Identifiability and sensitivity analysis of a Photodynamic Therapy model

Simona Dobre (1) , Thierry Bastogne (1, 2) , Muriel Barberi-Heyob (1, 3, 4) , Alain Richard (1)

1 CRAN - Centre de Recherche en Automatique de Nancy
2 BIGS - Biology, genetics and statistics
3 CAV - Centre Alexis Vautrin
4 PHOTOMED - Médicaments Photoactivables - Photochimiothérapie

Photodynamic therapy (PDT) is an alternative treatment for cancer that involves the administration of a photosensitizing agent, which is activated by light at a specific wavelength. The photo-toxic phase of this therapy can be described by a dynamic model composed of six nonlinear differential equations. The model parameters can be used to compare photosensitizing agents in their capability to produce cytotoxic species. The practical issue is their estimation from in vivo experimental data. In this paper, a new approach is proposed to analyze the photophysical parameters estimability through a local practical identifiability study combined with a global sensitivity analysis. Results show that only three parameters can reasonably be estimated in a given and realistic experimental framework. Input design (light signal) and model reduction are currently in progress.

Mots clés: Identifiability; sensitivity analysis; nonlinear systems; biomedical systems

Domaines: Automatique / Robotique; Cancer

Simona Dobre, Thierry Bastogne, Muriel Barberi-Heyob, Alain Richard. Identifiability and sensitivity analysis of a Photodynamic Therapy model. 6th International Conference on Sensitivity Analysis of Model Output, SAMO 2010, Jul 2010, Milan, Italy. pp.CDROM. ⟨hal-00518270⟩