Pascal Nguyen will defend his PhD thesis entitled “Fluorescence Diffuse Optical Tomography in Near-Infrared II window for murine models : from physics-based model to interpretable AI-enhanced reconstruction” on Thursday, June 18 at 02:00 p.m. in Ponte amphitheater, campus St Jérôme, Marseille.
The presentation and the slides will be in English.
Jury composition :
– Bruno MONTCEL, CREATIS, Univ Lyon 1 – President
– Nicolas DUCROS, CREATIS, INSA Lyon – Reviewer
– Tanja TARVAINEN, Univ. Eastern Finland, Kuopio, Finland – Reviewer
– Erving XIMENES, Univ. Autónoma de Madrid, Spain – Examiner
– Anabela DA SILVA, Institut Fresnel, CNRS, Marseille – Thesis Director
– Xavier LE GUÉVEL, IAB, CNRS, Grenoble – Thesis co-director
– Julien WOJAK, Institut Fresnel, AMU, Marseille – Thesis Advisor
– Paul DORVAL, Kaer Labs, Nantes – Invited member
– Rémi ANDRÉ, Institut Fresnel, AMU, Marseille – Invited member
Abstract : Fluorescence Diffuse Optical Tomography (FDOT) enables 3D molecular quantification in vivo, which is particularly useful in preclinical research, notably for monitoring tumor progression in murine model. However, its spatial resolution is limited by strong light scattering in tissues. The use of the second near-infrared window (NIR-II, 1000-2000 nm) allows for improved spatial resolution but raises major new challenges: strong tissue absorption that invalidates classical propagation models, coupled with complex instrumental noise. To overcome these challenges, this thesis introduces an efficient processing framework that allows the use of the Radiative Transfer Equation, validated by simulation and experimentally on ex vivo murine models. This thesis also presents a hybrid artificial intelligence model embedding the physics model to improve and automate the manual tuning required by classical models. Both of these works make NIR-II FDOT viable for translational medicine.
