Prognostic significance of bone marrow micrometastases in patients with gastric cancer

Background: Monoclonal antibodies (mabs) against components of the cytoskeleton such as cytokeratins allow single disseminated epithelial carcinoma cells to be detected in the bone marrow, The aim of this study was to examine the prognostic relevance of these cells in patients with gastric cancer and to evaluate by multivariate analysis their predictive value compared with conventional risk factors. Patients and Methods: A total of 1 x 10(6) cells from bone marrow aspirates were screened immunoctochemically for the presence and absolute number of disseminated tumor cells using mab CK2 to cytokeratin component no. 18. Patients were monitored prospectively for 30.6 +/- 15.2 months. Results: Between one and 122 CK2-positive cells per 1 million mononuclear bone marrow cells were present in 95 of 180 patients (53%). A similar prevalence of 51% was found in curatively operated patients (55 of 109). Comparison with conventional prognostic risk factors showed a correlation of cell dissemination with pathohistologic tumor (pT) stage (P = .07) and Bormann classification (P = .022). Tumor-cell content in the bone marrow predicted disease-free and overall survival in curatively resected patients (P = .007 and P = .049, respectively). Multivariate analysis, which included established risk factors, showed that extent of tumor-cell dissemination was an independent prognostic parameter for disease-free survival in T1/2 turners (P = .014; relative risk [RR], 1.84; 95% confidence interval [CI], 1.35 to 2.52), in intestinal type carcinomas according to Lauren (P = .008; RR, 1.62; 95% CI, 1.23 to 2.12), and in patients without lymph node involvement (P = .004; RR, 2.43; 95% CI, 1.22 to 4.82). Conclusion: Presence of disseminated tumor cells in bone marrow is indicative of systemic disease even in early-stage gastric cancer. The extent of tumor-cell presence in bone marrow correlates with prognosis in curatively resected patients. Therefore, ct positive bone marrow finding may be a selection criteria for adjuvant treatment because of minimal residual tumor load. (C) 1996 by American Society of Clinical Oncology.