eprintid: 10202948 rev_number: 14 eprint_status: archive userid: 699 dir: disk0/10/20/29/48 datestamp: 2025-03-28 13:57:04 lastmod: 2025-03-28 13:57:04 status_changed: 2025-03-28 13:57:04 type: thesis metadata_visibility: show sword_depositor: 699 creators_name: Ripart, Mathilde title: Automated Detection of MRI Pathology in Epilepsy ispublished: unpub divisions: UCL divisions: B02 divisions: D13 divisions: G26 keywords: MRI, Machine-learning, Epilepsy note: Copyright © The Author 2025. 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. abstract: Epilepsy affects one in 100 people, with one third of patients experiencing seizures that are not adequately controlled by medication. For patients whose seizures are caused by a structural brain lesion - such as focal cortical dysplasia (FCD) or hippocampal sclerosis (HS) - surgical removal of the lesion may offer a path to seizure freedom. A key step in the presurgical planning is localising these lesions on MRI scans. However, some lesions are subtle and evade visual detection, leading to delays, costly and invasive additional testing, and lower rates of post-surgical seizure freedom. Computational tools have the potential to aid the detection of epilepsy lesions on MRI scans. This thesis describes the development of automated AI-based tools to detect subtle epilepsy lesions on MRI scans, to improve presurgical planning and ultimately ease the burden on patients and families. The thesis aimed to: 1. Develop automated AI tools to help detect subtle epilepsy lesions on MRI scans 2. Tailor these tools for use in clinical settings, to offer a radiological adjunct for the presurgical planning of patients with drug-resistant focal epilepsy. 3. Release these tools and code as open-source resources, to facilitate widespread adoption and accelerate future research. The thesis introduces three distinct automated tools: - MELD FCD: a machine-learning model for the detection of focal cortical dysplasia (Chapters 2 & 3) - AID-HS: a machine-learning model for the automated and interpretable detection (AID) of hippocampal sclerosis (Chapter 4) - MELD FE: a deep-learning model for the detection of a wide range of lesions in focal epilepsy (FE) (Chapters 5 & 6). This thesis also details the collation of large neuroimaging datasets of patients with focal epilepsy, which were used for the training and evaluation of these AI models. Finally, the thesis showcases how these tools were used in research practice around the world and highlights their potential to improve presurgical planning of patients with drug-resistant epilepsy and subtle MRI lesions. The figure below offers a visual abstract of the thesis. date: 2025-01-28 date_type: published full_text_type: other thesis_class: doctoral_embargoed thesis_award: Ph.D language: eng verified: verified_manual elements_id: 2346767 lyricists_name: Ripart, Mathilde lyricists_id: MRIPA96 actors_name: Ripart, Mathilde actors_id: MRIPA96 actors_role: owner full_text_status: restricted pagerange: 2-227 pages: 227 institution: UCL (University College London) department: Developmental Neurosciences Dept thesis_type: Doctoral citation: Ripart, Mathilde; (2025) Automated Detection of MRI Pathology in Epilepsy. Doctoral thesis (Ph.D), UCL (University College London). document_url: https://discovery.ucl.ac.uk/id/eprint/10202948/1/Ripart_10202948_thesis.pdf