BACKGROUND: Different modalities of cytostatic lung perfusion were compared regarding plasma and tissue drug concentrations to assess the efficacy of an endovascular blood flow occlusion technique. METHODS: A cytostatic lung perfusion study with doxorubicin hydrochloride was performed on large white pigs (n = 12). Plasma and tissue concentrations of doxorubicin were compared for isolated lung perfusion with open cannulation (ILP), blood flow occlusion perfusion with open cannulation of the pulmonary artery alone (BFO), and intravenous drug administration (i.v.). In a fourth group, thoracotomy-free BFO perfusion was performed by endovascular balloon catheterization of the pulmonary artery (endovascular BFO). The 3 animals in this group were used to compare the doxorubicin-perfused pulmonary tissue with the contralateral nonperfused lobes after 1 month. RESULTS: The mean lung tissue doxorubicin concentration at the end of perfusion was 19.8 +/- 1.6 microg/g after ILP, 27.6 +/- 2.2 microg/g after BFO (p = not significant), and 3.0 +/- 0.8 microg/g after i.v. perfusion (p < 0.01). Whereas doxorubicin was not detectable in the plasma in the ILP group, concentrations ranged from not detectable to 0.44 microg/mL in the BFO group and from 0.31 to 0.84 microg/mL in the i.v. group (p < 0.05). Mean myocardial tissue concentration was not significantly different after BFO than i.v. perfusion (1.1 +/- 0.5 microg/g and 1.8 +/- 0.1 microg/g, respectively). In the endovascular BFO group, balloon-blocked pulmonary artery perfusion was successfully performed in all animals, and after 1 month, lung tissue showed no cytostatic-induced histologic changes. CONCLUSIONS: Compared with ILP, BFO cytostatic lung perfusion produced an insignificantly higher lung-tissue concentration, corresponding to a sixfold to ninefold higher level than after i.v. perfusion. Plasma drug levels during BFO perfusion were lower than during i.v. perfusion. Endovascular BFO may be a promising technique for repeated cytostatic lung perfusion.