Imagerie de population et IRM de diffusion pour caractériser la sclérose en plaques pour la moélle épinière humaine

Haykel Snoussi 1, 2
2 Empenn
IRISA-D5 - SIGNAUX ET IMAGES NUMÉRIQUES, ROBOTIQUE, Inria Rennes – Bretagne Atlantique , INSERM - Institut National de la Santé et de la Recherche Médicale
Abstract : Quantitative MRI has huge potential to provide intrinsic and normative value to tissue properties useful for diagnosis, prognosis and ultimately clinical trials in multiple sclerosis (MS) which is an inflammatory disorder of the central nervous system. Complementary to brain imaging, investigating how the spinal cord is damaged using quantitative imaging, and in particular diffusion MRI, becomes an acute challenge. Acquiring and processing this type of data present inherent challenges due to the susceptibility distortion, the small crosssectional area of the spine and the lack of visible anatomical landmarks to help identification of tracts or vertebral level. In this context, we propose several contributions for the processing and statistical analysis of this data. First, we propose novel geometric metrics to evaluate and compare different distortion correction methods by measuring the alignment of the reconstructed diffusion model with the apparent centerline of the spine. Second, using a cohort of MS patients and healthy controls, we study the link between diffusion measures and the presence or absence of lesion in a given vertebral level and we show that we can predict the latter with good accuracy by learning a multivariate linear classifier. Last, we show the feasibility of longitudinal study of the evolution of diffusion MRI metrics by performing a reproducibility study using a test-retest dataset and apply it to the 2 first timepoints (M0 and M12) of our cohort of MS patients.
Document type :
Theses
Complete list of metadatas

Cited literature [174 references]  Display  Hide  Download

https://tel.archives-ouvertes.fr/tel-02285896
Contributor : Abes Star <>
Submitted on : Friday, September 13, 2019 - 10:57:28 AM
Last modification on : Tuesday, September 17, 2019 - 10:41:51 AM

File

SNOUSSI_Haykel.pdf
Version validated by the jury (STAR)

Identifiers

  • HAL Id : tel-02285896, version 1

Citation

Haykel Snoussi. Imagerie de population et IRM de diffusion pour caractériser la sclérose en plaques pour la moélle épinière humaine. Medical Imaging. Université Rennes 1, 2019. English. ⟨NNT : 2019REN1S022⟩. ⟨tel-02285896⟩

Share

Metrics

Record views

132

Files downloads

39