The post-microinfusion transport of cis-diamminedichloroplatinum(II) (cisplatin) in rat brain has been modeled as a linear diffusion-reaction-permeation process. The model has been used to analyze the experimental data of Kroin and Penn to obtain the macromolecular binding constant of cisplatin in the brain, k = 0.0050 +/- 0.0023 min-1, and the capillary permeability, p = (9.0 +/- 4.4) X 10(-7) cm/s. Inclusion of saturation effects led to the same p value and a higher k value of 0.007 min-1. The corresponding diffusion length is 0.8 mm. The reaction constant is similar to those reported for plasma (0.008 min-1) and muscle (0.004 min-1), and the permeability value is within the range predicted by correlation with the permeability-octanol/water partition coefficient. Fits to data were accomplished with mathematical expressions giving the average total platinum concentration in saggital cerebellar sections which were not subdivided. Both time-dependent and steady-state solutions were obtained for the transport model, the former predicting a half-time to steady state of 3 h. Boundary effects were also investigated. Concentration profiles, calculated for a point source and for a 23-gauge cannula, were shown to differ by 7%. Similar comparisons between two profiles, one computed for an infinite diffusion range and another computed for drug diffusion into a flowing cerebrospinal fluid (CSF) at a finite range of 3 mm, showed differences of less than 3%. Free and bound drug forms, protein turnover, and CSF uptake have been accounted for as well as the percent infusate recoveries at 100 and 160 h reported by Kroin and Penn.(ABSTRACT TRUNCATED AT 250 WORDS)