Seventy-three cases of malignant, atypical, and multicentric granular cell tumors of soft tissue were studied to clarify criteria for malignancy and prognostic factors. Six histologic criteria were assessed: necrosis, spindling, vesicular nuclei with large nucleoli, increased mitotic activity (> 2 mitoses/10 high-power fields at 200x magnification), high nuclear to cytoplasmic (N:C) ratio, and pleomorphism. Neoplasms that met three or more of these criteria were classified as histologically malignant; those that met one or two criteria were classified as atypical; and those that displayed only focal pleomorphism but fulfilled none of the other criteria were classified as benign. Hence, 46 cases were classified as histologically malignant, 21 as atypical (3 were multicentric), and 6 as benign (all were multicentric). The patients with benign multicentric and atypical granular cell tumors had no metastases and there were no tumor deaths. In contrast, 11 of 28 patients (39%) with malignant granular cell tumor with follow-up information died of disease at a median interval of 3 years; 8 of 28 (29%) were alive with disease, and 9/28 (32%) were disease free (median intervals, 2 and 7 years, respectively). There were local recurrences in 9 of 28 malignant cases (32%) and metastases in 14 of 28 (50%) (median intervals, each 2 years). Forty-eight cases were studied immunohistochemically; 100% expressed vimentin, 98% S-100 protein, 98% neuron-specific enolase, 69% CD57, and 65% CD68. Alpha-smooth muscle actin, desmin, epithelial membrane antigen (EMA), cytokeratins (with CAM 5.2 and KL-1), chromogranin, and HMB45 were not detected. The proliferative index with Ki67 (MIB 1) was 10-50% in 14 of 25 malignant tumors (56%), and immunostaining for p53 was detected in 50% or more of tumor cells in 17 of 25 (68%); both of these factors were statistically significant with regard to the histologic classification as benign, atypical, or malignant. Ultrastructural examination of 13 benign, atypical, and malignant granular cell tumors showed engorgement of the cytoplasm with complex granules and lysosomes, as well as Schwannian features. By flow cytometric DNA analysis, two of six malignant tumors were aneuploid, two were hyperdiploid, and two were diploid. One atypical tumor was aneuploid and all 11 benign tumors were either diploid (9 cases) or hyperdiploid (2 cases). Statistically significant adverse prognostic factors with regard to survival included local recurrence, metastasis, larger tumor size, older patient age, histologic classification as malignant, presence of necrosis, increased mitotic activity, spindling of tumor cells, vesicular nuclei with large nucleoli, and Ki67 values greater [corrected] than 10%. This study defines clinical and morphologic criteria for malignancy in granular cell tumors and shows that malignant granular cell tumor is a high-grade sarcoma with a high rate of metastases and a short survival.

The authors studied the familial occurrence of tumors in 154 individuals with gastric cancer by reviewing the clinical data and the genealogical tree of all patients registered in 1986 through 1987 in the Local Health Care District of Modena, Italy, for cancer of the stomach. Crude and age-adjusted (world population) incidence rates of gastric cancer were 34.0 and 21.4 new cases/100,000/year, respectively, in men, and 24.5 and 10.9 in women, respectively. Among first-degree relatives of the registered patients there were 30 cases of gastric carcinoma versus 15 cases in a control group matched for age and sex (Mantel-Haenszel odds ratio [M-H OR] 3.14, P less than 0.01). This excess of gastric neoplasms was observed in siblings (17 versus 7, M-H OR 4.33, P less than 0.02) but not in parents (13 versus 8, not significant). Besides gastric cancer, there was no significant excess of other type of tumors in case families. The familial occurrence of gastric cancer tended to be more frequent in patients with "diffuse" carcinoma (52%) than in subjects with "intestinal" cancer (33%), although the difference was not statistically significant. In conclusion, the current investigation suggests that a "family history" for gastric neoplasms is usually observed in approximately 10% to 15% of the registered cases. As already described for other common malignancies, therefore, the familial occurrence of gastric carcinoma suggests the existence of a genetic susceptibility to cancer of the stomach, at least in a fraction of these patients.

Aberrant crypt foci (ACF) have been identified in the methylene-blue stained mucosa of the human colon. Some lines of evidence suggest that ACF may be precursors of colon cancer. The objective of the present study was to establish morphological criteria able to define and classify ACF in histological sections. Twenty-four colectomy specimens were collected after operation for colorectal cancer and fixed in 10% formalin. Strips of grossly normal mucosa were stained in a 0.2% solution of methylene blue in saline for 5-10 min. The strips were measured, put on a glass slide and observed under a light microscope at x25. One hundred and fourteen ACF identified by topology were sectioned parallel to the muscularis mucosae. Eighty-four ACF were evident at histological examination and could be classified into three main groups: group A (61 ACF, 72.6%) including foci whose epithelial cells had regular nuclei, with only mild or focal crowding but no stratification, no mucin depletion and no dysplasia; group B (16 ACF, 19.1%), in which features of hyperplasia were evident; and group C (seven ACF, 8.3%) including foci with enlarged, crowded and stratified nuclei, mucin depletion, frequent mitoses, and evident dysplasia, diffuse or focal (mild in five cases, moderate in two) representing microadenomas. Finally, hyperplastic foci were significantly larger than foci of group A and C. Group B ACF were also more frequent in the rectum than in the colon. In conclusion, selected histological features allow the definition of groups of ACF, which may represent sequential steps in the development of human colorectal tumours.