Recently, several authors reported on the Protein Chip approach to analyze serum. They used SELDI-TOF-MS (surface enhanced laser desorption/ionization-time of flight-mass spectrometry) to identify patients with cancers of various origins in a highly sensitive and specific manner. In the current study, a similar approach was employed to analyze the serum of patients with various stages of gastric cancer. Control serum specimens from patients with gastritis (n = 19) and those with gastric cancer (Stage I: n = 6, Stage II or III: n = 6, Stage IV: n = 6, total: n = 18) were collected and analyzed by the Protein Chip biomarker system (Bio-Rad, Japan), a platform for SELDI-TOF-MS, and protein profiles were obtained and compared. The cation exchange chip (CM10) and the anion exchange chip (Q10) were used for processing before TOF-MS. nine proteins were significantly over-expressed (P < 0.05, Student t-test) in patients with gastric cancer compared to patients with gastritis. Among them, four protein masses with 2929 m/z, 3293 m/z, 3371 m/z, and 4213 m/z were found to be differentially expressed solely in patients suffering from peritoneal dissemination. All peaks were processed on CM10 chips. Employing one data mining method, CART (classification and regression trees), gastric cancer patients with peritoneal dissemination were successfully separated from those who had no peritoneal seeding. A validation study with a larger number of samples is mandatory; however, the detected peaks here might be candidates for biomarkers of peritoneal dissemination and/or gastric cancer. Moreover, further analysis of these four proteins might be helpful in revealing the mechanism of peritoneal dissemination, which at present has no cure.