In a European joint project carried out in 6 laboratories a disease-oriented program was set up consisting of a panel of 7 tumor types, each represented by 4 to 8 different human tumor lines, for secondary screening of promising anticancer drugs. Human tumor lines were selected on the basis of differences in histology, growth rate, and sensitivity to conventional cytostatic agents. Xenografts were grown s.c. in nude mice, and treatment was started when tumors reached a mean diameter of 6 mm in groups of mice where at least 6 tumors were evaluable. Drugs were given at the maximum tolerated dose. For evaluation of drug efficacy, median tumor growth curves were drawn, and specific growth delay and treated/control x 100% were calculated. Doxorubicin (8 mg/kg i.v. days 1 and 8) was effective (treated/control < 50%, and specific growth delay > 1.0) in 0 of 2 breast cancers, 1 of 3 colorectal cancers, 2 of 5 head and neck cancers, 3 of 6 non-small cell lung cancers, 4 of 6 small cell lung cancers, 0 of 3 melanomas, and 3 of 6 ovarian cancer lines. Amsacrine (8 mg/kg i.v. days 1 and 8) was not effective, while datelliptium (35 mg/kg i.p. days 1 and 8) was active against 2 of 6 small cell lung cancer lines. Brequinar sodium (50 mg/kg i.p. days 1-5) showed efficacy in 4 of 5 head and neck cancers, 5 of 8 non-small cell lung cancers, and 4 of 5 small cell lung cancer lines. The project has been shown to be a feasible approach. Clinical activity for doxorubicin and inactivity for amsacrine against solid tumor types was confirmed in the human tumor xenograft panel. Additional anticancer drugs will be studied in the European joint project to further define the reliability of this novel, promising screening approach.