Perinucleolar compartment prevalence has an independent prognostic value for breast cancer.

The perinucleolar compartment (PNC) is a multicomponent nuclear structure enriched with RNAs transcribed by RNA pol III and RNA binding proteins. Studies in cultured cells showed an association between PNC and transformed phenotype. To evaluate the relationship between structure and malignancy in vivo, we examined PNC prevalence (the percentage of cells containing at least one PNC) in normal and cancerous paraffin-embedded breast tissues using immunohistochemistry against a PNC-associated protein. Five hundred nuclei in the most active area of each sample were scored for PNC prevalence. The results show that PNC prevalence significantly correlates with the progression of breast cancer (by the criteria of staging). PNC prevalence in primary tumors, lymph nodes, and distant metastases shows a stepwise increase from a median of 23% in primary tumors to approximately 100% in distant metastases. In addition, univariate and multivariate (controlling for tumor size and grade) analyses show that early-stage patients with invasive ductal carcinomas containing a higher PNC prevalence have a significantly poorer prognosis. These findings link PNC prevalence with the progression of breast cancer in vivo and suggest that PNC-containing cells have metastatic advantages. These findings also show the potential of PNC prevalence as a prognostic marker for breast cancer.

Expression of cell adhesion molecules, CD44s and E-cadherin, and microvessel density in invasive micropapillary carcinoma of the breast.

AIMS: Invasive micropapillary carcinoma (IMPC) is a rare variant of ductal carcinoma of the breast and is characterized by high metastatic potential and an aggressive clinical course. This tumour is hence ideal for studying the mechanism underlying tumour biological behaviour, especially metastasis. Cell adhesion molecules, such as CD44 and E-cadherin (Ecad), and angiogenesis are considered important in the invasion and metastasis of tumours. METHODS AND RESULTS: We immunohistochemically analysed 23 IMPCs for expression of a standard form of CD44 (CD44s), Ecad, and CD34 to measure microvessel density (MVD). Results are compared with the changes observed in 23 tubular carcinomas (TCs), another variant of ductal carcinoma that rarely metastasizes. Evaluation of haematoxylin and eosin (H&E) sections showed a higher prevalence of lymph-vascular invasion (19/23, 83%) and regional lymph node involvement (12/15, 80%) in IMPCs; whereas no lymph-vascular invasion or lymph node metastasis was identified in TCs. Loss or reduction of CD44s immunoreactivity was significantly frequent in IMPC (39%) compared with TC (4%) (P = 0.0098), and was associated with positive axillary lymph nodes and lymph-vascular invasion. All cases of IMPC and TC strongly expressed Ecad. MVD (in five 200x fields) was significantly higher in IMPC (88 +/- 37) than in TC (57 +/- 16) (P = 0.001). In the IMPC group, MDV was higher in cases with positive lymph node(s) (P = 0.048), and cases with loss or reduction of CD44s expression (P = 0.011). The same trend was also demonstrated in cases with lymph-vascular invasion (P = 0.077). Moreover, the vessels in IMPC had much smaller calibres with thinner walls than those in TC. CONCLUSIONS: Loss of the CD44 adhesion molecule and high MVD may play a significant role in the high incidence of lymph-vascular permeation and metastasis in IMPC.

Expression of epithelial mucins Muc1, Muc2, and Muc3 in ductal carcinoma in situ of the breast.

Epithelial mucins are glycoproteins secreted by epithelial cells and their carcinomas. At least nine mucin genes have been identified, and their products (MUC1-MUC9) are expressed in various epithelia. MUC1 is a mucin expressed in breast epithelial cells, whereas MUC2 and MUC3 are primarily intestinal mucins. Although MUC1 and MUC2 expression has been documented in invasive ductal carcinoma of the breast, mucin expression in pure ductal carcinoma in situ (DCIS) has not been investigated. Sixty-one of 105 cases of DCIS without coexisting infiltrating carcinoma diagnosed during a 30-month period were selected as having sufficient tissue for study. Paraffin-embedded tissue sections were stained using immunohistochemical techniques with mouse monoclonal anti-MUC1, anti-MUC2, and rabbit-specific polyclonal anti-MUC3 antibodies. Immunoreactive epitopes of MUC1, MUC2, and MUC3 were expressed in DCIS in 61, 19, and 16 of 61 cases, respectively. MUC2 and MUC3 staining intensity in DCIS was markedly less than that observed for MUC1. Luminal and/or cytoplasmic patterns of staining were observed for MUC1. MUC2 and MUC3 showed only cytoplasmic staining. Cytoplasmic-only staining of MUC1 was associated with a higher grade of DCIS. Any MUC2 staining was also associated with a higher grade of DCIS. Coexpression of MUC2 and MUC3 was present in only 6 of 61 cases, and MUC3 staining was unrelated to the grade of DCIS. Cytoplasmic expression of MUC1 and MUC2 appears to be associated with a higher grade of DCIS. MUC3 expression appears to be independent of grade and expression of MUC1 and MUC2. The relationship of mucin expression and grade warrants further study.

Are malignant cells displaced by large-gauge needle core biopsy of the breast?

OBJECTIVE: The purpose of this paper is to determine the rate of tumor displacement resulting from large-gauge needle core biopsy in patients with breast carcinoma. MATERIALS AND METHODS: Three hundred fifty-two cancer excisions in patients who had undergone large-gauge needle core biopsy were evaluated for evidence of tumor displacement. Three needle procedures were compared: vacuum-assisted, automated gun, and core biopsy guided by palpation. Needle track visualization, presence and amount of tumor displacement, tumor morphology, and interval between core biopsy and surgical excision were recorded for each case. RESULTS: Seventy-six cases showed tumor displacement of one or two cell clusters, and 38 cases-showed displacement of multiple tumor fragments. Tumor displacement was identified in 37% of automated gun specimens, 38% of specimens obtained with palpable guidance, and 23% of specimens obtained with a vacuum-assisted needle. Tumor displacement was seen in 42% of patients with an interval between biopsy and excision of less than 15 days, in 31% of patients with an interval of 15-28 days, and in 15% of tumors excised more than 28 days after core biopsy (p < .005). CONCLUSION: Tumor cell displacement was observed in 32% of patients who had undergone large-gauge needle core biopsy. The incidence and amount of tumor displacement was inversely related to the interval between core biopsy and excision. This relation suggests that tumor cells do not survive displacement.

Diagnostic discrepancies in breast specimens subjected to gross reexamination.

To determine the accuracy of gross examination of breast specimens from a large university pathology service, 1120 breast specimens submitted from 1995 to 1997 that had residual tissue after submission of tissue sections were reexamined for diagnostic discrepancies. A total of 520 mastectomies, 143 wire localization excisions, 156 lumpectomies, and 301 mammoplasties were reexamined. Fifty-three (5%) major and 65 (6%) minor diagnostic discrepancies were detected. Major discrepancies included eight additional positive lymph nodes, 37 missed cancers, four upstagings by size, and four skin invasions. Forty-four of the major discrepancies were in mastectomy specimens. First-year residents accounted for slightly more than one half of all discrepancies. In contrast, review of original slides of 733 breast cancer cases revealed only 11 (1.5%) major discrepancies: three changes of margin status, six missed carcinomas, one positive lymph node, and one upstaging by size. Most discrepancies occurred because a specimen was not thoroughly inspected. The second most common cause was failure to recognize lesions. Our findings suggest that gross dissection performed by first-year residents is more prone to error and that such discrepancies are amenable to instruction and supervision.

Imaging features of focal breast fibrosis: mammographic-pathologic correlation of noncalcified breast lesions.

OBJECTIVE: Our objective was to describe the spectrum of imaging and histologic findings of focal breast fibrosis with an emphasis on noncalcified lesions, thereby offering a means of confirming mammographic-pathologic concordance on core biopsy of this increasingly encountered diagnosis. MATERIALS AND METHODS: Retrospective review of 610 core needle biopsies revealed the histologic diagnosis of focal fibrosis in 89 (15%). Thirty-nine cases were excluded: 17 in which focal fibrosis was not the primary diagnosis and 22 in which calcifications were the main imaging findings. The 50 remaining patients with noncalcified lesions that proved on histology to be focal fibrosis constituted the basis of the study. RESULTS: Mammographically, focal fibrosis presented as a mass in 68% of patients (n = 34), architectural distortion in 12% (n = 6), and asymmetric density in 10% (n = 5); focal fibrosis was mammographically occult in 10% (n = 5). Sonographically, 72% (n = 36) of cases of focal fibrosis presented as masses with three echo texture patterns: hypoechoic, isoechoic, and centrally echogenic with a peripheral hypoechoic rim. The sonographic margins were well circumscribed (n = 21), lobulated (n = 10), or ill defined (n = 5). Histologic review revealed three morphologic patterns of collagen deposition: perilobular, septal, and haphazard fibrosis. Correlation with the imaging findings showed that septal and perilobular fibrosis most often presented as hypoechoic or centrally echogenic masses, whereas the haphazard form was more often seen with architectural distortion. CONCLUSION: Focal fibrosis often presents as a noncalcified mass on mammography or sonography. The diagnosis of focal fibrosis on core needle biopsy can be considered concordant for a mass exhibiting well-circumscribed or partially obscured margins. Imaging findings discordant with focal fibrosis, such as marginal spiculation, require excisional biopsy.

Immunohistochemical detection of breast cancer cells in paired peripheral blood progenitor cell specimens collected after cytokine or cytokine and myelosuppressive chemotherapy.

Mobilized peripheral blood progenitor cells (PBPC) from 30 patients with advanced breast cancer were studied for the presence of tumor cell contamination using a highly sensitive immunohistochemical technique with the capacity to detect one tumor cell in one million mononuclear cells. Aliquots of PBPC were obtained after 4 days of G-CSF and/or GM-CSF and again during G-CSF-stimulated recovery from myelosuppressive doses of cyclophosphamide. The overall incidence of tumor cell contamination was 23%, occurring in PBPC specimens from seven of 30 patients. All four cases in which tumor cells were detected after mobilization with cytokine alone also had tumor cells detected in PBPCs collected following chemotherapy and G-CSF. There were three cases in which malignant contamination was detected only in the specimens collected after cyclophosphamide. There was a greater frequency of tumor cell contamination in aphereses performed during G-CSF-stimulated recovery from cyclophosphamide than in collections primed by cytokine alone (13% vs 23%; P = 0.08), although this did not reach statistical significance. This trend suggests that collection of PBPC during cytokine-stimulated recovery from myelosuppressive chemotherapy may be associated with a greater risk of contamination with malignant cells than apheresis during mobilization with cytokines in the steady state.

Sonographic evaluation of infiltrating lobular carcinoma.

OBJECTIVE: Infiltrating lobular carcinoma (ILC), which accounts for 7-10% of all breast malignancies, often poses diagnostic difficulties. The purpose of our study was to correlate the clinical, mammographic, and sonographic findings in each histologic subtype of ILC and to evaluate the sensitivity of sonography in its diagnosis. MATERIALS AND METHODS: We reviewed 208 cases of invasive lobular carcinoma. In 81 of these tumors, sonography was performed to further examine a mammographically invisible palpable abnormality or a mammographically subtle lesion. A dedicated breast pathologist classified each of these tumors as pure invasive lobular carcinoma or mixed invasive lobular and ductal carcinoma. Pure ILC tumors were further subclassified as one of five histologic subtypes. We retrospectively studied the clinical, mammographic, and sonographic findings in each histologic tumor subtype. RESULTS: The most common sonographic appearance of ILC was a heterogeneous, hypoechoic mass with angular or ill-defined margins and posterior acoustic shadowing, which was seen in 60.5% (49/81) of tumors. Of the remaining 32 tumors, 15% (12/81) showed focal shadowing without a discrete mass, 12% (10/81) appeared as a lobulated, well-circumscribed mass, and 12% (10/81) were sonographically invisible. Although considerable overlap occurred among histologic subtypes, classic ILC tended to present as focal shadowing without a discrete mass; pleomorphic ILC typically was seen as a shadowing mass; and, of all the tumor subtypes, signet, alveolar, and solid ILC were most likely to be revealed on sonography as a lobulated, well-circumscribed mass. In the 81 mammographically subtle or invisible lesions, sonography detected the tumor in 87.7% (71/81). The sensitivity of sonography in tumors smaller than 1 cm was 85.7% (12/14). CONCLUSION: High-resolution sonography of the breast is a useful adjunct in the evaluation of ILC, a neoplasm that frequently presents a clinical and mammographic diagnostic challenge.

Risk of concurrent prostate cancer in cystoprostatectomy specimens is related to volume of high-grade prostatic intraepithelial neoplasia.

OBJECTIVES: To assess the relationship of prostatic intraepithelial neoplasia (PIN) with both incidental and clinical carcinoma of the prostate. METHODS: We retrospectively reviewed prostate histology in 48 men (group 1) who underwent surgical removal of the prostate for diagnoses other than prostate cancer, as well as in 64 men (group 2) who underwent radical prostatectomies. Both groups were assessed for the presence and extent of high-grade (HG-) PIN and compared with respect to patient age, Gleason score, and volume of prostate cancer. RESULTS: HG-PIN was present in 40 of 48 (83%) group 1 cases. Forty-six percent of these cases (22 of 48) had incidental prostate cancer. Twenty-nine of 48 (60%) group 1 patients with HG-PIN had multifocal or extensive disease. Twenty of 22 (91%) incidental prostate cancers were present in 29 prostates with multifocal or extensive HG-PIN. In contrast, only 2 of 19 (11%) cases with absent to focal HG-PIN had prostate cancer. The association of multifocal or extensive HG-PIN with incidental prostate cancer was significant (P = 0.001); the relationships of extent of HG-PIN and cancer volume (P = 0.06) or high Gleason score (P = 0.017) were not significant. HG-PIN was present in 61 of 64 (95%) group 2 cases. The associations of extent of HG-PIN and cancer volume (P = 0.169) or high Gleason score (P = 0.156) were not significant. CONCLUSIONS: Both the low rate of incidental prostate cancer in specimens with absent to focal HG-PIN and the high rate of cancer in specimens with multifocal or extensive HG-PIN suggest that HG-PIN is a marker for concurrent prostate cancer and that the risk of concurrent prostate cancer is related to the volume of HG-PIN in the prostate gland.

Cell type specific expression of steroid 5 alpha-reductase 2.

Isozymes of steroid 5 alpha-reductase (5 alpha-reductase) have crucial roles in androgen physiology by synthesizing the potent hormone dihydrotestosterone. The expression pattern of the 5 alpha-reductase type 2 isozyme was determined in genital and extragenital tissues by developing an immunohistochemical assay using formalin-fixed tissue and affinity purified polyclonal antibodies that specifically recognize this isozyme. Expression was detected in basal epithelial and stromal cells of the normal prostate but not in luminal epithelial cells. Stromal cells of the seminal vesicle also expressed the type 2 isozyme. In contrast, staining was detected in epithelial cells of the epididymis but not in the surrounding stroma. Myofibroblasts in foreskin samples of normal and hypospadiac individuals expressed antigen and were distributed in bands throughout the prepuce, suggesting a clonal origin. In most cells the type 2 isozyme exhibited a perinuclear subcellular distribution. However, in liver hepatocytes the protein was distributed throughout the intracellular membrane compartment.

Expression and regulation of steroid 5 alpha-reductase 2 in prostate disease.

The androgen dihydrotestosterone is synthesized by the enzyme steroid 5 alpha-reductase, and it is required for growth and development of the prostate. We used immunohistochemistry to examine the expression of the type 2 isozyme of 5 alpha-reductase in benign prostatic hyperplasia and prostate cancer. The type 2 isozyme is highly expressed within stromal cells in both disease states. No type 2 isozyme is detectable in a lymph node metastasis. Immunoblotting studies show that androgen ablation therapies substantially decrease isozyme expression in the epididymis but have a lesser effect on expression in the prostate. Finasteride therapy (2 weeks to 3 years) did not abolish expression of the prostatic type 2 isozyme nor did this drug treatment induce expression of the type 1 isozyme.

Primary leiomyosarcoma of the seminal vesicle.

A case of leiomyosarcoma of the seminal vesicle is described in a 68-year-old man. Digital rectal examination and pelvic computed tomography (CT) scan disclosed a large pelvic mass in the region of the prostate, whereas magnetic resonance imaging (MRI) suggested that the mass arose from the right seminal vesicle. Biopsy of the mass revealed a high-grade malignancy, thus a radical cystoprostatectomy was performed. Pathologic examination revealed a leiomyosarcoma arising from the right seminal vesicle. The patient is well and free of recurrent disease 13 months following surgery.

Detection of Mycobacterium avium-intracellulare complex in bone marrow specimens of patients with acquired immunodeficiency syndrome.

Thirty-seven bone marrow core biopsy specimens from 21 human immunodeficiency virus-infected patients with Mycobacterium avium-intracellulare complex bacteremia were stained using rabbit polyclonal antibodies against Mycobacterium bovis strain Bacillus-Calmette-Guerin (BCG) and Mycobacterium duvalii, as well as Kenyon and Fite stains, to compare sensitivities of these techniques and evaluate possible response to therapy. The patients in this study had participated in a phase I/II trial of liposome-encapsulated gentamicin therapy. Two biopsy specimens had inadequate tissue for evaluation. Thirty-two specimens demonstrated bacilli with anti-M duvalii, 33 with anti-BCG, 20 with Kenyon, and 23 with Fite. Two were negative with all stains. Fifteen biopsy specimens had epithelioid granulomas, 12 had histiocytic granulomas, and 1 had a granuloma of indeterminate type. The remaining seven biopsy specimens had no granulomas. Four of these seven demonstrated bacilli with anti-M duvalii, 5 with anti-BCG, 1 with Kenyon, and 2 with Fite. The number of M avium-intracellulare organisms per milliliter of blood decreased in 14 of 21 patients after liposome-encapsulated gentamicin therapy. However, none of the 11 patients whose pre- and post-therapy bone marrow core biopsy specimens were both evaluable demonstrated a reduction in the number of M avium-intracellulare organisms. The authors concluded that anti-M duvalii and anti-BCG are more sensitive than acid-fast stains for identifying M avium-intracellulare infection in bone marrow core biopsy specimens of patients who have acquired immunodeficiency syndrome (AIDS) with M avium-intracellulare bacteremia. Bone marrow core biopsy specimens may provide a perspective on M avium-intracellulare infection in AIDS patients that differs from the one provided by blood cultures.

Pagetoid spread of intratubular germ cell neoplasia into rete testis: a morphologic and histochemical study of 100 orchiectomy specimens with invasive germ cell tumors.

Intratubular germ cell neoplasia (ITGCN) is now considered to be the preinvasive phase of testicular germ cell tumors with the exceptions of spermatocytic seminoma, pure yolk sac tumor, and mature teratoma. Pagetoid spread of ITGGN into rete testis is a common yet unpublished finding in these cases. We reviewed 100 cases of testicular germ cell tumors from the Surgical Pathology service of Parkland Memorial Hospital (Dallas, TX) to evaluate the frequency of this pattern of spread. Additional sections were obtained from selected cases and were stained with anti-placental alkaline phosphatase, anti-low molecular weight keratin (clone AE1), and various lectins to highlight the process. Pagetoid spread of ITGCN into rete testis was identified in 24 of 60 cases (40%) in which histologic sections contained both ITGCN and rete testis. The incidence of pagetoid ITGCN involvement of the rete testis was lower in pure seminoma (seven of 25 cases [28%]) than in testes containing nonseminomatous germ cell tumors (17 of 35 cases [49%]). AE1 stained the epithelial cells of the rete testis but not the cells of the ITGCN, whereas placental alkaline phosphatase stained the neoplastic cells but not the epithelial cells of the rete testis. These stains were useful in delineating two cases in which the pagetoid involvement was so extensive that they were misdiagnosed as invasive seminomas. Pagetoid spread of ITGCN is a relatively common finding in testicular germ cell tumors and rarely can be mistaken for invasive seminoma. Immunohistochemistry can be helpful in distinguishing florid pagetoid spread from invasive seminoma.

Opportunistic events and p17 expression in the bone marrow of human immunodeficiency virus-infected patients.

Bone marrow biopsies from 114 human immunodeficiency virus (HIV)-infected patients were stained with an anti-p17 monoclonal antibody to detect active HIV replication and associated factors. Immunoreactive p17 was found as virus-like particles in macrophages and dendritic cells and occasionally in megakaryocytes in 62 of 114 marrows and was considered evidence of active HIV replication. Immunoreactive p17 was not found significantly more often in the marrows of patients with lower CD4 cell counts; however, it was found significantly more in the marrows of patients with concurrent mycobacterial or fungal infections or lymphoma (chi 2 = 12.1, P < .001). Immunoreactive p17 was even more frequent when these opportunistic diseases were found in the biopsied marrow (chi 2 = 20.5, P < .001). The association of active HIV replication with certain opportunistic diseases, but not with lower CD4 cell counts, raises the possibility that these opportunistic diseases may under some circumstances be a cause as well as a consequence of active HIV replication.

The tubular variant of adenoid cystic carcinoma of the Bartholin's gland.

The authors report two examples of the tubular variant of adenoid cystic carcinoma of the Bartholin's gland. Each occurred in young women (both aged 34 years) and presented with a painful indurated mass, clinically thought to be of inflammatory nature. On microscopic examination, the most distinctive feature was the haphazard proliferation of bland-appearing tubular structures often lined by two cell layers. A thin, periodic acid-Schiff-positive basement membrane immunoreactive for collagen IV surrounded the tubules, but hyaline globules were virtually absent. The inner cell layer was strongly and diffusely cytokeratin positive, whereas epithelial membrane antigen reactivity was seen only along the luminal borders. Focal positivity for S100 protein identified the peripheral myoepithelial cells, which were confirmed by electron microscopy. Focal perineural invasion was seen. The histologic, ultrastructural, and immunohistochemical features of these tumors are similar to those of the tubular variant of adenoid cystic carcinoma arising in major and minor salivary glands. In one of the cases, which was followed for 6 years, the tumor metastasized, despite the low histologic grade, and the patient died. Similarities and differences between the tubular, cribriform, and solid variants and other vulvar tumors are discussed.

Sweat gland adenomas: immunohistochemical study with emphasis on myoepithelial differentiation.

Thirty-one dermal appendage tumors of sweat gland differentiation including 7 spiradenomas (SPA), 8 cylindromas (CYL), 8 acrospiromas (ACS), and 8 chondroid syringomas (CS) were analyzed using antibodies to epithelial membrane antigen (EMA), cytokeratin (AE1, AE3, CAM 5.2, 34BE12), S-100 protein, actin (ACT), and desmin (DES) to characterize the immunocytochemical profile of benign sweat gland tumors. Cytokeratin expression was variable; AE1, 34BE12, AE3, and CAM 5.2 were present in 31, 24, 23, and 22 tumors respectively; 29 tumors contained EMA. Seventeen tumors, (6 SPA, 8 CYL, 2 ACS, 1 CS) stained with antibody to alpha smooth muscle actin, and 26 (7 SPA, 7 CYL, 4 ACS, 8 CS) expressed S-100 protein. Although some prior studies had reported actin filaments on electron microscopy in both spiradenoma and cylindroma, these tumors have previously been considered to be negative for myoepithelial differentiation. All spiradenomas and cylindromas we studied demonstrated actin and/or S-100 protein positivity in basal epithelial cells, consistent with myoepithelial differentiation. The organization of actin and S-100 protein positivity displayed by the spiradenomas and cylindromas we studied suggests that the tumors are differentiated towards the secretory portion of the eccrine sweat gland.

Monoclonal antibody to Pneumocystis carinii. Comparison with silver stain in bronchial lavage specimens.

Monoclonal 3F6 anti-Pneumocystis carinii antibody (MAB-3F6) was used to stain cell blocks from 164 bronchial lavage specimens from patients with the acquired immune deficiency syndrome (AIDS) and AIDS-related complex and compared with slides stained with Grocott's modification of the Gomori methenamine silver stain. Pneumocystis organisms were present in 83 of 164 cases using MAB-3F6 stain, whereas Grocott's modified silver stain demonstrated Pneumocystis organisms in 48. MAB-3F6 demonstrated Pneumocystis organisms in 38 cases with negative silver stains, whereas silver stain identified Pneumocystis organisms in only three MAB-3F6-negative cases. Of 70 patients with clinical Pneumocystis pneumonia at the time of the specimen was obtained, 59 had MAB-3F6-positive specimens, whereas 39 had organisms detected using Grocott's modified silver stain. Of 37 patients without clinically apparent Pneumocystis pneumonia any time in their course, 4 had abundant organisms and 33 had negative stains with MAB-3F6. MAB-3F6 detected Pneumocystis organisms in 22 of 31 cases of Pneumocystis pneumonia that had no organisms identified using Grocott's silver stain (X2 = 5.76, P = 0.016). MAB-3F6 immunochemical staining is a more sensitive method than Grocott's modified silver stain to detect Pneumocystis organisms.

Fibrous histiocytoma of the parotid gland.

The light microscopic, electron microscopic, and immunocytochemical characteristics of a case of a fibrous histiocytoma arising in the parotid gland is presented. This neoplasm is very rare in this site and must be distinguished from other spindle cell tumors of the parotid gland, particularly those of epithelial and myoepithelial origins. Histologic characteristics similar to those displayed by dermatofibromas and dermatofibrosarcoma protuberans help to differentiate this tumor from other spindle cell tumors. The absence of cytochemical epithelial markers and the electron microscopic demonstration of fibroblasts and histiocytes are useful in establishing the diagnosis. This tumor, which was confined completely to the parotid gland, appears to have arisen from mesenchymal elements within the gland.