Software for empirical building of biokinetic models for normal and decorporation-affected data.

This paper describes software ("RATDOSE") developed to analyze data from animal experiments investigating the efficacy of chelating agents. An empirical model building approach is used where, starting from the simplest model structures, one minimizes χ(2) by varying transfer rates in the model. Model complexity is increased as needed until the minimum attained value of χ(2) per data point decreases to about 1. This approach requires careful treatment of data uncertainties and independent checks of data self-consistency. The biokinetic models can include second-order kinetics to describe the chelation chemical reaction. The radiation dose to the animal is also calculated using S quantities specific for the animal, although the tissue weighting factors used to calculate the effective dose are those for the human.