Computational models of antibody-based tumor imaging and treatment protocols

We present improved computational models for investigating monoclonal antib ody-based protocols for diagnostic imaging and therapy of solid tumors. Our earlier models used a boundary condition (Dirichlet) that specified concen trations of diffusing molecular species at the interface between a prevascu lar tumor nodule and surrounding normal tissue. Here we introduce a concent ration-dependent Bur boundary condition with finite rates of diffusion in t he normal tissue. WP then study the effects of this new condition on the tu mor's temporal uptake and spatial distribution of radiolabeled targeting ag ents. We compare these results to ones obtained with the Dirichlet boundary condition and also conduct parameter sensitivity analyses. Introducing fin ite diffusivity for any molecular species in normal tissue retards its deli very to and removal from the tumor nodule. Effects are protocol- and dose r egimen-dependent. generally, however? mean radionuclide concentration and t umor-to-blood ratio declined, whereas relative exposure and mean residence time increased, Finite diffusivity exacerbates the negative effects of anti gen internalization, Also, the sensitivity analyses show that mean concentr ation and tumor-to-blond ratio are quite sensitive to transcapillary permea bility and limphatic efflux values, yet relatively insensitive to precise v alues of diffusion coefficients. Our analysis underscores that knowledge of antigen internalization rates and doses required to saturate antigen in th e tumor will be important for exploiting antibody-based imaging and treatme nt approaches. (C) 2001 Biomedical Engineering Society.