Gary L. Ellis

Mucoepidermoid carcinoma of the major salivary glands

The authors had previously conducted an investigation of minor salivary gland mucoepidermoid carcinoma, in which they demonstrated that certain clinical and histopathologic features were useful in predicting biologic outcome. The current study investigated the usefulness of these features in determining the prognoses of patients with mucoepidermoid carcinomas of the major salivary glands.

Mucoepidermoid carcinoma of intraoral salivary glands evaluation and application of grading criteria in 143 cases

The histopathologic criteria most useful for grading of mucoepidermoid carcinomas are controversial. To identify those histologic features most important in the grading of intraoral mucoepidermoid carcinomas, 143 cases of this disease with clinicopathologic correlations were studied. Twelve histopathologic features of each tumor and their clinical presentation were correlated with patient outcome. Seven patients died of disease, 5 had regional metastases only, 10 had recurrences only, and 121 had no additional problems. Clinical features suggesting aggressive behavior were short duration, presence of clinical symptoms, and location of tumor in the tongue and floor of the mouth. The histopathologic features that indicated high‐grade behavior were an intracystic component of less than 20%, four or more mitotic figures per ten high‐power fields, neural invasion, necrosis, and cellular anaplasia. The simultaneous assessment of these features showed improved prognostic correlation over individual parameters. A quantitative grading system was devised using these features. Tumors with a point score of 0 to 4 were considered low grade, and none of 122 patients with scores in this range died of their tumor, although 9 had recurrences only and 3 had regional metastases. Point scores of 7 or above indicated highly aggressive behavior. Six of ten patients with these high scores died of tumor. Most of these six patients had recurrences and regional metastases, and all had distant metastases. Two other patients had regional metastases only. Scores of 5 to 6 were considered intermediate between low‐grade and high‐grade scores because only 1 of 13 patients with these scores died of disease. Three of the five patients with regional metastasis had low‐grade tumors, indicating the inability of the grading system to identify them. Nonetheless, with an average follow‐up on these patients of 10 years after treatment of the metastasis, no patient had additional problems. The relative objectivity of our proposed grading system for intraoral mucoepidermoid carcinomas may help achieve more accurate and consistent grading of these rare tumors.

Acinic cell adenocarcinoma. A clinicopathologic analysis of 294 cases

Two hundred and ninety‐four cases of acinic cell adenocarcinoma were reviewed for the purpose of defining the clinical parameters and determining the distribution of the four histomorphologic tissue patterns and five cell types for correlation to biologic behavior. The vast majority occurred in the parotid gland. There was a male predominance and a peak incidence in the third decade of life. The tumors were usually less than 3 cm in diameter and were slow growing. Pain was a common symptom, but was not indicative of prognosis. Nearly one half of the neoplasms exhibited multiple tissue growth patterns, and three fourths of the tumors displayed more than one cell type. The microcystic pattern was seen most frequently, regardless of the biological behavior of the tumors. The well‐differentiated acinic cell was the most prevalent cell type except in cases with metastases, where the intercalatedduct cell type was slightly more frequent. Follow‐up of 244 cases revealed a recurrence rate of 12%, a metastatic rate of 7.8%, and death rate of 6.1%. Since all histomorphologic patterns and cell types were manifest in tumors which recurred, metastasized, or caused the death of the patients, it seems appropriate to consider these neoplasms as low‐grade adenocarcinomas rather than essentially benign with occasional unpredictable malignant behavior.

Calcifying odontogenic cyst. A review of ninety-two cases with reevaluation of their nature as cysts or neoplasms, the nature of ghost cells, and subclassification.

Ninety-two cases of calcifying odontogenic cyst (COC) were reviewed with special consideration of their nature as cysts or neoplasms, the nature of ghost cells, and classification on the basis of clinicopathologic features. The cases were divided into 79 (85.9%) cysts and 13 (14.1%) neoplasms. The cysts occurred as four variants: (1) nonproliferative COC (35 cases), characterized by a simple unicystic structure; (2) proliferative COC (17 cases), characterized by a cystic structure with multiple daughter cysts, extensive ghost cell formations, and marked tendency for calcification; (3) ameloblastomatous COC (11 cases), characterized by ameloblastoma-like, cyst-lining epithelium with ghost cells and calcifications; and (4) COC associated with odontoma (16 cases), which combined features of COC and odontoma. The neoplasms occurred as three variants: (1) ameloblastoma ex COC (two cases), which showed unifocal and multifocal intraluminal and intramural ameloblastoma proliferating from the COC-lining epithelium; (2) peripheral epithelial odontogenic ghost cell tumor (eight cases), which occurred in the gingiva and resembled peripheral ameloblastoma except for clustered ghost cells in the central portion of epithelial islands and the presence of juxtaepithelial dentinoid; and (3) central epithelial odontogenic ghost cell tumor (three cases). The latter showed ameloblastomatous or adenomatoid odontogenic tumor-like epithelial clusters with ghost cell formation and juxtaepithelial dentinoid. The clinical features of cystic and neoplastic variants were tabulated and described. On the basis of histopathologic features and their immunohistochemical reaction to polyclonal antikeratin antibody, it is suggested that ghost cells might be the result of coagulative necrosis.

A clinical and histomorphologic comparison of the central giant cell granuloma and the giant cell tumor

The clinical, histologic, and histomorphometric features of 42 giant cell tumors (GCT) of long bones and 49 central giant cell granulomas (CGCG) of the jaws were compared. These findings were also correlated with the clinical behavior of 25 cases of CGCG for which follow-up information was available. There was a female predilection for both lesions. The mean ages of patients with CGCG and GCT were 21 and 25 years, respectively. In contrast to CGCG, GCT rarely occurred in persons below the age of 10 years. The only statistically significant quantitative difference between the lesions at the histologic level was the greater number of nuclei in the giant cells of the GCT. There were four significant histologic differences between the two lesions, but 26% of the GCTs were histologically similar to most of the CGCGs and 10% of the CGCGs were histologically similar to most of the GCTs. Five of the 25 patients with CGCG for whom follow-up information was available had recurrences. The average age of those five patients was 11 years, compared to 29 years for those patients without recurrence. All five patients with recurrence were under 17 years of age, and they constituted 45% of the patients in this age group with follow-up. There were no quantitative or histologic differences between the recurrent and nonrecurrent CGCGs that were useful in predicting the likelihood of recurrence. Our findings suggest that the GCT and the CGCG represent a spectrum of a single disease process modified by the age of the patient and the site of occurrence.

Dental follicular tissue: Misinterpretation as odontogenic tumors

The dental follicle is radiographically and histologically observed in association with unerupted or impacted teeth. However, this normal tissue structure is often confused with odontogenic tumors by pathologists with limited experience in evaluating jaw lesions. This study was designed to evaluate the incidence and possible reasons for incorrect interpretation of dental follicles. From 1970 to 1988, 847 dental follicles and/or dental papillas from 663 patients were submitted to the Armed Forces Institute of Pathology (AFIP) by medical pathologists seeking diagnostic consultation. Nearly 84% of patients were in the second and third decades of life. The male to female ratio was 1.4:1.0. Over 70% of specimens were obtained from around impacted third molar teeth. Fifty-three percent of specimens were correctly interpreted by the contributing pathologists. Only a descriptive interpretation was given for 17%, no diagnosis was made in 10%, and 20% were incorrectly diagnosed. In descending order, the most frequent incorrect diagnoses were odontogenic cyst, odontogenic myxoma, odontogenic fibroma, ameloblastic fibroma, odontoma, and ameloblastoma. Dental papillas were most frequently misdiagnosed as odontogenic myxomas. The histologic features and diagnostic pitfalls are discussed, as well as the need to consider the clinical, radiographic, and microscopic features in arriving at a final diagnosis.

Spindle-cell carcinoma of the aerodigestive tract. An immunohistochemical analysis of 21 cases.

Immunohistochemical analysis of 21 prototypic mucosal spindle-cell carcinomas of the aerodigestive tract was performed at the Armed Forces Institute of Pathology (AFIP) to establish the usefulness of selected immunohistochemical markers in distinguishing spindle-cell carcinoma from other mucosal spindel-cell neoplasms. Immunoreactive keratin could be demonstrated in only 13/21 (62%) of cases. Coexpression of keratin and vimentin was demonstrated in 10/17 (59%) of the tumors evaluated for both of these intermediate filaments. All spindle-cell carcinomas lacked S100 protein, which is an immunoreactivity we would expect to find in spindle-cell malignant melanoma, one of the principal considerations in a differential diagnosis. Both alpha-1-antitrypsin (AAT) and alpha-1-antichymotrypsin (ACT) were demonstrated in the tumor cells in all cases. However, albumin had a similar distribution in the tumors, which suggested that passive uptake was a serious confusing factor. The results of this study indicate that AAT and ACT are unreliable markers for distinguishing spindle-cell carcinomas from malignant fibrous histiocytomas

Metastasizing mixed tumor of salivary glands. A clinicopathologic and flow cytometric analysis

Among salivary gland neoplasms are a group of rare tumors that are histologically identical to benign mixed tumors that inexplicably metastasize; they have been called metastasizing mixed tumor (MZMT) of salivary glands. We report the clinicopathologic features and flow cytometric findings for 11 cases of MZMT. At the time of discovery of metastatic disease, the patients, six women and five men, ranged in age from 20 to 83 years. Primary sites of involvement included the parotid gland (eight cases), submandibular gland (two cases), and the nasal septum (one case). With one exception, all the patients had at least a single recurrence of their primary mixed tumor, but two or more recurrences were the norm before development of metastatic foci. The metastases were discovered from six to 52 years following the occurrence of the primary tumor. Metastatic deposits were identified in bone, lung, regional lymph nodes, skin, kidney, retroperitoneum, oral cavity, pharynx, calvarium, and central nervous system. The metastases either occurred simultaneously with an episode of recurrent mixed tumor (n=5) or from 5 to 29 years after a recurrence (n=6). The treatment of the primary, recurrent, and metastatic neoplasms was surgical excision. Follow-up, ranging from 8 months to 16 years following the diagnosis of MZMT, revealed seven patients to be alive without disease (64%) and two dead of causes unrelated to metastatic disease (18%). Two patients (18%) died as a direct result of metastatic tumor at 3 and 2 years after metastasis of their mixed tumors. Flow cytometric analysis revealed a diploid DNA cell population in the primary and/or metastatic tumors in nine cases. Aneuploid DNA cell content was identified in two of the cases. DNA ploidy levels and cell proliferation rates were compared with those of conventional benign mixed tumors and also with malignant mixed tumors. Retrospective analysis of histologic parameters (mitotic rate, cellular pleomorphism, infiltrative growth, vascular or lymphatic invasion) and flow cytometric analysis failed to identify criteria to predict the development of metastasis in these neoplasms.

Salivary gland cystadenocarcinomas : A clinicopathologic study of 57 cases

Current classification schemes for salivary gland neoplasms categorize cystadenocarcinomas on the basis of a recurring histomorphologic pattern of cystic, and often, papillary growth without features of other specific types of salivary gland tumors. To ascertain the clinicomorphologic spectrum and biologic behavior of this tumor, the clinicopathologic features of 57 cystadenocarcinomas from the files of the Armed Forces Institute of Pathology were studied. Excluding five Veterans Administration military cases, men and women were equally affected. Patients ranged in age from 20 to 86 years (mean, 58.8; median, 64), and patients aged over 50 years accounted for 71% of cases. Thirty-seven tumors (65%) occurred in major salivary glands, 35 in the parotid, and two in the sublingual glands. The 20 minor salivary gland tumors (35%) involved, in descending order, the lips, buccal mucosa, palate, tongue, retromolar area, and floor of mouth. Grossly, the lesions were cystic or multicystic masses that ranged in size from 0.4 to 6.0 cm. Microscopically, all tumors demonstrated an invasive, cystic growth pattern, and 75% had a conspicuous papillary component. The predominant cell type varied among tumors and included small cuboidal cells (35 cases), large cuboidal cells (nine cases), and tall columnar cells (seven cases). Six cases exhibited an admixture of cell types. Ruptured cysts with hemorrhage and granulation tissue were common. All 40 patients with follow-up data were either alive or had died of other causes and were free of tumor a mean interval of 59 months after their initial surgery. Three tumors recurred locally (mean interval, 76 months). Three tumors were metastatic to regional lymph nodes at the time of diagnosis, and one patient developed a regional lymph node metastasis after 55 months. Salivary gland cystadenocarcinomas represent a distinct group of malignancies that have an indolent biologic behavior.

Melanotic neuroectodermal tumor of infancy. Clinicopathological, immunohistochemical, and flow cytometric study.

Twenty cases of melanotic neuroectodermal tumor of infancy (MNTI) are reported. The patients (13 females, seven males), whose ages ranged from 1 to 9 months (mean, 5 months), typically presented with a rapidly growing mass. Tumor sites included the maxilla (13 cases), mandible (three cases), dura (two cases), brain (one case), and skull/orbit (one case). The mean tumor size was 3.5 cm (range, 1.0-10.0 cm). Follow-up was obtained on 12 cases. Five tumors (45%) recurred within 4 months of diagnosis, but none metastasized. One surgical death occurred Histologic appearance was distinctive, with tubular or alveolar formations of large melanin-containing cells around nests of smaller neuroblastic cells possessing scant or fibrillar cytoplasm. Twelve tumors were studied immunohistochemically; tumor was positive for cytokeratin in 12 of 12, for HMB 45 in 12 of 12, for vimentin in seven of eight, and for epithelial membrane antigen (EMA) in four of nine tumors, mainly in the large cells. Neuron-specific enolase (NSE) (seven of 12) and Leu 7 (nine of 12) were positive in small and large cells; some tumors also expressed synaptophysin (four of 12), glial fibrillary acidic protein (GFAP, three of 12 tumors), or S-100 protein (two of 12 tumors). No staining was found for chromogranin, desmin, or carcinoembryonic antigen (CEA). Eight of 10 tumors studied had interpretable results on flow cytometry (FCM) (four DNA diploid, three DNA aneuploid, and one DNA diploid with a prominent shoulder). Tumor recurred locally in two of five cases with follow-up, and we were unable to demonstrate the usefulness of FCM in predicting recurrences. Further studies are necessary to define better the potential usefulness of FCM in predicting aggressive behavior. Distinctive morphology and multiphenotypic (epithelial, neural, melanocytic) expression distinguish MNTI from melanoma and metastatic neuroblastoma