BackgroundTo study the pharmacokinetics and clinical outcome of gemcitabine (2′-2′-difluoro-deoxcytidine [dFdC]) during intra-arterial versus intravenous delivery in locally advanced and regionally metastatic pancreatic cancer.Patients and methodsSeven patients with unresectable pancreatic cancer received escalating intra-arterial doses of gemcitabine ranging from 800 to 1400 mg/m2, after selective embolisation of all pancreatic blood supply, except for the tumour-feeding arteries. Four patients received intravenous gemcitabine (control). Venous blood samples at different time intervals were taken throughout 270 minutes for pharmacokinetic analyses of gemcitabine and its inactive metabolite 2′-2′-difluorodeoxyuridine (dFdU).ResultsPharmacokinetic data revealed differences in plasma concentrations between intra-arterial and intravenous delivery routes. The plasma concentration-time curve of gemcitabine during and after cessation of intra-arterial pancreatic target administration through the proximal splenic artery showed a profile with an area under the plasma concentration-time curve from 0 to 270 minutes (intra-arterial 29.0 ± 0.4 vs intravenous 331.0 ± 2.7 ng · min/mL; p < 0.0001) and peak plasma concentration (intra-arterial 1.1 ± 0.2 vs intravenous 7.6 ± 2.0 ng/mL; p < 0.0001) significantly lower than that for the corresponding systemic intravenous route. A plot of ln (% of dose) versus time showed a bi-compartmentalised metabolic model for intravenous administration of gemcitabine, one indicating rapid conversion of gemcitabine to dFdU, and another at a significantly lower affinity resulting in no conversion. Hence, this could be the main reason why dFdU was not detected in the systemic circulation during pancreatic intra-arterial target delivery. Furthermore, during intravenous administration a pseudo first-order rate constant (≈0.20 min–1) forin vivoconversion of gemcitabine to dFdU was estimated, indicating a rapid cellular deamination which was not shown in the intra-arterial route. Clinically, one patient had a partial response and six patients had a stable disease after intra-arterial administration of gemcitabine. The median time to disease progression was 4 months and the median overall survival was 5 months. One patient survived for 26 months. No grade III or IV toxicity was documented.ConclusionIntra-arterial administration of gemcitabine has a major advantage related to reduced toxicity as increasing the dose through this administration route will eventually result in pancreatic cellular drug target delivery prior to systemic availability. Despite the low number of patients recruited, the clinical results are encouraging and this approach should be tested in a randomised study.