Annonce

Les commentaires sont clos.

Signal and Image Processing Engineer

16 Janvier 2023


Catégorie : Post-doctorant


Context

Radiotherapy (RT) is a main treatment of glioblastoma (GB), which metabolic heterogeneity can be depicted by MR Spectroscopic Imaging (MRSI). Indeed, MRSI is able to characterize abnormal metabolic areas, which are predictive of GB relapse site. Moreover, these areas are interestingly enriched in GB stem-like cells (GSC), which present a higher bioenergetic state and a differential metabolic response after RT. This suggests that specific metabolic adaptations after RT can occur in these GSC leading to radioresistance and recurrence processes.

The aim of this project is to decipher the metabolic shift after RT in these GSC and to validate this shift in clinic through multimodal data. Metabolites will be studied and quantified in blood samples and MRSI spectra acquired longitudinally from patients included in a clinical trial:

- Advanced and new signal processing methodologies will contribute to detect subtle and possibly new metabolites unrevealed and unstudied so far in the metabolic shift. Denoising methods, combined with model-based methods (blind source separation techniques such as Bayesian models with adapted prior distributions and Bernouilli-Laplace prior with two-level sparsity in the target signal) and learning-based methods (with both handcrafted features and Deep Learning) will be investigated.

- Multimodal relapse profiles (from biologic and imaging data) will be studied by classification methods and features combinations will be investigated to describe the different relapse patterns. Deep Learning techniques will be investigated to extract implicit features (recurrent and convolutional networks will be considered). To overcome difficulties encountered with low sample size, generative models, data augmentation and self-learning techniques will be investigated.

 

Context

Radiotherapy (RT) is a main treatment of glioblastoma (GB), which metabolic heterogeneity can be depicted by MR Spectroscopic Imaging (MRSI). Indeed, MRSI is able to characterize abnormal metabolic areas, which are predictive of GB relapse site. Moreover, these areas are interestingly enriched in GB stem-like cells (GSC), which present a higher bioenergetic state and a differential metabolic response after RT. This suggests that specific metabolic adaptations after RT can occur in these GSC leading to radioresistance and recurrence processes.

 

Objectives

The aim of this project is to decipher the metabolic shift after RT in these GSC and to validate this shift in clinic through multimodal data. Metabolites will be studied and quantified in blood samples and MRSI spectra acquired longitudinally from patients included in a clinical trial:

- Advanced and new signal processing methodologies will contribute to detect subtle and possibly new metabolites unrevealed and unstudied so far in the metabolic shift. Denoising methods, combined with model-based methods (blind source separation techniques such as Bayesian models with adapted prior distributions and Bernouilli-Laplace prior with two-level sparsity in the target signal) and learning-based methods (with both handcrafted features and Deep Learning) will be investigated.

- Multimodal relapse profiles (from biologic and imaging data) will be studied by classification methods and features combinations will be investigated to describe the different relapse patterns. Deep Learning techniques will be investigated to extract implicit features (recurrent and convolutional networks will be considered). To overcome difficulties encountered with low sample size, generative models, data augmentation and self-learning techniques will be investigated.

This translational project will reveal and validate new radioresistance mechanisms involving RT-induced GSC metabolism shift.

 

Skills

Minimum MSc degree or equivalent diploma in Physics, Mathematics or Informatics and a strong expertise in signal processing and image analysis, Python and Matlab programming skills are key prerequisites. Knowledge in Magnetic Resonance Imaging of the brain and its physiopathology will also be strongly valued. The candidate should be fluent in English (and preferably in French, which will be the working language), have a good communication skills and organizational skills. The candidate is expected to be highly motivated and to work independently with a strong work ethic. Applications including a cover letter, a detailed CV including a publication list if relevant and recommendation letters of up to three referees should be sent.

 

Location

Institut de Recherche en Informatique de Toulouse (IRIT) – Team MINDS

Université Paul Sabatier Toulouse, France

118 route de Narbonne 31062, Toulouse Cedex 9

And

Centre de Recherche en Cancérologie de Toulouse (CRCT)- Team RADOPT

2 Avenue Hubert Curien, 31100 Toulouse

 

Supervision

CHAARI Lotfi, PhD-HDR

lotfi.chaari@toulouse-inp.fr

Institut de Recherche en Informatique de Toulouse (IRIT) – Equipe MINDS –ENSEEIHT (UMR 5505)

Toulouse INP

 

KEN Soleakhena, PhD-HDR

ken.soleakhena@iuct-oncopole.fr

Centre de Recherche en Cancérologie de Toulouse (CRCT)- Equipe RADOPT

Institut Universitaire du Cancer de Toulouse – Oncopole (IUCT-O)

 

Complementary information

Key words: Glioblastoma, Magnetic Resonance Imaging, MR Spectroscopy, Metabolic Shift, Stem Cells Radioresistance, Multimodal relapse profile

Contract: 24 months, preferred starting date beginning of July 2023

Salary Depending of experiences and according to CNRS grid

Funds Fondation ARC; project MSrGB: Metabolic Shift in radioresistance of GlioBlastoma

Contact lotfi.chaari@toulouse-inp.fr and ken.soleakhena@iuct-oncopole.fr