Using the radial basis function model for the Brdicka curve fitting

The Brdicka curve proved to be a good source of biomarkers, especially of cancer, due to its relation to the concentration of metalothionein protein. Various approaches for the automatic processing of the Brdicka curve were proposed and showed promising results. The Brdicka curve was examined from various points of view, but there are still missing pieces to the whole image of the multi-stage process of the Brdicka curve utilization. This paper presents a first attempt to the understanding of the content of Brdicka curve by the means of regression modeling. A radial basis function model was used to fit the Brdicka curve. This model proved to be complex enough to precisely fit the Brdicka curve. Unfortunately, in a follow-up experiment designed to validate the appropriateness of the model, the radial basis function model was found not to model the phenomena taking place during the Brdicka curve measurement.