A new form of dosage (MMC-CH) was composed of activated carbon particles adsorbing mitomycin C. Intraperitoneal administration of MMC-CH was tested clinically for prophylactic and therapeutic effects on peritoneal carcinomatosis of gastric cancer. The criteria of MMC-CH's administration were equal or less than 70 years old, more than 40 kg in body weight, no disfunction of liver and kidney, no particular findings in electrocardiography, S2 or S3 in the grade of serosal invasion, P0, P1, P2 or P3 in the grade of peritoneal dissemination, according to the General Rules for the Gastric Cancer Study in Surgery and Pathology by the Japanese Research Society for Gastric Cancer. MMC-CH was given to 44 patients undergoing gastrectomy for gastric cancer in our department from 1985 to 1988. The 44 patients were composed of 12 patients with P0 findings (P0 patients), 8 patients with P1 findings (P1 patients), 12 patients with P2 findings (P2 patients), and 12 patients with P3 findings (P3 patients). MMC-CH at 50 mg/person in terms of mitomycin C was administered intraperitoneally before the operation wound was closed. Fifty-seven patients in our department from 1983 to 1987 for whom the same criteria were applicable and did not receive MMC-CH therapy, served as the control group. The 57 patients were composed of 23 P0 patients, 21 P1 patients, 10 P2 patients, and 3 P3 patients. There was statistically with chi 2 test no significant difference of age, sex, depth of infiltration macroscopically and microscopically defined progression of lymph-nodal metastases between the MMC-CH group and the control group. Survival rate was calculated with Kaplan-Meier's method in the overall patients in each of the MMC-CH group or the control group. The overall survival rate in the MMC-CH group was statistically significantly (p less than 0.01-0.05) higher from day 460 to day 552 and from day 736 to day 800 than that in the control group. Next, the patients were classified into two subgroups, namely the subgroup composed of P0 patients and the subgroup composed of P1, P2, and P3 patients, in order to examine each of the MMC-CH's prophylactic effects on subsequent dissemination or its therapeutic effects on established dissemination. Survival rate was calculated with Kaplan-Meier's method in the two subgroups through the same procedures used for the overall survival rate.(ABSTRACT TRUNCATED AT 400 WORDS)