Rat 9L gliosarcoma cells infiltrating the normal brain have been shown previously to accumulate only approximately 30% as much boron as the intact tumor after administration of the boronated amino acid p-boronophenylalanine (BPA). Long-term i.v. infusions of BPA were shown previously to increase the boron content of these infiltrating tumor cells significantly. Experiments to determine whether this improved BPA distribution into infiltrating tumor cells after a long-term i.v. infusion improves tumor control after BNCT in this brain tumor model and whether it has any deleterious effects in the response of the rat spinal cord to BNCT are the subjects of the present report. BPA was administered in a fructose solution at a dose of 650 mg BPA/kg by single i.p. injection or by i.v. infusion for 2 h or 6 h, at 330 mg BPA/kg h(-1). At 1 h after the end of either the 2-h or the 6-h infusion, the CNS:blood (10)B partition ratio was 0.9:1. At 3 h after the single i.p. injection, the ratio was 0.6:1. After spinal cord irradiations, the ED(50) for myeloparesis was 14.7 +/- 0.4 Gy after i.p. administration of BPA and 12.9 +/- 0.3 Gy in rats irradiated after a 6-h i.v. infusion of BPA; these values were significantly different (P < 0.001). After irradiation with 100 kVp X rays, the ED(50) was 18.6 +/- 0.1 Gy. The boron compound biological effectiveness (CBE) factors calculated for the boron neutron capture dose component were 1.2 +/- 0.1 for the i.p. BPA administration protocol and 1.5 +/- 0.1 after irradiation using the 6-h i.v. BPA infusion protocol (P < 0.05). In the rat 9L gliosarcoma brain tumor model, the blood boron concentrations at 1 h after the end of the 2-h infusion (330 mg BPA/kg h(-1); n = 15) or after the 6-h infusion (190 mg BPA/kg h(-1); n = 13) were 18.9 +/- 2.2 microg 10B/g and 20.7 +/- 1.8 microg 10B/g, respectively. The irradiation times were adjusted individually, based on the preirradiation blood sample, to deliver a predicted 50% tumor control dose of 8.2 Gy ( approximately 30 photon-equivalent Gy) to all tumors. In the present study, the long-term survival was approximately 50% and was not significantly different between the 2-h and the 6-h infusion groups. The mode of BPA administration and the time between administration and irradiation influence the 10B partition ratio between the CNS and the blood, which in turn influences the measured CBE factor. These findings underline the need for clinical biodistribution studies to be carried out to establish 10B partition ratios as a key component in the evaluation of modified administration protocols involving BPA.