UCL Discovery
UCL home » Library Services » Electronic resources » UCL Discovery

Quantitative MRI and machine learning for the diagnosis and prognosis of Multiple Sclerosis

Ricciardi, Antonio; (2021) Quantitative MRI and machine learning for the diagnosis and prognosis of Multiple Sclerosis. Doctoral thesis (Ph.D), UCL (University College London). Green open access

[thumbnail of ARicciardi_PhD_thesis.pdf]
Preview
Text
ARicciardi_PhD_thesis.pdf

Download (19MB) | Preview

Abstract

Multiple sclerosis (MS) is an immune-mediated, inflammatory, neurological disease affecting myelin in the central nervous system, whose driving mechanisms are not yet fully understood. Conventional magnetic resonance imaging (MRI) is largely used in the MS diagnostic process, but because of its lack of specificity, it cannot reliably detect microscopic damage. Quantitative MRI provides instead feature maps that can be exploited to improve prognosis and treatment monitoring, at the cost of prolonged acquisition times and specialised MR-protocols. In this study, two converging approaches were followed to investigate how to best use the available MRI data for the diagnosis and prognosis of MS. On one hand, qualitative data commonly used in clinical research for lesion and anatomical purposes were shown to carry quantitative information that could be used to conduct myelin and relaxometry analyses on cohorts devoid of dedicated quantitative acquisitions. In this study arm, named bottom-up, qualitative information was up-converted to quantitative surrogate: traditional model-fitting and deep-learning frameworks were proposed and tested on MS patients to extract relaxometry and indirect-myelin quantitative data from qualitative scans. On the other hand, when using multi-modal MRI data to classify MS patients with different clinical status, different MR-features contribute to specific classification tasks. The top-down study arm consisted in using machine learning to reduce the multi-modal dataset dimensionality only to those MR-features that are more likely to be biophysically meaningful with respect to each MS phenotype pathophysiology. Results show that there is much more potential to qualitative data than lesion and tissue segmentation, and that specific MRI modalities might be better suited for investigating individual MS phenotypes. Efficient multi-modal acquisitions informed by biophysical findings, whilst being able to extract quantitative information from qualitative data, would provide huge statistical power through the use of large, historical datasets, as well as constitute a significant step forward in the direction of sustainable research.

Type: Thesis (Doctoral)
Qualification: Ph.D
Title: Quantitative MRI and machine learning for the diagnosis and prognosis of Multiple Sclerosis
Event: UCL (University College London)
Open access status: An open access version is available from UCL Discovery
Language: English
Additional information: Copyright © The Author 2021. Original content in this thesis is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International (CC BY-NC 4.0) Licence (https://creativecommons.org/licenses/by-nc/4.0/). Any third-party copyright material present remains the property of its respective owner(s) and is licensed under its existing terms. Access may initially be restricted at the author’s request.
UCL classification: UCL
UCL > Provost and Vice Provost Offices > School of Life and Medical Sciences
UCL > Provost and Vice Provost Offices > School of Life and Medical Sciences > Faculty of Brain Sciences
UCL > Provost and Vice Provost Offices > School of Life and Medical Sciences > Faculty of Brain Sciences > UCL Queen Square Institute of Neurology
UCL > Provost and Vice Provost Offices > School of Life and Medical Sciences > Faculty of Brain Sciences > UCL Queen Square Institute of Neurology > Neuroinflammation
URI: https://discovery.ucl.ac.uk/id/eprint/10136247
Downloads since deposit
186Downloads
Download activity - last month
Download activity - last 12 months
Downloads by country - last 12 months

Archive Staff Only

View Item View Item