Tumor and liver drug uptake following hepatic artery and portal vein infusion.

Anatomic dye injection studies of the blood supply of colorectal hepatic metastases suggest that tumors are supplied predominantly by the hepatic artery. Using 13N amino acids with dynamic gamma camera imaging in patients with colorectal hepatic metastases, it has been shown that hepatic artery infusion results in a significantly greater nutrient delivery to tumor compared with portal vein infusion. However, direct measurements of drug levels in tumor following hepatic artery and portal vein infusion in humans have not previously been reported. Patients with metastatic colorectal cancer confined to the liver received fluorodeoxyuridine (FUdR) through the hepatic artery or through the portal vein. All patients had previously failed systemic chemotherapy. Five patients with hepatic artery catheters were matched (by age, serum lactic dehydrogenase levels, percent hepatic replacement, and tumor size) with five patients with portal vein catheters. At operation, 3H-FUdR (1 microCi/kg) and 99mTc-macroaggregated albumin (MAA) (6 mCi) were injected into the hepatic artery or portal vein. Liver and tumor biopsies were obtained two and five minutes later. 3H and 99mTc were measured per gram tissue by scintillation and gamma counting. The mean liver levels following hepatic artery infusion (23.9 +/- 11.4 nmol/g) and portal vein infusion (18.4 +/- 14.5 nmol/g) did not differ. However, the mean tumor FUdR level following hepatic artery infusion was 12.4 +/- 12.2 nmol/g, compared with a mean tumor FUdR level following portal vein infusion of 0.8 +/- 0.7 nmol/g (P less than .01). This low level of tumor drug uptake after portal vein infusion of FUdR predicts minimal tumor response to treatment via this route. Thus, regional chemotherapy for established colorectal hepatic metastases should be administered through the hepatic artery.