Natural killer cells in the host response to melanoma.

Using an immunoperoxidase technique, we have investigated natural killer (NK) cells in the host response to malignant melanomas and melanocytic nevi in frozen sections. Eight primary melanomas, 12 metastatic melanomas, and 31 dysplastic nevi were studied. NK cells were identified phenotypically using an antibody, B73.1, against an Fc receptor present only on NK cells and neutrophils. Rare NK cells were identified in three of 31 dysplastic nevi and in one of eight melanomas. In contrast, significant numbers of NK cells were identified in ten of 12 metastases.

Immunoperoxidase localization of lymphocyte subsets in the host response to melanoma and nevi.

Using an avidin:biotin immunoperoxidase system with monoclonal antibodies to lymphocyte subsets, we have investigated the host response to malignant melanoma and melanocytic nevi in frozen sections. Eight primary melanomas, eight metastases, three dysplastic nevi, and two dermal nevi were studied with antibodies T11, T4, T8, and B1. Sections were read in a semiquantitative manner by two observers. Virtually all lymphocytes in these lesions were T-cells (T11 positive). In all primary melanomas, in the majority of metastases, and in all dysplastic nevi, both T4- and T8-positive cells were present. In two of eight metastases, tumor cells stained with T4, and in one case, melanoma cells stained with B1 antibody. The host response to melanoma involves primarily T-cells and includes both the helper:inducer (T4) and suppressor:cytotoxic (T8) subsets.

Invasive thymoma: the role of mediastinal irradiation following complete or incomplete surgical resection.

To evaluate the role of mediastinal irradiation (RT) following surgery for invasive thymomas, a clinical and pathologic review of 117 patients with the diagnosis of thymoma was completed. Fourteen cases were excluded because of the lack of histologic criteria for a thymic tumor, and the remaining 103 were classified according to a staging system as follows: stage I, completely encapsulated (43); stage II, extension through the capsule or pericapsular fat invasion (21); stage III, invasion of adjacent structures (36); and stage IV, thoracic dissemination or metastases (3). The 5-year actuarial survival and relapse-free survival rates were 67% and 100% for stage I, 86% and 58% for stage II, and 69% and 53% for stage III. No recurrences occurred among stage I patients after total resection without RT. However, eight of 21 patients with invasive (stage II or III) thymomas had mediastinal recurrence as the first site of failure following total resection without RT. The 5-year actuarial mediastinal relapse rate of 53% in this group compares unfavorably with the mediastinal relapse rate seen among stage II or III cases following total resection with RT (0%) or following subtotal resection/biopsy with RT (21%). Despite attempted salvage therapy, five of eight patients with mediastinal relapse following total resection alone died of progressive disease. No significant difference was observed in the local relapse rate, overall relapse rate, or survival between those patients undergoing biopsy and RT v subtotal resection and RT for invasive thymomas (stages II and III). Total resection alone appears to be inadequate therapy resulting in an unacceptably high local failure rate with poor salvage therapy results.

Myasthenia gravis and primary squamous cell carcinoma of the thymus: a case report.

We detail a case of primary squamous cell carcinoma of the thymus in a patient with myasthenia gravis. The clinical course of the patient and the gross and histologic appearance of the thymus are discussed. Appropriate therapy for patients with these tumors includes resection of the tumor followed by radiation therapy and continued surveillance for recurrence. Review of the literature indicates less than 50% survival at 10 years and 22% recurrence rate at 5 years for patients with this tumor.

Diagnosis of occult meat aspiration by fiberoptic bronchoscopy.

Cytologic examination of bronchial washings from a patient with a persistent localized pulmonary infiltrate revealed large numbers of striated muscle fibers. The patient died shortly after bronchoscopy, and postmortem examination provided evidence of recurrent aspiration pneumonias. Since skeletal muscle fibers are not likely to enter the tracheobronchial tree from any endogenous source, it is proposed that this unusual cytologic finding is virtually diagnostic of recent food aspiration.

Immunohistology of the benign lymphoepithelial lesion in AIDS-related lymphadenopathy: a case report.

We report a case of a benign lymphoepithelial lesion of the parotid gland in a patient with progressive generalized lymphadenopathy (PGL) related to human immunodeficiency virus. Serologic studies and immunoperoxidase studies for lymphocyte subsets, which have not previously been reported in this lesion, suggest involvement of an intraparotid lymph node by PGL rather than Sjogren's syndrome.

Role of intraoperative cytopathology in pediatric surgical pathology.

Studies regarding the efficacy of intraoperative cytopathology (IOC) of surgical specimens in the pediatric population are almost non-existent, despite their recent popularity in adults. To determine the utility of IOC in children, we examined 58 cases from 54 pediatric patients (neonate to 18 years of age) who had cytologic smears performed in addition to or instead of frozen section (FS) examination during their surgical procedure. Knowing only the patient's age, sex, anatomic site, and any pertinent radiographic or historic data, three pathologists independently reviewed and issued a diagnosis using only the IOC smears. Subsequently, in 28 cases that also had accompanying FS examination, both IOC and FS analysis were interpreted for a composite final diagnosis. Three cases were judged unsatisfactory because of sparse cellularity of smears. Correct classification of the smears as being benign or malignant for each pathologist was 98%, 94%, and 94% using IOC alone and 98%, 94%, and 96% using combined IOC and FS examination. The most frequent anatomic sites were bone (15 cases) and lymph node (14 cases). The most common diagnoses were malignant small round cell tumor (22 cases) and benign lymphoid tissue (10 cases). The records of all 55 cases were reviewed in the second phase of our study. Twenty-seven cases (49%) were found in which IOC diagnoses were rendered without a concurrent FS examination. These were correctly interpreted in 26 of 27 cases (96%) in the determination of a benign versus malignant disease process. The tissue sample measured < or = 2 cm in 15 of 27 cases (56%) in this latter group. Intraoperative cytopathology diagnoses in this group were rendered by various faculty members on call for FS examination and, in some cases, by fifth-year residents with faculty supervision. Without minimizing the degree of difficulty in the interpretation of pediatric IOC, we conclude that it serves as a useful supplement in FS diagnosis and, in some situations (particularly when tissue is limited), can replace histologic FS examination.

Extrarenal Wilms' tumor: an analysis of four cases.

During a review of Wilms' tumor, four located external to the kidney were identified. Patient age ranged from 7 months to 4 years; three were female. One neoplasm was located in the parametrial connective tissue to the left of the uterus; both kidneys were radiographically normal. Three neoplasms were located in the right retroperitoneum adherent to the surface of the ipsilateral kidney, but separated from the parenchyma by a thick fibrous capsule. Two were attached to the upper pole, while the third was attached to the midportion of the kidney. Radiologic studies showed displacement of all three kidneys, but intravenous pyelogram (IVP) revealed no calyceal distortion. All four neoplasms were favorable histology Wilms' tumor: one was monophasic epithelial type, one was monophasic blastemal type, and two had a mixture of stromal, epithelial, and blastemal tissue. No teratomatous elements were present. Immunoperoxidase staining for cytokeratin (AE-1, AE-3, CAM 5.2), vimentin, and epithelial membrane antigen (EMA) showed the strongest focal positive staining of tubular structures with CAM 5.2, and slight staining with EMA. Staining reaction to vimentin was variable, but negative in most areas. Three tumors extracted from paraffin were diploid by quantitative flow cytometric DNA analysis; in one case, flow cytometry could not be performed. Clinical follow-up from 2 years to 6 years showed all children to be alive without evidence of disease. Based on the similarity to conventional renal Wilms' tumor, these findings support the hypothesis of displaced mesonephric/metanephric rests for the origin of extrarenal Wilms' tumor.

Cortical versus medullary thymomas: a useful morphologic distinction?

We have tested the hypothesis that thymomas can be classified solely on the basis of epithelial cell morphology and that this distinction is prognostically useful. One hundred thymic epithelial tumors were classified according to the morphologic resemblance to cortical or medullary thymic epithelium and to the traditional classification (lymphocytic, epithelial, mixed, and spindled). Follow-up data was obtained on 78 patients. Fifty-eight percent of the tumors were invasive. Nineteen of the invasive tumors relapsed and none of the non-invasive tumors relapsed (P less than .0001). Four of nine tumors with microscopic invasion through the capsule recurred. Statistical analysis showed no significant differences in relapse-free survival for any of the histologic categories. Ninety-four percent of the tumors studied were keratin positive and all were chromogranin negative. Carcinoembryonic antigen was negative for all but one cytologically malignant tumor; of the tumors 75% were epithelial membrane antigen positive, 80% were Leu-7 positive, and 11% were neuron specific enolase positive. Seven of 12 tumors tested expressed HLA-DR. There was no correlation between immunoreactivity and classification. The morphologic cortical/medullary distinction is conceptually attractive but appears clinically to be no more advantageous than the traditional classification.

CD5 labeling of thymic carcinomas and other nonlymphoid neoplasms.

We studied 109 thymic tumors and 423 other neoplasms for immunocytochemical expression of CD5 in paraffin-embedded tissue using 2 monoclonal antibodies. One of the antibodies (clone CD5/54/B4) labeled 7 (29%) of 24 thymic carcinomas but did not stain any of the other neoplasms. In contrast, the other antibody (clone 4C7) labeled 15 (62%) of 24 thymic carcinomas, 2 (2%) of 84 thymomas, 3 (8%) of 37 lymphomas, and 14 (4%) of 386 other tumors. Among the lymphomas, 4C7 labeled 2 of 4 peripheral T-cell lymphomas and 1 of 4 anaplastic large cell lymphomas. Clone 4C7 produced strong labeling of reactive T lymphocytes, whereas CD5/54/B4 staining of these cells was weak or absent. No other normal cells seemed to stain with either antibody. In conclusion, one CD5 antibody clone (CD5/54/B4) specifically identified 30% of thymic carcinomas. The other CD5 antibody (clone 4C7) identified 62% of thymic carcinomas but also labeled occasional other neoplasms.

Antibody NCL-CD5 fails to detect neoplastic CD5+ cells in paraffin sections.

Many low grade B-cell lymphomas, and most T-cells lymphomas, express CD5 on their surface. This expression has been demonstrated in fresh cells or frozen sections. Recently, a new monoclonal antibody to CD5, NCL-CD5, has been introduced that detects CD5 in paraffin-embedded tissues. Paraffin-embedded tissues from five patients with small lymphocytic lymphoma (SL), four patients with chronic lymphocytic leukemia (CLL), and one patient each with large granular lymphocytic leukemia, diffuse large cell lymphoma (T cell), and mycosis fungoides were stained with the NCL-CD5 antibody after unmasking the antibody with the steam/citrate technique. All these cases had CD5 positivity demonstrated by flow cytometry or on cytospin or frozen section preparations. In addition, one case of angiofollicular lymph node hyperplasia (CD5+), two cases of SL/CLL (CD5-), and one case CD5- T-cell lymphoma were also investigated. Of the 12 CD5+ malignancies, only 1, an SL, was positive with the NCL-CD5 antibody. In seven of these cases, both B-5 and formalin-fixed tissues were tested; the one positive case was positive only in the formalin tissue. The three CD5- malignancies were also negative in paraffin sections (P = .001; McNemar test). However, reactive T cells did stain in these biopsy sections. The case of Castleman's disease showed many CD5+ cells with the NCL-CD5 antibody. Although NCL-CD5 does indeed stain reactive T cells in paraffin sections, it does not appear to stain neoplastic CD5+ cells.

CD5 immunoreactivity of epithelial cells in thymic carcinoma and CASTLE using paraffin-embedded tissue.

Because the histologic features may resemble those of other mediastinal malignancies, thymic carcinoma can be difficult to diagnose, particularly if the primary site is uncertain. In an effort to facilitate this diagnosis, the authors have evaluated the use of immunohistochemistry with an antibody to CD5 (NCL-CD5). Nine thymic carcinomas, 15 thymomas, 8 lymphomas, 10 poorly differentiated lung carcinomas, 10 breast carcinomas, 1 mediastinal seminoma, and 1 thyroid carcinoma showing thymus-like differentiation (CASTLE) were studied. Four of 9 poorly differentiated carcinomas of the thymus were CD5 positive. The one CASTLE was CD5 positive. All other tumors were negative. CD5-positive lymphocytes were internal controls in every case. CD5 labels some thymic carcinomas in paraffin-embedded tissue, whereas other tumors studied were negative. CD5 immunoreactivity of CASTLE appears to support thymic derivation of this tumor.

Dendritic reticulum cells and immunophenotype in aspiration biopsies of lymph nodes. Value in the subclassification of non-Hodgkin's lymphomas.

Twenty-seven lymph node aspirates were identified for which histologic confirmation of non-Hodgkin's lymphoma was subsequently obtained. Fifteen aspirates interpreted as reactive hyperplasia were also examined. All aspirates were studied by immunoperoxidase on cytospin preparations with the use of antibodies DRC1, kappa, lambda, CD3, CD5, and CD20. The follicular lymphomas could not be identified reliably by morphologic examination of aspirate smears. Clusters of DRC1-positive (DRC1+) cells were present in seven of seven follicular lymphomas, one of one mantle zone lymphoma, and one of seven small lymphocytic lymphomas. Rare DRC1+ cells were present in one of one diffuse mixed and one of seven large cell lymphomas. One lymphoblastic, one Burkitt's, and two diffuse small cleaved cell lymphomas had no DRC1+ cells. None of the seven follicular lymphomas was CD5 positive (CD5+), whereas five of the seven small lymphocytic lymphomas were CD5+. Conversely, all seven follicular lymphomas were CD20-positive (CD20+), but only one of seven small lymphocytic lymphomas was CD20+. Nineteen of the lymphomas, including all 7 of the follicular lymphomas, were either kappa or lambda positive. The other eight lymphomas were T-cell (1), B-cell (1), true histiocytic (1), or "null" cell (5). The reactive aspirates had both kappa- and lambda-positive B-cells. Seven of the 15 had clusters of DRC1+ cells. To further evaluate these antibodies, the authors studied 29 additional, surgically biopsied, non-Hodgkin's lymphomas that had not been aspirated. Similar results were obtained, except that three of five diffuse small cleaved cell lymphomas had DRC1+ cells. DRC1, in conjunction with antibodies to CD5, CD20, kappa, and lambda, helps to distinguish follicular lymphoma from small lymphocytic lymphoma. DRC1 is not useful in separating reactive hyperplasia from follicular lymphoma.

Lymphoglandular bodies in fine-needle aspiration cytology smears.

Lymphoglandular bodies (hyaline bodies or lymphoid globules), when found in cytology smears from fine-needle aspirates, have long been accepted as being diagnostic of lymphoid tissue. To investigate the validity of this association as it relates to malignant tumors, we examined cytologic smears from 132 fine-needle aspirates of malignant neoplasms. Three experienced observers independently scored Diff-Quik-stained smears as to cellularity and number and size of lymphoglandular bodies. Discrepancies were resolved by consensus. Results of the fine-needle aspiration biopsies revealed 6 of 104 nonlymphoid malignancies with easily identifiable lymphoglandular bodies (defined as > 2 lymphoglandular bodies per high-power field) and 3 with numerous lymphoglandular bodies (> 10 per high-power field). These tumors consisted of two cases of small-cell carcinoma, four non-small-cell carcinomas, one ganglioneuroblastoma, one melanoma, and one seminoma. The tumors had few, if any, lymphocytes. Of the 28 lymphomas, 5 had easily identifiable lymphoglandular bodies and 19 had numerous lymphoglandular bodies. Although lymphoglandular bodies in the background of cytologic smears taken from malignant tumors are useful in alerting the pathologist to the possibility of lymphoma, there are exceptions.

Leu M1 and S100 in Hodgkin's disease and non-Hodgkin's lymphomas.

Leu M1 positivity of Reed-Sternberg (RS) cells has been reported. The authors studied the specificity and sensitivity of Leu M1 in Hodgkin's disease (HD) and non-Hodgkin's lymphoma (NHL). Within NHL, they particularly selected cases that were confused with HD. The authors also studied S100 antigen to determine the pattern of staining in HD and NHL. Paraffin-embedded sections of 23 HD cases (3 lymphocyte predominate, 10 nodular sclerosing, 10 mixed cellularity) and 22 NHL cases (13 diffuse large cell, 5 diffuse mixed small and large cell, 4 others) were studied using an ABC technic. In 20 of 23 HD cases, RS cells and variants were Leu M1+; most cases contained prominent paranuclear positivity; some had diffuse cytoplasmic staining; and some had apparent staining of the cell surface. Neutrophils were intensely positive for Leu M1 and occasional histiocytes also were labeled. In two of the three negative cases (MC), the neutrophils were only weakly positive, thus suggesting a problem with tissue preparation. Of 22 NHL cases, 15 were totally Leu M1 negative. In six cases, rare or occasional tumor cells contained Leu M1 positivity in either a weak punctate, granular, or surface pattern. In an additional case, extensive pleomorphic cell staining was seen indistinguishable from that observed in RS cells; this case was the fourth recurrence of a primary skin NHL which began two years earlier as a pure small cleaved cell NHL. A total of three cases had positive pleomorphic cells. Some carcinomas were also Leu M1 positive. Concerning S100 antigen, the authors found scattered non-neoplastic cells throughout both HD and NHL samples; no tumor cells stained with this antigen. The negative S100 reaction of RS cells fails to support the argument for a dendritic cell origin. In properly prepared tissue, Leu M1 staining is quite sensitive for RS cells and variants, displaying a characteristic pattern. However, occasional Leu M1 positivity identified in NHL raises doubt as to its complete specificity.

Neuroendocrine tumors of the thymus: a clinicopathological and prognostic study.

BACKGROUND: Neuroendocrine tumors of the thymus are rare, histologically diverse neoplasms with an unpredictable clinical behavior. This study provides a useful clinicopathological classification and determines the relevance of specific prognostic factors. METHODS: Ten neuroendocrine tumors of the thymus were analyzed for specific clinical and pathological features. Prognostic factors of these cases and 71 previously published cases were evaluated by Kaplan-Meier survival curves and Cox multivariate hazard model. RESULTS: There were 7 males and 3 females, with ages ranging from 26 to 77 years. Cases were classified as carcinoid tumor (2), atypical carcinoid tumor (6), and small cell carcinoma (2). An advanced clinical stage was evident in all instances with frequent recurrence (4) and metastases (8), and a short disease-free survival. Overall mortality was 60%. Statistical analysis of current and previously published cases (n = 81 total) revealed that unresectability (p = 0.0001), extent of surgical resection (p = 0.0002), and advanced clinical stage at presentation (p = 0.03) were associated with higher mortality. By multivariate Cox regression analysis, unresectability (p = 0.02) and advanced clinical stage (p = 0.03) were associated with decreased survival. CONCLUSIONS: Neuroendocrine tumors of the thymus can be classified into distinct clinicopathological entities, and specific factors have prognostic relevance.

Epidermal growth factor receptor expression in squamous cell carcinoma of the hypopharynx.

Expression of the epidermal growth factor receptor (EGFR) has been demonstrated in normal and malignant squamous epithelia. Its presence has been suggested to be important in the pathophysiology and prognosis of epithelial cancers. Archival tumour specimens from 57 patients with squamous cell carcinoma of the hypopharynx were studied using OM-11-951, a new murine anti-EGFR monoclonal antibody which recognizes the receptor on deparaffinized tissue. By visual inspection, 28 (49%) tumours were EGFR negative; 29 (51%) tumours were EGFR positive. While patients whose tumours were EGFR positive were younger, there was no significant correlation with other clinical or pathological variables (including grade and stage). Patients whose tumours were EGFR negative had a median survival of 21 (95% CI 4.3-37.7) months compared with a median survival of 17 (95% CI 11.4-22) months for patients whose tumours were EGFR positive. The difference was not statistically significant. A multiple regression analysis did not demonstrate EGFR status to be important in predicting survival. These data cast doubt on the prognostic significance of EGFR expression in this neoplasm.

CD10 expression in follicular lymphoma versus reactive follicular hyperplasia: evaluation in paraffin-embedded tissue.

The differentiation of reactive follicular hyperplasia (RFH) from follicular lymphoma (FL) may be a diagnostic challenge and require the use of adjuvant studies. Antibodies to CD10 have been used with flow cytometry and frozen tissue sections and recently have become available for use with paraffin-embedded tissue. We sought to investigate the utility of CD10 in the differentiation of RFH from FL in paraffin-embedded tissue. Twenty-three cases of FL and 19 cases of RFH were evaluated after immunostaining with antibodies to CD3, CD20, BCL2, and CD10 antigens. Intensity of CD10 expression was graded. CD10 expression was present in 87% of cases of FL. Follicles demonstrated strong positivity in 70%. One CD10-positive case of FL showed equivocal staining for BCL2 and one an absence of staining for BCL2. Weaker but definite staining was present in 17%, and 13% were negative for CD10. In contrast, none of the cases of RFH was strongly positive for CD10 expression, with 26% demonstrating weak immunostaining, one case showing equivocal staining, and 68% being entirely negative. CD3, CD20, and BCL2 stained appropriately in all cases. These findings indicate that immunohistochemical staining of paraffin-embedded tissue sections for CD10 may help differentiate reactive from neoplastic follicular lesions.

Unusual biclonal plasma cell dyscrasia with crystallike inclusions producing colonic tumors (plasma cell polyposis) and terminating in T-cell lymphoma.

A 62-year-old man underwent a partial colectomy for persistent rectal bleeding. Multiple submucosal lesions up to 7 cm in diameter were identified and were composed of plasmacytoid and histiocytoid cells with crystallike inclusions. Ultrastructurally, inclusions were bounded by rough endoplasmic reticulum. Immunohistochemical studies demonstrated a monoclonal population of IgA kappa-positive plasma cells. Immunofixation using a saline extract of the tumor revealed two distinct bands: one for IgG kappa and another for IgA kappa. Southern blotting revealed a germline configuration for the immunoglobulin heavy-chain gene and the T-cell antigen receptor beta-chain gene. A bone marrow aspirate contained 12% plasma cells, some of which had cytoplasmic inclusions. Eight months after surgery, the patient developed a large mediastinal mass. Histologic and immunohistochemical studies indicated that this mass was an undifferentiated neoplasm. However, gene rearrangement studies indicated a T-cell lymphoma. This patient illustrates a rare association between a plasma cell neoplasm and a T-cell lymphoma.

Kaposiform hemangioendothelioma of the thymus.

Kaposiform hemangioendothelioma is a rare pediatric neoplasm that presents most commonly in the soft tissues. We report the case of a 1-month-old infant who presented with stridor and was found to have a diffusely infiltrating tumor in the thymus that extended into the pericardium and up the carotid sheaths. Histologic examination revealed a vascular tumor infiltrating among the lobules of the lymphocyte-depleted thymus. The lesion had features of both a capillary hemangioma and Kaposi sarcoma. Immunoperoxidase studies on formalin-fixed, paraffin-embedded tissue demonstrated the neoplastic endothelial cells to be positive for vascular markers CD31 and CD34. Antibody to factor VIII-related antigen labeled feeding vessels, but failed to stain the lobules of tumor. Although these tumors have been treated in a fashion similar to capillary hemangiomas in the past, it may be important to differentiate Kaposiform hemangioendotheliomas because of their association with Kasabach-Merritt syndrome and recent success with more aggressive chemotherapy regimens.