This study evaluated the effects of an association of ethiodized oil (Lipiodol Ultra Fluide, Laboratoires Guerbet, Aulnay-sous-Bois, France), with or without gelatin sponge, with doxorubicin (Adriamycin, Adria Laboratories, Columbus, OH) on the biodistribution and kinetics of doxorubicin during intraarterial injection. Eighteen patients with hepatocellular carcinoma on cirrhotic liver received a therapeutic injection into the hepatic artery of 50 mg of doxorubicin alone (Group 1; n = 4), or emulsified in 10 ml of ethiodized oil and 2.5 ml of ioxaglate (Hexabrix, Laboratoires Guerbet) with (Group 2; n = 7) or without (Group 3; n = 7) gelatin sponge embolization. Before treatment, the absence of intrahepatic shunts was verified by an injection of technetium-labeled albumin macroaggregates. The biodistribution of doxorubicin was studied on two fronts: (1) pharmacokinetic--by measurement of the doxorubicin blood level during the 48 hours after injection; and (2) scintigraphic (2 mg of doxorubicin were labeled with 2 mCi of iodine 131)--by examination of the scintigrams and calculation of the following parameters: tumours liver/nontumorous liver binding ratio (T/NT ratio), liver/liver+lungs+abdomen binding ratio, and doxorubicin half-life in tumorous tissue. Pharmacokinetics results showed the following: the peak plasma concentration was significantly higher in Group 1 as compared with Groups 2 or 3 (Group 1: 2.1 +/- 0.9 mg/ml; Group 2: 0.9 +/- 0.3 mg/ml; Group 3: 0.5 +/- 0.2 mg/ml); the area under curve calculated from time zero to 1 hour was lower in Groups 2 and 3 compared with Group 1. Examination of the scintigrams showed the following: diffuse activity throughout the organism (Group 1), diffuse activity with strong hepatic and tumorous binding (Group 2), and mostly hepatic and tumoral binding (Group 3). The liver/liver+lungs+abdomen binding ratio was 28% +/- 1% in Group 1, 36% +/- 5% in Group 2, and 63% +/- 7% in Group 3. The T/NT ratios were 1.0 +/- 0 (Group 1), 1.5 +/- 0.1 (Group 2), and 4.7 +/- 0.5 (Group 3). The doxorubicin half-lives in tumourous tissue were 0.7 +/- 0.1 days (Group 1), 1.8 +/- 0.2 days (Group 2), and 2.6 days (n = 1; Group 3). This study shows (1) that the association of ethiodized oil with doxorubicin lowers the peak concentration of doxorubicin and increases the intratumoral concentration and half-life of doxorubicin, and (2) that these kinetic ameliorations are even more pronounced after embolization. Therefore, from a kinetic standpoint, the doxorubicin-ethiodized oil-gelatin sponge association is the best.