A phase I and pharmacokinetic comparison of hepatic arterial and peripheral vein infusions of bisantrene for liver cancer.

Bisantrene (NSC-337766) was administered to five patients with cancer of the liver (one case of hepatocellular carcinoma, two of metastatic carcinoma of unknown primary, two of metastatic colorectal carcinoma). Under fluoroscopic guidance, percutaneous hepatic venous catheters were placed in five patients and percutaneous hepatic arterial catheters in four. A fifth patient's hepatic arterial catheter was implanted at laparotomy. Hepatic plasma flow was estimated by the Fick principle using peripheral vein indocyanine green infusion. On the first day of treatment, patients received a 2- or 4 h hepatic arterial infusion of bisantrene (130 mg/m2); peripheral venous, hepatic arterial, and hepatic venous timed blood samples were drawn during and for 18 h after drug infusion. On the second day of treatment, 2- or 4 h peripheral vein infusion of bisantrene (130 mg/m2) was followed by the same blood sampling schedule. Patients were followed weekly for toxicity. Four patients received only one course of treatment, while a fifth received two courses. All patients experienced leukopenia (median nadir 2400/mm3; range 1400-2700/mm3). Two patients developed fever after drug infusion. No antitumor responses were observed. Plasma bisantrene concentrations were measured by HPLC. Pharmacokinetic analyses are reported for four patients. The hepatic extraction ratio ranged from 15% to 49%, hepatic plasma clearances were 0.029-0.353 1/min/m2; peripheral vein areas under the concentration-time curve during hepatic arterial infusion ranged from 35% to 50% of peripheral vein areas under the curve during peripheral vein infusion. We conclude that hepatic arterial bisantrene infusion offers only modest pharmacokinetic advantage to the target organ or to the systemic circulation over peripheral vein infusion.