The most appropriate route for regional administration of chemotherapeutic drugs to liver tumours was studied in a standardized rodent model: cells of Novikoff hepatoma were transplanted into the central liver lobe of Sprague-Dawley rats. From day 5 to day 12 after transplantation, the liver was continuously perfused with 420 mg/kg 5'-fluoro-2-deoxyuridine by subcutaneous osmotic micropumps via the hepatic artery (n = 20), the portal vein (n = 20) or both vessels together (n = 12). The tumour multiplication factor (TMF) and the vascularization of the tumour were evaluated. Arterial and combined infusion led to a highly significant reduction in TMF, but combined infusion was not more effective than arterial alone. Portal infusion had no significant effect. There was no correlation between vascularization and tumour response in arterial infusion, but a strong correlation in portal infusion. Thus chemotherapy via the portal route may be effective in selected tumours with considerable portal vascularisation.

The effect of continuous and discontinuous locoregional chemotherapy with Floxuridine (FUdR) was studied in a standardized and controlled animal model, using the transplantable Novikoff hepatoma in Sprague-Dawley rats. The liver was perfused after transplantation with a total of 420 mg/kg FUdR, via a fully implanted osmotic minipump or miniport and catheter in the hepatic artery, either continuously (n = 22) from day 5 to day 12, or discontinuously (n = 28) on days 5 and 8. Viable tumor volume and peritumorous cell infiltration were measured. No reduction in tumor volume was attained using discontinuous therapy, in contrast to a highly significant reduction using continuous therapy (P less than 0.001). Significantly less cell infiltration was found after discontinuous than after continuous therapy. In conclusion, continuous locoregional chemotherapy with FUdR was the superior administration method on measurable tumor effect, when compared to discontinuous infusion.