Project

Back to overview

Non-invasive histology of the brain microstructure in vivo using advanced modeling techniques and multi-contrast MRI data

Applicant Canales-Rodríguez Erick Jorge
Number 185814
Funding scheme Ambizione
Research institution Laboratoire de traitement des signaux 5 EPFL - STI - IEL - LTS5
Institution of higher education EPF Lausanne - EPFL
Main discipline Other disciplines of Engineering Sciences
Start/End 01.08.2020 - 31.07.2024
Approved amount 600'664.00
Show all

All Disciplines (2)

Discipline
Other disciplines of Engineering Sciences
Biomedical Engineering

Keywords (6)

Quantitative models; Microstructure; White matter; T2 relaxometry; Magnetic Resonance imaging (MRI); Diffusion MRI

Lay Summary (Italian)

Lead
Istologia non invasiva della microstruttura cerebrale in vivo mediante tecniche di modellazione avanzate e dati RMI multi-contrasto
Lay summary

La risonanza magnetica per immagini (RMI) è una tecnica non invasiva che produce immagini dettagliate del tessuto nervoso cerebrale. Consente l'acquisizione di immagini risultanti da diversi meccanismi di contrasto, tra cui la diffusione dell'acqua ed i tempi di rilassamento T1 e T2 che forniscono informazioni sull'anatomia cerebrale. L'obiettivo principale di questo progetto è lo sviluppo di avanzati modelli biofisici di risonanza magnetica e di nuove tecniche di stima parametrica per ottenere descrittori quantitativi più accurati della microstruttura del tessuto cerebrale. La novità del progetto è legata all'utilizzo ed alla combinazione di due diverse modalità RMI: RMI a diffusione e multi-eco T2 (MET2). Quest'idea è motivata dalla natura complementare delle informazioni disponibili con queste tecniche, che possono aiutare a superare le attuali limitazioni. La proposta è divisa in tre sottoprogetti, separati ma correlati, ognuno incentrato su una diversa tecnica di imaging. Il primo sottoprogetto si occupa dello sviluppo di un nuovo modello di risonanza magnetica per spiegare la dipendenza nel tempo del processo di diffusione quando tale processo avviene nello spazio extra-assonale della sostanza bianca cerebrale. Lo studio sarà esteso allo spazio intra-assonale, ottenendo un modello di diffusione più accurato per studiare la materia bianca. Il secondo sottoprogetto mira a sviluppare un nuovo algoritmo di ricostruzione per stimare la distribuzione dei tempi di rilassamento T2 dai dati MET2. Consentirà una stima accurata della frazione d'acqua della mielina, un importante biomarcatore per varie patologie cerebrali. Infine, nel terzo sottoprogetto verrà sviluppato un nuovo quadro teorico che combina l'analisi di entrambe le modalità. La tecnica proposta verrà applicata ai dati provenienti da uno scanner MRI CONNECTOM all'avanguardia. Il principale contributo di questo progetto sarà una stima affidabile e congiunta dei parametri rilevanti della microstruttura cerebrale, comprese le frazioni volumetriche dei compartimenti intra-assonale, extra-assonale e mielinico, nonché il rapporto g, le diffusività principali ed i tempi di rilassamento T2. Il potenziale di questa metodologia per caratterizzare le lesioni cerebrali verrà dimostrato utilizzando i dati acquisiti dalla stessa coorte di soggetti reclutati nel secondo sottoprogetto. Le caratteristiche del tessuto nervoso rivelate tramite queste tecniche hanno il potenziale di essere specifiche e sensibili ad alcune patologie cerebrali come sclerosi multipla, ictus, morbo di Alzheimer e disturbi mentali.
Direct link to Lay Summary Last update: 17.04.2020

Lay Summary (English)

Lead
Non-invasive histology of the brain microstructure in vivo using advanced modeling techniques and multi-contrast MRI data
Lay summary
Magnetic resonance imaging (MRI) is a non-invasive technique that produces detailed internal images of the brain nervous tissue. It enables the creation of various image contrasts, including water diffusion, and T1 and T2 relaxation times, which provide information about the brain's anatomy. The primary aim of this project is the development of advanced MRI biophysical models and novel model parameter estimation techniques to obtain more accurate quantitative descriptors of the brain tissue microstructure. The novelty of the project is related to using and combining two different MRI modalities: diffusion MRI and multi-echo T2 (MET2). This idea is motivated by the complementary nature of the information available with these techniques, which can help to overcome the current limitations. The proposal is divided into three separate but interrelated sub-projects, each one focused on a different imaging technique. The first sub-project addresses the development of a novel diffusion MRI model to explain the time-dependent diffusion process in the extra-axonal space of the brain's white matter. It will be extended to include the intra-axonal space, and will thus provide a more accurate diffusion model to study the white matter. Both models will be validated using a realistic Monte Carlo simulator and real data acquired from tissue-mimetic phantoms. The second sub-project aims at developing a new reconstruction algorithm for estimating the distribution of T2 relaxation times from MET2 data. It will enable an accurate estimation of the myelin water fraction, an important biomarker for various brain pathologies. To translate the resulting methodologies into a clinical environment, real MET2 datasets from a population of healthy subjects and patients with multiple sclerosis will be collected and their myelin water fraction maps compared. Finally, a new theoretical framework combining the analysis of both modalities will be developed in the third sub-project. The proposed technique will be applied to data coming from a cutting-edge CONNECTOM MRI scanner. The major contribution of this project will a reliable, joint estimation of relevant brain microstructure parameters, including the volume fractions of the intra-axonal, extra-axonal and myelin compartments, as well as the g-ratio, main diffusivities, and T2 relaxation times. The potential of this methodology to characterize brain lesions will be demonstrated using data acquired from the same cohort of subjects recruited in the second sub-project. The nervous tissue features revealed by these techniques have the potential to be specific and sensitive to some brain pathologies such as multiple sclerosis, stroke, Alzheimer's disease, and mental disorders. Future studies could use these features as non-invasive biomarkers to monitor illness progression, to assess the efficacy of new treatments, to design clinical trials, and to answer fundamental neuroscientific questions.
Direct link to Lay Summary Last update: 17.04.2020

Responsible applicant and co-applicants

Employees

Abstract

Magnetic resonance imaging (MRI) is a non-invasive technique that produces detailed internal images of the brain nervous tissue. It enables the creation of various image contrasts, including water diffusion, and T1 and T2 relaxation times, which provide information about the brain's anatomy. Diffusion MRI (dMRI) is the main technique for studying the spatial organization of the brain's white matter in vivo. A number of relevant models for dMRI have been proposed in the last years to quantify the microstructure of white matter tissue. Despite all the previous work, many open problems still remain. For instance, the diffusion process in the extra-axonal space is time-dependent, and except for a few limiting cases, there is no general, analytical solution to explain the measurements. Moreover, the dMRI signal is 'weighted' by the unknown T2 values of each microstructure compartment (e.g., myelin water, intra- and extra-axonal spaces). This confounding factor is further enhanced by dMRI' s lack of sensitivity to myelin due to the very short T2 of the myelin water. In order to overcome these difficulties, it is crucial to develop new dMRI models, which should be complemented with T2 measurement techniques.The primary aim of this project is the development of advanced MRI biophysical models and novel model parameter estimation techniques to obtain more accurate quantitative descriptors of the brain tissue microstructure. The novelty of the project is related to using and combining two different MRI modalities: dMRI and multi-echo T2 (MET2). This idea is motivated by the complementary nature of the information available with these techniques, which can help to overcome the current limitations.The proposal is divided into three separate but interrelated sub-projects, each one focused on a different imaging technique. The first sub-project addresses the development of a novel dMRI model to explain the time-dependent diffusion process in the extra-axonal space of the brain's white matter. It will be extended to include the intra-axonal space, and will thus provide a more accurate diffusion model to study the white matter. Both models will be validated using a realistic Monte Carlo dMRI simulator (in collaboration with the host group LTS5, EPFL, Lausanne) and real dMRI data acquired from tissue-mimetic phantoms (in collaboration with DRCMR, Denmark). The second sub-project aims at developing a new reconstruction algorithm for estimating the distribution of T2 relaxation times from MET2 data. It will enable an accurate estimation of the myelin water fraction, an important biomarker for various brain pathologies such as multiple sclerosis. Moreover, the relationship between the morphological features of the biomimetic phantoms and their T2 values will be investigated. In order to translate the resulting methodologies into a clinical environment, real MET2 datasets from a population of healthy subjects and patients with multiple sclerosis will be collected and their myelin water fraction maps compared (in collaboration with the University of Basel and Siemens Healthcare). Finally, after overcoming some of the current limitations of dMRI and MET2, a new theoretical framework combining the analysis of both modalities will be developed in the third sub-project. The proposed technique will be applied to data coming from a cutting-edge CONNECTOM MRI scanner (in collaboration with CUBRIC, UK). The major contribution of this project will a reliable, joint estimation of relevant microstructure parameters, including the volume fractions of the intra-axonal, extra-axonal and myelin compartments, as well as the g-ratio, main diffusivities, and T2 relaxation times. The latter will be also linked to other morphological features of the sample. The potential of this methodology to characterize brain lesions will be demonstrated using data acquired from the same cohort of subjects recruited in the second sub-project.This proposal builds on the PI's previous work on modeling in medical imaging and extends it to encompass an interdisciplinary combination of biophysical modeling, radiology, and diagnostic imaging. The nervous tissue features revealed by these techniques have the potential to be specific and sensitive to a number of brain pathologies such as multiple sclerosis, stroke, Alzheimer's disease, and mental disorders. Future studies could use these features as non-invasive biomarkers to monitor illness progression, to assess the efficacy of new treatments, to design clinical trials, and to answer fundamental neuroscientific questions.
-