during the study period from March 1995 through March 1996. All 12 patients first had an FNA aspiration that was preliminarily evaluated by the FNA cytopathologist while the patient was at the clinic. If the cytopathologist was unable to render a definitive diagnosis of the lesion type or if a repeat FNA was requested by the clinician because of a previous non-definitive FNA result, a core needle biopsy was then performed. RESULTS When the FNA and the core needle biopsies were reviewed together, a correct definitive diagnosis was made in 10 of the 12 cases. In contrast, review of the FNA material alone yielded a correct definitive diagnosis in only five cases. CONCLUSIONS The supplemental core needle biopsy was found to be especially useful in: 1) providing a definitive diagnosis of infiltrating carcinoma in those cases in which the FNA diagnosis was reported as “suspicious;” 2) providing ample tissue for ancillary studies; and 3) differentiating a phyllodes tumor from a fibroadenoma. It is the authors' opinion that the FNA cytopathologist is the physician best qualified to perform the combined FNA/core needle biopsy procedure should he/she deem it necessary. Cancer (Cancer Cytopathol) 1997; 81:33-9. © 1997 American Cancer Society

BACKGROUND This study addresses the controversial prognostic and therapeutic aspects of phyllodes tumor of the breast. METHODS Records of 170 women with phyllodes tumor of the breast were reviewed. On the basis of the criteria proposed by Azzopardi and Salvadori et al., including estimation of tumor margin, growth of the connective tissue component, mitoses, and cellular atypia, the entire series was divided into three histotypes of phyllodes tumor, i.e., benign (92 cases, 54.1%), borderline (19 cases, 11.2%), and malignant (59 cases, 34.7%). Ninety-eight patients (57.6%) were treated by wide local excision (79 benign, 15 borderline, and 4 malignant), 43 (25.3%) by simple mastectomy (13 benign, 4 borderline, and 26 malignant), and 29 (17.1%) by radical mastectomy (all malignant). RESULTS Of the 170 treated patients, 141 (82.9%) survived 5 years without evidence of disease. In the Cox multivariate analysis the histotype of the tumor was the only independent prognostic factor: 5-year NED survival was observed in 95.7% of the patients with benign phyllodes tumor, 73.7% with borderline phyllodes tumor, and 66.1% with malignant phyllodes tumor. After a wide local excision 98.7% of the patients with benign tumor, and 80% with borderline tumor, were cured. Local recurrence was found in 14 patients (8.2%) (4 benign, 3 borderline, and 7 malignant); 10 of these underwent reoperation (7 wide local excision, 3 radical mastectomy) and survived 5 years NED. CONCLUSIONS The histotype of phyllodes tumor (benign, borderline, and malignant), assesed on the basis of the criteria proposed by Azzopardi and Salvadori et al., was the only prognostic factor in our group of patients. Based on the data from literature and our own observations, we observed that a wide local excision, with an adequate margin of normal breast tissue, is the preferred initial therapy for phyllodes tumor of the breast. Cancer 1996;77:910-6.

Background. The histogeneses of fibroadenoma and phyllodes tumor of the breast appear to be closely related, but it is still unclear whether fibroadenoma can progress directly to phyllodes tumor. Methods. This issue was studied by conducting clonal analysis of fibroadenoma and phyllodes tumors that were obtained sequentially from the same patient. One patient developed local recurrence of phyllodes tumor twice, and the other two patients each developed a phyllodes tumor after excision of a primary fibroadenoma. The method for clonal analysis was based on trinucleotide repeat polymorphism of the X chromosome-linked androgen receptor (AR) gene and on random inactivation of the gene by methylation. Results. Clonal analysis revealed that all the three primary fibroadenomas were monoclonal and all four recurrent phyllodes tumors were also monoclonal in origin. In addition, the same allele of the AR gene was inactivated in fibroadenoma and phyllodes tumor(s) in each patient. The probability that phyllodes tumors of different origin happen to inactivate the same allele of the AR gene as fibroadenomas in every case is quite low. Rather, it is more reasonable to assume that the phyllodes tumor has the same origin as fibroadenoma. Conclusions. These results identified monoclonal fibroadenomas that can progress to phyllodes tumors. Cancer 1995; 76:1779–85.

Background. Cystosarcoma phyllodes (CP) is an uncommon fibroepithelial breast neoplasm that rarely involves the central nervous system (CNS). Methods. The authors encountered two patients with CNS metastases and reviewed the literature to identify reports of CP affecting the nervous system. Patients must have been reported in sufficient detail to characterize clinical course, extent of tumor burden, and histologic type of the primary tumor and metastasis to be included in the study. Results. Eleven cases were identified from the literature, but only six were presented in sufficient detail for analysis. Including the current patients, the mean age of patients with CP metastases was 52 years. The time between initial diagnosis and CNS involvement was 5 years. Parenchymal brain metastases respond poorly to treatment, with a mean survival of 29 days, whereas survival is significantly longer with CNS involvement secondary to epidural extension of tumor from adjacent bony structures. Local recurrence or widespread metastases nearly always were present at the time of CNS involvement. Conclusions. Brain metastases from CP are refractory to therapy and carry a dismal prognosis. Patients with CP and brain metastases should undergo brain biopsy to exclude metastases from associated epithelial breast tumors, which have significantly better prognoses. Patients with compressive lesions arising from bone have a better prognosis and may benefit from aggressive management. Cancer 1993; 72:126–130.

Background. Malignant cystosarcoma phyllodes tumors of the breast are rare fibroepithelial tumors. The epidemiologic features of these malignancies in a defined population have not previously been described. Methods. Incidence data were collected from 1972–1989 by the Cancer Surveillance Program, the population-based cancer registry for Los Angeles County. Results. The average annual age-adjusted incidence rate of malignant cystosarcoma phyllodes is 2.1 per 1 million women. Latina whites have a higher risk of this cancer than other racial-ethnic groups. The age-incidence curve shows that risk of cystosarcoma phyllodes peaks in the 45–49-year-old age group, but the age patterns vary by race-ethnicity, with Asian and Latina patients significantly younger, on average, than non-Latina white patients. The incidence rates of malignant cystosarcoma phyllodes were substantially higher in the 1980s than in the 1970s, particularly among Latina whites and Asians. For Latina whites, birthplace is a significant predictor of risk. Women born in Mexico and Central and South America are at threefold to fourfold greater risk of malignant cystosarcoma phyllodes tumors than Latina whites born in the United States. Conclusions. The epidemiology of malignant cystosarcoma phyllodes tumors of the breast is strikingly different from that of the more common histologic types of breast cancer. The high risk in Latina immigrants may offer an important clue as to the cause of this rare tumor.

An unique case of high-grade phyllodes tumor (cystosarcoma phyllodes) arising in the seminal vesicle is reported. A 61-year-old man had symptoms of urinary obstruction, and ultrasonography and computed tomography showed a large mass posterior to the bladder. Using cystoprostatoseminovesiculectomy, a high-grade phyllodes tumor of the left seminal vesicle was found that did not involve the bladder or prostate. Histologically, the tumor consisted of benign irregular slit-like glands set in a mitotically active cellular sarcomatous stroma. The epithelium displayed intense cytoplasmic immunoreactivity with all keratin proteins (AE1/AE3, CAM 5.2, and high-molecular-weight keratin [clone 34 beta E12]). The stromal cells were strongly positive for vimentin, and approximately 30% were positive for muscle-specific actin and desmin. Four years after resection, a lung metastasis was removed, which was histologically and immunohistochemically identical to the seminal vesicle tumor. This case represents the malignant end of the range of phyllodes tumors (cystosarcoma phyllodes). Like similar tumors in the breast and prostate, phyllodes tumors of the seminal vesicle should be considered high-grade lesions (malignant) in which there is significant mitotic activity, stromal pleomorphism, and stromal overgrowth.

BACKGROUND: Cytology and core-needle biopsies are not always sufficient to exclude malignancy in benign breast lesions (BBL) that are at risk of developing cancer, and open biopsy often is mandatory. In screening programs, open biopsies performed for lesions that are at risk of developing malignancy are considered benign. The authors of this report evaluated the impact of the screen-detected BBL at risk of developing cancer that were counted in the quota of benign breast open biopsies in the Breast Cancer Screening Program of Verona. METHODS: Benign open biopsies were subdivided into 4 groups according to their risk of developing cancer: Histo1, normal histology; Histo2, ‘pure’ BBL (fibroadenoma, fibrocystic disease, mastitis, adenosis); Histo3, BBL with a low risk of developing cancer (radial scar, papilloma, papillomatosis, phyllodes tumor, mucocele-like lesion); and Histo4, BBL with a high risk of developing cancer (atypical columnar cell hyperplasia, atypical ductal hyperplasia, atypical lobular hyperplasia). RESULTS: Of 510 open biopsies, 83 biopsies were benign, and the ratio of benign to malignant biopsies was 1:5. Histo1 was observed in 4.8% of all benign open biopsies, Histo2 was observed in 37.4%, Histo3 was observed in 31.3%, and Histo4 was observed 26.5%. CONCLUSIONS: BBL at risk of developing cancer may be numerous in screening programs. It is inappropriate to include BBL at risk of developing cancer in the overall benign open biopsy rate. The authors propose separating pure BBL from lesions at higher risk of developing cancer. To date, there is no evidence to support the premise that detecting high-risk proliferative lesions leads to benefits in terms of reduced mortality; however, these lesions need to be counted separately for future evaluations. Cancer 2009. © 2008 American Cancer Society.

The authors studied prognostic factors in 77 patients with primary cystosarcoma phyllodes (CSP) of the breast. Median patient age was 50 years of age, and the median follow-up time was 8 years. Sixteen patients (21'/0) had distant metastases and subsequently died of CSP. Clinical variables such as age, symptom duration, clinical tumor size, and type of surgery were not of prognostic value. Local recurrence was more common among patients treated with breast-conserving surgery than among those treated with mastectomy. However, there was no significant difference between these two subgroups in terms of distant metastasisfree survival or overall survival. The prognostic significance of several histopathologic parameters was also assessed, e.g., stromal cellularity, stromal cellular atypism, mitotic activity, atypic mitoses, stromal overgrowth, tumor contour, tumor necrosis, and heterologous stromal elements. In a multivariate Cox analysis, the only features that were found to be independent prognostic factors were tumor necrosis (P < 0.05) and presence of stromal elements other than fibromyxoid tissue (P < 0.01). In summary, additional studies of prognostic factors in CSP are warranted because of the conflicting results in published reports. Cancer 68:2017-2022,1991. YSTOSARCOMA PHYLLODES (CSP) is a relatively un

An extremely rare neoplasm of the prostate, commonly known as phyllodes type of atypical prostatic hyperplasia, is described in a 57-year-old man. Over a 2-year period, the atypical smooth muscle component has transformed into leiomyosarcoma, and eventually metastasized to the bones and the lungs. Such an evolution has never been documented for this tumor.

BACKGROUND. Malignant phyllodes tumor is a rare and potentially aggressive breast neoplasm. Little information is available regarding the optimal management of these lesions and rarer still are data regarding survival. The current study used a large population database to determine prognostic factors that predict cause-specific survival (CSS). METHODS. Data were obtained from the Surveillance, Epidemiology, and End Results Program (SEER) for the years 1983–2002. Women receiving resection for primary nonmetastatic malignant phyllodes tumor of the breast were included (n = 821). Analyses of patient, pathologic, and treatment characteristics were performed using univariate and multivariate Cox regression analyses for the CSS endpoint. RESULTS. With a median follow-up of 5.7 years, CSS was 91%, 89%, and 89%, at 5, 10, and 15 years, respectively. Mastectomy was performed in 428 women (52%) and wide excision or lumpectomy in 393 (48%). Women undergoing mastectomy were significantly older (P = .004) and had larger tumors (P = .009). Wide excision was associated with equivalent or improved CSS relative to mastectomy on univariate and multivariate analyses. Older age predicted for cause-specific mortality on multivariate analysis. Adjuvant radiotherapy (RT) predicted for worse CSS when implemented compared with surgery alone. CONCLUSIONS. Mastectomy was not found to provide a benefit in CSS compared with wide excision in malignant phyllodes tumor of the breast. Women undergoing wide excision had at the minimum similar cancer-specific mortality compared with those who received mastectomy. The role of adjuvant RT is uncertain and requires further investigation. Cancer 2006. © 2006 American Cancer Society.

BACKGROUND The authors reviewed smears from fine-needle aspiration biopsies (FNAB) diagnosed as “papillary lesions” and correlated the cytologic findings with the final diagnoses at excision. The objective of the current study was to determine the accuracy of FNAB diagnosis of a papillary lesion in distinguishing true papillary from nonpapillary proliferations and to evaluate cytologic criteria for the distinction of papillomas from true papillary malignancies and their cytologic look-alikes. METHODS The cytopathology database at the New York University Medical Center was searched for women who underwent surgical excision after a breast FNAB diagnosis of a papillary lesion. The FNAB smears and corresponding slides from excisional biopsies were reviewed. The smears were evaluated and graded for the following features: cellularity, architecture, presence of fibrovascular cores, single cells, columnar cells, cellular atypia, myoepithelial cells, foamy histiocytes, and apocrine cells. The F test was used to determine the statistical significance of differences between true benign papillary lesions (papilloma) and adenocarcinomas (in situ and invasive). RESULTS At the time of excision, 46 (66 %) cases were benign (23 solitary intraductal papillomas, 6 intraductal papillomatosis, 11 examples of fibrocystic change, and 6 fibroadenomas) and 24 (34 %) were malignant (1 low-grade phyllodes tumor [PT], 23 ductal in situ and invasive carcinomas). Of the 23 carcinomas, 3 (13 %) were classified as benign papillary lesions on FNAB and 19 (86 %) were classified as either atypical or suspicious. One case of low- grade PT originally was classified as benign on FNAB. There were four false-negative diagnoses; two were due to sampling and two to interpretative errors. A portion of the lesions classified as papillary were fibroadenomas and examples of fibrocystic change on excision and all of these were correctly classified as benign on FNAB. Of the histologically proven papillomas, 62% were correctly classified as benign on FNAB and none were designated as being positive for malignancy. Statistically significant features of distinction between papillomas and carcinomas included cellularity (P = 0.016), cellular atypia (P = 0.0053), and the presence of cytologically bland columnar cells (P = 0.04). Low-grade ductal carcinoma in situ (cribriform and micropapillary types) and tubular carcinoma represented the most difficult differential diagnostic problems. CONCLUSIONS A significant portion of lesions displaying a papillary pattern on FNAB are nonpapillary on follow-up. Among benign processes, fibrocystic change and fibroadenoma may closely simulate papilloma on cytology. However, in spite of the overlapping features of true papillary lesions and their cytologic look-alikes, the majority can be classified accurately into benign or atypical (and above) categories by FNAB. Lesions that fall short of a definitive benign diagnosis should be placed into an indeterminate category. This approach will guide the surgeon to provide better patient management. Cancer (Cancer Cytopathol) 2003;99:156–65. © 2003 American Cancer Society.

Cystosarcoma phyllodes is a rare sarcoma of the breast. Although surgical removal is the mainstay of treatment, the extent of surgery required (excision vs. mastectomy) and the need for additional local therapy, such as radiotherapy, are unclear. The current study evaluated the rate of local and distant failure, as well as potential prognostic factors, to better define appropriate treatment strategies.One hundred one patients treated primarily for cystosarcoma phyllodes of the breast were evaluated. These tumors were classified histologically into benign (58%), indeterminate (12%), and malignant (30%) based on well defined criteria. Stromal overgrowth (29%) was considered separately. Surgery was comprised of local excision with breast conservation (47%) or mastectomy (53%). Microscopic surgical margins were negative in 99% of cases. Six patients received adjuvant radiotherapy.Overall survival for the 101 patients was 88%, 79%, and 62% at 5, 10, and 15 years, respectively. For patients with nonmalignant (benign or indeterminate) and malignant cystosarcoma phyllodes, the overall survival was 91% and 82%, respectively, at 5 years, and 79% and 42%, respectively, at 10 years. Similar rates were observed based on the presence or absence of stromal overgrowth. Local recurrence occurred in 4 patients, with an actuarial 10-year rate of 8%. Eight patients developed distant metastases, with an actuarial 10-year rate of 13%. Multivariate analysis using Cox proportional hazards regression revealed stromal overgrowth to be the only independent predictor of distant failure.Local failure in this group of largely margin negative patients with cystosarcoma phyllodes of the breast was low, showing that breast-conserving surgery with appropriate margins is the preferred primary therapy. The current study data do not support the use of adjuvant radiotherapy for patients with adequately resected disease. Patients with stromal overgrowth, particularly when the tumor size was5 cm, were found to have a high rate of distant failure; such patients merit consideration of a trial that examines the efficacy of systemic therapy.

Fine-needle aspiration biopsy (FNA) has been successful in diagnosing epithelial lesions of the breast. Its role in the evaluation of spindle cell and mesenchymal lesions of the breast, which include a variety of benign and malignant conditions, is less clear. This article discusses the cytologic features and differential diagnosis of these lesions, as well as the potential diagnostic pitfalls associated with them.FNAs of the breast, in which a spindle cell or mesenchymal component was a key or dominant feature, were retrieved. Fibroadenomas without cellular stroma and typical lipomas were excluded.Forty-six aspirates (0.87%) in a series of 5306 breast FNAs contained a significant spindle cell or mesenchymal component. The aspirates were classified into 4 categories: 1) reactive conditions, including 2 diabetic mastopathies, 3 granulation tissue specimens, and 7 granulomatous lesions; 2) benign neoplastic conditions, including 1 mammary hamartoma, 1 dermatofibroma, 1 fibromatosis, 2 granular cell tumors, 2 angiolipomas, and 7 cellular fibroadenomas; 3) low grade malignant neoplastic lesions, including 10 low grade phyllodes tumors; and 4) high grade malignant neoplastic lesions, including 1 metaplastic carcinoma with chondroid stroma, 1 pleomorphic liposarcoma, 2 malignant fibrous histiocytomas, 2 osteosarcomas, and 4 metastatic melanomas. A specific diagnosis was rendered in 38 cases (82.6%). The mammary hamartoma was diagnosed as fibrocystic changes; the dermatofibroma as benign spindle cell lesion, not otherwise specified (NOS); and the primary osteosarcoma as an atypical spindle cell proliferation, NOS. The reactive ductal epithelial cells in one of the granulomatous mastitis specimens, as well as the hyperplastic ductal epithelial cells in one of the phyllodes tumors, were interpreted as atypical ductal proliferation. The marked cytologic atypia displayed by one granular cell tumor was interpreted as low grade adenocarcinoma and the primary liposarcoma as poorly differentiated carcinoma.Breast lesions with a significant spindle cell or mesenchymal component are rarely encountered in FNA and constitute a heterogeneous group that may pose a diagnostic dilemma. FNA should be the initial diagnostic procedure for investigating these lesions, as a specific diagnosis was rendered in the majority of cases. Cancer (Cancer Cytopathol)

The global DNA methylation of 136 breast lesions (117 primary invasive carcinomas, 5 benign phyllodes tumors, 11 fibroadenomas, and 3 sclerosing adenosis) and their respective adjacent parenchyma was analyzed using an in vitro enzyme assay.In the group of patients with breast carcinoma, DNA hypomethylation was correlated with clinical and pathologic parameters known to affect disease prognosis. Histopathologic type, disease stage, and tumor grade were evaluated according to the World Health Organization classification, the TNM system, and the criteria of Elston and Ellis' criteria, respectively. DNA flow cytometry was performed in fresh/frozen samples stained with propidium iodide. Hormone receptor (estrogen and progesterone receptor) status was determined by immunocytochemistry.The comparative study of DNA methylation showed that the DNA of breast carcinomas was statistically significantly less methylated than the DNA of the respective adjacent parenchyma (P=0.0001), the DNA of breast benign lesions (P=0.0002), and the DNA of normal parenchyma (P0.0001). A statistically significant correlation was found between the global DNA hypomethylation and the disease stage (P=0.0009), tumor size (P=0.0026), and histologic grade (P=0.0097) of malignant neoplasms. A trend for DNA from breast carcinomas with positive axillary lymph nodes (N1) to be more hypomethylated than those without nodal involvement (NO) (P=0.055) was verified. In contrast, no significant association was found between DNA methylation and histologic type of tumors, hormone receptors, DNA ploidy, and S-phase fraction.The current shows that DNA hypomethylation is increased in breast carcinomas, playing a potentially important role in tumor development. These findings also suggest that DNA methylation status may be a biologic marker with prognostic significance in this group of neoplasms.

The histogeneses of fibroadenoma and phyllodes tumor of the breast appear to be closely related, but it is still unclear whether fibroadenoma can progress directly to phyllodes tumor.This issue was studied by conducting clonal analysis of fibroadenoma and phyllodes tumors that were obtained sequentially from the same patient. One patient developed local recurrence of phyllodes tumor twice, and the other two patients each developed a phyllodes tumor after excision of a primary fibroadenoma. The method for clonal analysis was based on trinucleotide repeat polymorphism of the X chromosome-linked androgen receptor (AR) gene and on random inactivation of the gene by methylation.Clonal analysis revealed that all the three primary fibroadenomas were monoclonal and all four recurrent phyllodes tumors were also monoclonal in origin. In addition, the same allele of the AR gene was inactivated in fibroadenoma and phyllodes tumor(s) in each patient. The probability that phyllodes tumors of different origin happen to inactivate the same allele of the AR gene as fibroadenomas in every case is quite low. Rather, it is more reasonable to assume that the phyllodes tumor has the same origin as fibroadenoma.These results identified monoclonal fibroadenomas that can progress to phyllodes tumors.

Cystosarcoma phyllodes (CP) is an uncommon fibroepithelial breast neoplasm that rarely involves the central nervous system (CNS).The authors encountered two patients with CNS metastases and reviewed the literature to identify reports of CP affecting the nervous system. Patients must have been reported in sufficient detail to characterize clinical course, extent of tumor burden, and histologic type of the primary tumor and metastasis to be included in the study.Eleven cases were identified from the literature, but only six were presented in sufficient detail for analysis. Including the current patients, the mean age of patients with CP metastases was 52 years. The time between initial diagnosis and CNS involvement was 5 years. Parenchymal brain metastases respond poorly to treatment, with a mean survival of 29 days, whereas survival is significantly longer with CNS involvement secondary to epidural extension of tumor from adjacent bony structures. Local recurrence or widespread metastases nearly always were present at the time of CNS involvement.Brain metastases from CP are refractory to therapy and carry a dismal prognosis. Patients with CP and brain metastases should undergo brain biopsy to exclude metastases from associated epithelial breast tumors, which have significantly better prognoses. Patients with compressive lesions arising from bone have a better prognosis and may benefit from aggressive management.

An unique case of high-grade phyllodes tumor (cystosarcoma phyllodes) arising in the seminal vesicle is reported. A 61-year-old man had symptoms of urinary obstruction, and ultrasonography and computed tomography showed a large mass posterior to the bladder. Using cystoprostatoseminovesiculectomy, a high-grade phyllodes tumor of the left seminal vesicle was found that did not involve the bladder or prostate. Histologically, the tumor consisted of benign irregular slit-like glands set in a mitotically active cellular sarcomatous stroma. The epithelium displayed intense cytoplasmic immunoreactivity with all keratin proteins (AE1/AE3, CAM 5.2, and high-molecular-weight keratin [clone 34 beta E12]). The stromal cells were strongly positive for vimentin, and approximately 30% were positive for muscle-specific actin and desmin. Four years after resection, a lung metastasis was removed, which was histologically and immunohistochemically identical to the seminal vesicle tumor. This case represents the malignant end of the range of phyllodes tumors (cystosarcoma phyllodes). Like similar tumors in the breast and prostate, phyllodes tumors of the seminal vesicle should be considered high-grade lesions (malignant) in which there is significant mitotic activity, stromal pleomorphism, and stromal overgrowth.

Background. The breast is rarely primary site for extranodal malignant lymphoma. Most reported primary non-Hodgkin malignant lymphomas of the breast (PBL) are of B-cell phenotype. Methods. Histologic and immunohistochemical analyses of seven patients with PBL and statistical analysis of 152 patients with PBL reported in the Japanese literature were performed. Results. Malignant lymphoma could not be predicted preoperatively with clinical and radiologic findings; breast carcinoma, fibroadenoma, and phyllodes tumor were the preoperative diagnoses. All patients were women; they ranged in age from 31 to 80 years (mean, 57.6 years). The right breast was involved initially in five patients. In four, only the breast was involved (Stage I), whereas in three, the ipsilateral axillary lymph nodes (Stage II) were involved at diagnosis. According to the Working Formulation, all patients belonged to the intermediate grade and were classified as having diffuse large cell (five patients) or mixed (two patients) lymphoma. Immunophenotypic analysis revealed that all patients had B-cell lymphoma. No patients had lymphoepithelial lesions, which is the characteristic feature in categorizing lymphoma as mucosa-associated lymphoid tissue (MALT) lymphoma. A statistical analysis of the patient reported in the Japanese literature has divided PBL into two types: bilateral type that affects younger women and unilateral type that has broad age distribution, but preponderantly occurs in older women. The age and stage at diagnosis were significant prognostic factors in predicting the survival time, but the location and size of the tumor at initial presentation, histopathologic type, terminal leukemic manifestation, and treatment modality were not. Conclusions. This study indicates that most PBL are diffuse large cell lymphoma of B-cell phenotype and that the age and stage at diagnosis are significant prognostic factors.

The known spatial interaction between normal breast epithelium and its surrounding stroma prompted an investigation of the spatial relationship between stromal mitoses and the epithelial component of fibroepithelial tumors of the breast. The authors applied a novel computerized morphometric technique to routinely processed histologic sections of 23 fibroepithelial tumors (13 fibroadenomas and 10 phyllodes tumors). The proportional area of epithelium in successive concentric annuli surrounding stromal mitoses was measured, and its distribution was compared with that around suitable control points.The authors found that stromal mitotic activity in these tumors was significantly more likely to occur close to rather than remote from the epithelial component, with a significant excess of epithelium around mitoses compared with control points within a range of 79 microns. Essentially similar findings were obtained when randomly identified fibroblast nuclei were used as control points, thus obviating variations in stromal cell density with distance from epithelium as an explanation for the findings.These findings support the hypothesis that stromal growth in fibroepithelial tumors depends, to a variable extent, on the epithelial component. An interaction in the opposite direction (i.e., the stroma providing the growth support to the epithelium) also may occur, but this was not investigated. It is suggested that there is an interdependence of growth between the epithelial and stromal components in these tumors that explains their complex morphology and that stromal dependence on epithelium is lost with increasing malignancy of the stromal elements.

Metastases from cystosarcoma phyllodes are rare, and treatment generally is ineffective. Four patients were treated with ifosfamide (alone in three and combined with doxorubicin in one). Two patients had complete remissions that lasted 26 and 61+ months. One other patient had a partial response that lasted 13 months. The complete responders were both treated as soon as metastases appeared, when they had only a small volume of disease. This appears to represent a significant improvement on other described regimens for this condition, and further trials of ifosfamide are warranted. Close follow-up of patients at high risk for metastases is suggested.

A retrospective study of 33 patients with cystosarcoma phyllodes was done. Eight of these patients had metastases, and the clinical and histologic criteria predicting the development of metastases were examined. The most reliable predictor was the presence of stromal overgrowth; this appears to be necessary for metastasis to occur. Other useful indicators of clinical behavior were the degree of mitotic activity, nuclear pleomorphism, and infiltrating margins. Based on these data and a literature review, the authors suggest close follow-up of patients whose primary tumors contain areas of stromal overgrowth because, in all series combined, the risk of metastatic spread in such patients was 72% within 5 years. Among these high-risk patients, local recurrence is another indication that metastasis is likely.

An extremely rare neoplasm of the prostate, commonly known as phyllodes type of atypical prostatic hyperplasia, is described in a 57-year-old man. Over a 2-year period, the atypical smooth muscle component has transformed into leiomyosarcoma, and eventually metastasized to the bones and the lungs. Such an evolution has never been documented for this tumor.

Biochemical study of fibroepithelial tumors of the female breast showed presence of progesterone receptor (PgR) in all five cystosarcoma phyllodes (two malignant, three benign), and in 11 of 13 fibroadenomas tested. Estrogen receptor (ER) was detected in only one of five cystosarcomas and 2 of 13 fibroadenomas. The relative volumes occupied by epithelium and stroma in each tumor were measured from histologic sections. The results were consistent with presence of PgR in the stroma and ER in the epithelium. Different types of cystosarcoma (benign and malignant) and different types of fibroadenomas (intracanalicular, pericanalicular, and mixed) did not differ significantly in content of PgR, and mean levels of PgR in cystosarcoma were comparable with those in fibroadenomas. The presence of PgR in cystosarcomas suggests that progestational therapy, and possibly other forms of hormonal therapy, should be tested in the treatment of advanced, malignant cystosarcoma phyllodes.

A retrospective clinicopathologic evaluation of 42 patients with cystosarcoma phyllodes was undertaken to determine if tumor size, contour, degree of stromal atypia and mitotic activity were reliable indicators of clinical behavior. Excluding size, the latter three determinants showed a positive correlation with prognosis and served as the basis of a classification in which 18 benign, 5 borderline and 19 malignant cystosarcomas were diagnosed. The tumors occurred in women averaging 44.3 years of age who most often presented with a palpable occasionally painful mass with a median diameter of 5 cm. Excision or simple mastectomy were the more frequent forms of therapy. Local recurrences were experienced by 6 patients and occurred in all 3 categories of tumor. Only malignant neoplasms developed systemic metastases which was observed in 4 patients 3 of which have died. One borderline tumor recurred 14 times and eventually proved fatal as a result of contiguous pulmonary involvement. A plea is made to label the stroma of malignant cystosarcomas as to the cell(s) of origin so future investigators may evaluate the effect of various soft tissue patterns on prognosis.

A retrospective study was made of recurrent mammary cystosarcoma phyllodes seen at Memorial Sloan-Kettering Cancer Center from 1932 to 1972. Clinicopathologic findings, clinical course, treatment, and survival are described. It is apparent that cystosarcomas may recur and several years may elapse between reucrrences. We found that 18% of the benign and 8% of the malignant cystosarcomas recurred and malignant "transformation" occurred in two of the 28 recurrent benign cystosarcomas.

Twenty-six cases of breast carcinoma demonstrating pseudosarcomatous metaplasia are described, and their clinical and pathologic aspects are correlated. The metaplastic elements include mature and immature bone, cartilage, myxoid stroma, loose fibromyxoid and dense spindle-cell stroma, and anaplastic stroma with giant cell features. Light and electron microscopic examination demonstrated an orderly sequence of transformation and dedifferentiation of epithelial cells to become undifferentiated mesenchymal cells. These tumors are more aggressive than purely epithelial carcinomas. The overall survival rate was 44%. Five-year survival figures for TNM Stages I, II, and III lesions were 56%, 26%, and 18%, respectively. The incidence of lymph node metastasis was 25% despite the large size of many of these tumors. Systemic metastases replicated the range of metaplastic elements seen in the primary site. Patients with tumors composed predominantly of pseudosarcomatous elements had worse prognoses than those with predominantly epithelial components (28% versus 62%, 5-year survival).

Cystosarcoma phyllodes tumors of the breast occasionally exhibit malignant behavior, including chest wall invasion, hematogenous spread or, rarely, metastasis by lymphatic routes. An unusual case of cystosarcoma is presented in which death was apparently caused by perineural and intraneural extension of the tumor along intercostal nerves, to the sympathetic chain, and then to the brain stem by spinal roots, with no evidence of embolic metastasis. Cancer 56: 1691-1695, 1985.

The clinical and pathologic findings in 94 patients with cystosarcoma phyllodes were studied to determine which pathologic characteristics were related to clinical behavior. The neoplasm recurred in 28 patients and 15 patients (17%) died of metastatic cystosarcoma. Most recurrences occurred within 2 years of initial surgery and all patients who died of cystosarcoma did so within 6 years. No tumor less than 4 cm in diameter or having fewer than 3 mitotic figures per 10 hpf in the areas of greatest mitotic activity proved fatal. Other microscopic features associated with a low risk of recurrence or of death were pushing margins (one death in 39 examples) and minimal cytologic atypism of the stromal cells (two deaths in 30 examples). No one feature was wholly reliable and a clearcut separation of benign from malignant tumors could not be made. Although axillary lymph nodes were enlarged in 17% of patients, metastasis to them occurred in not more than three instances (one histologically proved) and therefore radical mastectomy or routine axillary lymph node dissection is not worthwhile. Wide local excision for small cystosarcomas and simple mastectomy for larger ones is recommended. In addition, low axillary dissection should be considered in patients having clinically enlarged axillary lymph nodes and tumors larger than 4 cm if there is marked atypism of stromal cells or high mitotic activity, as determined by microscopic evaluation of the cystosarcoma.

Modified radical mastectomy was performed in 162 patients with minimal breast cancer between 1953 and 1972 (JAU). The majority of lesions were small infiltrating cancers (52), non-infiltrating duct cancer (37), lobular carcinoma in situ (49), Paget's disease with non-infiltrating cancer (12), and also a small group of miscellaneous low-grade cancers (7), and malignant cystosarcoma phyllodes (5). Major indications for surgery were the presence of a mass, nipple ulceration, positive mammograms, and the discovery of an occult cancer on contralateral biopsy. Mammography was accurate in 60% of patients with minimal breast cancer. Contralateral biopsy proved to be a very sensitive diagnostic technique, and was positive in 35 patients, most of whom were negative by clinical examination and mammography. Residual cancer was found in the mastectomy specimens in 52% after generous excisional biopsies. Sixteen percent of patients with minimal infiltrating cancer and clinically negative axillae had micrometastases (2mm or less) in axillary nodes. The crude survival rate free of disease was 93 of 95 patients (98%) at 5 years, and 42 of 44 patients (95%) at 10 years. There were no local recurrences. No postoperative radiation was used on these patients. Modified radical mastectomy is recommended as the ideal treatment for minimal breast cancer as it yields excellent long term salvage, complete local control, and is more acceptable cosmetically than the classical radical mastectomy. It is to be preferred over extended lumpectomy (tylectomy) with radiation therapy since it avoids the potential long-term carcinogenic effects of ionizing radiation. The excellent results obtained by adequate treatment of minimal breast cancers attest to the need for reliable mass screening techniques to detect occult, asymptomatic breast cancer, and for the availability of these techniques to the public at large.

Two cases of cystosarcoma phyllodes of the breast are presented with central nervous system (CNS) metastases appearing several years after mastectomy for the primary lesion. Unusual features in these cases include the widespread metastases themselves, neurologic symptomatology and CNS involvement, metastases into a uterine leiomyoma and into an area of hepatic adenomatous hyperplasia, and glomus-like structures in one of the primaries with similar structures resembling glomus cells in metastases. The distant metastases were of stromal cells only and frequently surrounded epithelial cells to isolate the indigenous glandular structures. The concept that cytosarcoma is a peculiar stromal neoplasm rather than a tumor of dual neoplastic origin is discussed.

Cystosarcoma phyllodes is an unusual breast neoplasm that rarely metastasizes. Most series report chemotherapy, radiation, and hormonal therapy to be ineffective. Three patients were treated with cisplatin and etoposide combination chemotherapy with effective palliation in two patients. Radiation therapy was effective in controlling symptomatic metastasis in all three patients. Hormonal therapy was ineffective in two patients despite the presence of positive hormone receptors. Chemotherapy and radiotherapy may be more effective in the treatment of this tumor than has been reported, although there is no apparent role for hormonal therapy. Functional hormone receptors are probably not present.

Biochemical study of fibroepithelial tumors of the female breast showed presence of progesterone receptor (PgR) in all five cystosarcoma phyllodes (two malignant, three benign), and in 11 of 13 fibroadenomas tested. Estrogen receptor (ER) was detected in only one of five cystosarcomas and 2 of 13 fibroadenomas. The relative volumes occupied by epithelium and stroma in each tumor were measured from histologic sections. The results were consistent with presence of PgR in the stroma and ER in the epithelium. Different types of cystosarcoma (benign and malignant) and different types of fibroadenomas (intracanalicular, pericanalicular, and mixed) did not differ significantly in content of PgR, and mean levels of PgR in cystosarcoma were comparable with those in fibroadenomas. The presence of PgR in cystosarcomas suggests that progestational therapy, and possibly other forms of hormonal therapy, should be tested in the treatment of advanced, malignant cystosarcoma phyllodes.

Eighty-one female patients with phyllodes tumors of the breast, surgically treated from 1974 to 1983, were studied. Their age ranged from 9 to 88 years. According to histology, the series was divided into three groups, of 28 (34.5%) benign tumors, 32 (39.5%) border-line tumors, and 21 (25.9%) malignant tumors. Because ten patients were lost to follow-up, only 71 women could be evaluated. All the patients had received surgical treatment: 51 women had been treated conservatively (11 enucleations, 40 wide resections), and 20 had undergone radical operations (13 underwent total and five underwent subcutaneous mastectomies, whereas one underwent modified and one underwent radical mastectomy). The mean follow-up, for the three groups, was 106 months for benign, 84 months for borderline, and 82 months for malignant tumors; in no case was radical surgery followed by local recurrence: of 51 women conservatively treated, 14 experienced local relapse, i.e., one of 24 women with benign, ten of 22 with borderline, and three of 8 with malignant lesions. Only two of 47 patients (4.2%) with borderline or malignant tumors developed distant metastasis and died from disease. No relationship between tumor size and risk of local recurrence could be demonstrated, and no difference could be identified between borderline and malignant lesions, in terms both of local and distant relapse. Local recurrences do not appear to affect survival: as a consequence, wide resection should be the primary treatment. Enucleation is to be proscribed. Total mastectomy has been indicated for very large tumors and for local recurrences of borderline and malignant lesions. Axillary dissection is not worthwhile.

We studied the DNA histograms obtained by flow cytometry from a series of six giant fibroadenomas and ten phyllodes tumors to determine if the analysis of DNA ploidy would help to predict clinical behavior. We were unable to document any relation between ploidy and histologic appearance, recurrence, metastasis, lesion size, or patient age. DNA aneuploid stem cell lines were seen in 75% of histologically benign phyllodes tumors, 50% of histologically malignant phyllodes tumors, and approximately 33% of giant fibroadenomas.

The unique combination of clinically unsuspected stromal sarcoma and non-infiltrating intraductal mammary carcinoma is presented. This association has not been previously documented. The clinicopathologic features of cystosarcoma phyllodes and mammary stromal sarcoma are reviewed to provide background perspectives to illuminate the unusual aspects of this case.

A retrospective study was made of recurrent mammary cystosarcoma phyllodes seen at Memorial Sloan-Kettering Cancer Center from 1932 to 1972. Clinicopathologic findings, clinical course, treatment, and survival are described. It is apparent that cystosarcomas may recur and several years may elapse between reucrrences. We found that 18% of the benign and 8% of the malignant cystosarcomas recurred and malignant "transformation" occurred in two of the 28 recurrent benign cystosarcomas.

Twenty-six cases of breast carcinoma demonstrating pseudosarcomatous metaplasia are described, and their clinical and pathologic aspects are correlated. The metaplastic elements include mature and immature bone, cartilage, myxoid stroma, loose fibromyxoid and dense spindle-cell stroma, and anaplastic stroma with giant cell features. Light and electron microscopic examination demonstrated an orderly sequence of transformation and dedifferentiation of epithelial cells to become undifferentiated mesenchymal cells. These tumors are more aggressive than purely epithelial carcinomas. The overall survival rate was 44%. Five-year survival figures for TNM Stages I, II, and III lesions were 56%, 26%, and 18%, respectively. The incidence of lymph node metastasis was 25% despite the large size of many of these tumors. Systemic metastases replicated the range of metaplastic elements seen in the primary site. Patients with tumors composed predominantly of pseudosarcomatous elements had worse prognoses than those with predominantly epithelial components (28% versus 62%, 5-year survival).

Twenty-six cases of cystosarcoma phyllodes diagnosed at M. D. Anderson Hospital were reviewed. The following criteria were evaluated for possible correlation with local recurrence, uncontrolled local recurrence, metastasis, and tumor death: tumor size, stromal overgrowth, tumor necrosis, mitotic rate, stromal cellularity, nuclear size and pleomorphism, the presence of specialized stroma, and initial therapy. Of the 26 tumors, seven caused death. Five patients developed metastatic spread, and all of them died of tumor. Five patients had local recurrence, which was uncontrolled in three (two patients died with uncontrolled recurrence alone, and one with uncontrolled recurrence and metastasis). Stromal overgrowth was present in eight cases. Six of the seven patients who died of tumor had stromal overgrowth, including all five with metastasis. Correlation of stromal overgrowth with metastatic spread and tumor death was significant at P levels of 0.0014 and 0.02, respectively. It is concluded that stromal overgrowth is a significant histologic indicator of malignant behavior in cystosarcoma phyllodes.

Two cases of metastatic cystosarcoma phyllodes, one in lung and the other in axillary lymph nodes, were studied with the light and the electron microscopes. The malignant element of these tumors appears to be a poorly differentiated mesenchymal cell. No evidence supporting the presence of epithelial cells was found. The tumor metastatic to the axilla exhibited intracellular virus-like particles similar to those described in adenocarcinoma of breast and other sarcomas.

We detected cytosol estrogen receptors in a malignant cystosarcoma phyllodes and succeeded in serial transplantations into athymic nude mice. The tumors of this transplantable strain (MC-3-JCK) have the same histologic features as the original tumors, and retain significant amounts of cytosol estrogen receptors. This strain appears to provide a useful experimental model for the study of biologic and therapeutic aspects of human cystosarcoma phyllodes of the breast.

Malignant alteration in cystosarcoma phyllodes is uncommon and almost always confined to the stromal component. A rare case of recurring cystosarcoma is reported. In the first recurrence, lobular carcinoma in situ was present within the tumor and the second recurrence revealed tubular carcinoma within the cystosarcoma. Epithelial proliferation of various sorts is not uncommon in cystosarcomas but only two previous instances of infiltrating carcinoma have been observed within these tumors. The problem of mammary carcinoma coexisting with cystosarcoma is reviewed and comparisons are made with a closely related neoplasm, the fibroadenoma.

A cell line was established from a portion of a 25-cm stromal sarcoma of the left breast of a 65-year-old woman. The clinical course was rapid with tumor recurrence on the chest wall less than 1 month after mastectomy. Other cutaneous and abdominal metastases occurred shortly thereafter, and death followed within 3 months despite chemotherapy. The cultured cells, designated RW-972, produced large amounts of acid mucopolysaccharides (hyaluronic acid) and mimicked the aggressive growth characteristics seen in the patient. After injection into nude mice, the tumor grew rapidly and occasionally produced metastases. This unique cell line, RW-972, presumably derived from the stromal component of a human malignant cystosarcoma phyllodes, might be useful in studies of experimental therapy of this rare tumor type and of lobular stromal cells of breast. It may also be used to investigate hyaluronic acid production by tumor cells.