Human tumors transplanted into nude mice have long been used to assess the effectiveness of antitumor drugs and yet there is still no established standard method in preclinical practice for screening new antitumor drugs in vivo using nude mice. Thus, a cooperative study on the feasibility of a human tumor/nude mouse system for the in vivo screening of drugs was conducted by the Japanese Research Society for Chemosensitivity of Cancer. Two human stomach cancers, H-111 and SC-6-JCK, and one human colon cancer, Co-4, were transplanted serially into nude mice and used as gastrointestinal tract tumors with stable tumor growth. The appropriate dosage of six well-known antitumor drugs [mitomycin C (MMC), cyclophosphamide (CPA), nimustine hydrochloride 1-(4-amino-2-methyl-5-pyrimidinyl) methyl-3-(2-chloroethyl)-3-nitrosourea hydrochloride (ACNU), cis-platinum (II) diaminodichloride (CDDP), adriamycin (ADM) and 5-fluorouracil (5-FU)] in human tumor-bearing nude mice was determined based on the maximum tolerance dose of the drug. The respective dosages were 6 mg/kg of MMC x 1 (i.p.), 120 mg/kg of CPA x 1 (i.p.), 30 mg/kg of ACNU x 1 (i.p.), 8 mg/kg of CDDP x 1 (i.p.), 8 mg/kg of ADM x 1 (i.v.), and 50 mg/kg of 5-FU q4d x 3 (i.p.). Three weeks after treatment, drug effectiveness was judged by the tumor growth inhibition rate. Treatment with these appropriate doses appeared to show the maximum effect of the respective drugs on the tumor-bearing nude mice.(ABSTRACT TRUNCATED AT 250 WORDS)