The purpose of this study was to establish a lung tumor model for the evaluation of chemopreventive agents against lung cancer in smokers. Lung tumor induction In A/J mice by 4-(methylnitrosamino)-l-(3-pyridyl)-l-butanone (NNK) and benzo[a]pyrene (BaP) was studied using protocols in which these two tobacco smoke carcinogens were given individually or in combination. Groups of female A/J mice were treated by either intragastric gavage (i.g.) or by intraperitoneal injection (i.p.) with various doses of NNK and/or BaP for 8 consecutive weeks. The mice were killed either 9 or 19 weeks later and tumors of the lung and forestomach were counted. The i.g. route of administration proved to be more satisfactory than i.p. administration, because it avoided complications due to tumor formation at the injection site and associated mortality. A dose-response relationship for lung tumor induction by i.g. administration of NNK and BaP in combination was established in the mice killed 9 or 19 weeks after completion of carcinogen treatment The highest total doses of NNK and BaP (a total of 24 μmol of each) induced more lung tumors than would have been expected by extrapolation from the lower doses. Comparisons of NNK and BaP given individually showed that BaP was more tumorigenic to the lung than NNK when given by the i.g. route; i.p. administrations of BaP were complicated by local tumor formation and mortality. The most favorable dosing regimen of NNK and BaP for evaluation of chemopreventive agents appears to be a total dose of 24 μmol of each, administered in eight weekly subdoses i.g., with sacrifice 9 weeks after completion of dosing. This regimen induced 103 ±4.4 lung adenomas/ mouse. A combination of benzyl isothiocyanate and phenethyl isothiocyanate, given 2 h prior to each gavage of NNK and BaP, was found to be an effective inhibitor of lung tumor formation, reducing the tumor multiplicity to 5.9±5.7 lung adenomas/mouse (P < 0.001) and completely Inhibiting forestomach tumor development The results of this study provide a convenient model for assessing the efficacy of chemopreventive agents against lung cancer induction by tobacco smoke carcinogens.