Tumor and normal tissue pharmacokinetics of 5-Fluorouracil (5-FU) in patients can be determined with positron emission tomography scanning. However, the data obtained are of limited value because of the inability to distinguish catabolites (inactive species) from parent 5-FU and anabolites (cytotoxic species). In this paper, we have blocked 5-FU catabolism in one arm of a paired study with eniluracil, an inactivator of dihydropyrimidine dehydrogenase, enabling catabolite correction and calculation of tissue pharmacokinetic parameters to be achieved. Using this novel approach, we report for the first time that the net clearance of 5-[(18)F]FU from plasma into tumors (liver metastases and pancreatic tumor) of patients is low (K(I) = 0.0033 +/- 0.0005 ml plasma/ml tissue/min). In contrast, the initial (up to 10 min) clearance through catabolism in liver was high (K(I) = 0.7313 +/- 0.092 ml plasma/ml tissue/min). In the absence of eniluracil, catabolites in tumors accounted for 83% of total tumor exposure (range, 66-91%), whereas catabolites in liver accounted for 96% of total liver exposure (range, 94-98%). This study provides definitive evidence that the cytotoxicity of 5-FU in patients with gastrointestinal cancer could be compromised by its intrinsically low uptake by tumors, as well as decreased systemic availability through hepatic catabolism.