Differentiation of diffuse malignant mesothelioma (DMM) from pleural carcinomatous metastases, e.g., of lung cancer (LC) may be difficult both for the clinician and the pathologist [1, 2] because of the limited diagnostic accuracy of radiologic criteria and conventional light microscopy including histochemistry [3,4]. Immunohistochemical detection of carcinoembryonic antigen (CEA) in tumor biopsy material has been shown to be a valuable diagnostic adjunct, as DMMs, unlike LCs and other carcinomas, express this antigen only rarely (9-11%) [5]. In some cases, however, the only method of obtaining suitable biopsy material is by open thoracotomy [2, 6]. Moreover, immunohistochemical examinations are expensive and not generally readily available. We therefore decided to find out what conclusions may be drawn from the much simpler determination of CEA levels in serum and in pleural effusion fluid (EF).We measured CEA in serum and EF using our own, and five commercially available radioimmunoassays or enzyme immunoassays (Abbott, Behringwerke, CIS, Roche, Pharmacia) with a serum level of 3 mg/L as the upper limit of reference values. In an initial partly retrospective study (learning phase), we examined patients with histologically proven DMM (n = 94) or LC (n = 79). Using ROC analysis, cutoff levels (serum CEA: 5.2 rag/L, EF CEA: 4.5 mg/L) were set so as to produce the most statistically significant differentiation (x2-test) between DMM and LC ( Table 1). The discrimination values specified and published [7] after the learning phase were then evaluated in a strictly prospective fashion in two additional sets of patients. From 1991 to 1992, we determined serum and EF CEA levels in 146 patients participating in the multicentric German DMM study, in all of whom the diagnosis of DMM had been reviewed by a panel of experienced pathologists. For comparison, we analyzed 124 patients who had presented with pleural effusions of unknown aetiology and had subsequently been shown to be suffering from metastatic carcinoma. There was a statistically significant difference between the DMM and the metastatic carcinoma group in terms of the frequency of elevated CEA levels (x2-test, p < 0.01) ( Table I). These results show that simple determination of CEA provides information that is useful for distinguishing between DMM

No abstract

No abstract

In a twelve-year period (1980-1992) 23 patients (9 men, 14 women; mean age 65 [25-82] years) with malignant non-Hodgkin lymphoma of the stomach, in localized stage I or II (Ann Arbor classification), were given chemo- and radiotherapy without preceding operation. The results were compared with those of a group of 34 patients (19 men, 15 women; mean age 55 [17-77] years) who had been admitted during the same period, having first been treated by surgical resection and most of them additionally by chemo- and/or radiotherapy. Three not previously operated patients with inadequate response to chemotherapy subsequently underwent laparotomy and remained in remission after further treatment. There was in each group one fatal recurrence. Five-year total probability of survival in the operated and non-operated groups was 90% and 92%, respectively; five-year total recurrence-free probability was 88% and 86%, respectively. There were no cases of severe bleeding or perforation. These results demonstrate that conservative management of localized gastric lymphoma by combined chemo- and radiotherapy does not bring about a rise in recurrence rate and does not seem to be inferior to surgical treatment.

Between March 1981 and February 1985, 93 out of 132 patients with a histologically confirmed diagnosis of malignant pleural mesothelioma were eligible for therapy and were prospectively assigned to receive either combined therapy or best supportive care, according to their personal preferences. Fifty-seven patients underwent multimodal therapy including surgical resection where possible, polychemotherapy, and radiation therapy in case of partial remission. Thirty-six patients received maximal supportive care only, as did 39 patients who were not eligible for treatment. The median survival was 13 months for treated patients compared to 7 for those receiving best supportive care and 5 for patients not amenable to treatment. Median progress-free survival was 6, 2, and 1 month respectively. Surgical resection did not prolong life expectancy within the treated group. In view of significant differences in the distribution of various cofactors over the two study groups, stepwise Cox model analyses were performed. Prognostic nontreatment variables related to prolonged survival were: good performance status, stage I and II, absence of chest pain, age below 50 years, and epithelial histology. Although in the Cox model analyses the survival improvement of patients being treated could be greatly attributed to other cofactors, multimodal treatment showed some prolongation of life expectancy.