Welcome to the UPF Digital Repository

A novel approach to assess Alzheimer’s Disease by using radiomics in TgF344-AD rats: Development and application of radiomic features as imaging biomarkers to evaluate the different stages of Alzheimer’s disease

Show simple item record

dc.contributor.author Sorribes Torrent, Elisabet
dc.date.accessioned 2021-07-20T09:03:56Z
dc.date.available 2021-07-20T09:03:56Z
dc.date.issued 2021-07
dc.identifier.uri http://hdl.handle.net/10230/48250
dc.description Tutors: Emma Muñoz Moreno, Gemma Piella Fenoy
dc.description.abstract Alzheimer’s disease (AD) is a complex neurodegenerative disease that leads to progressive cognitive decline and memory loss. Nowadays, it is still very challenging to diagnose AD due to its complexity, the inter-individual differences and the late symptoms appearance. However, the use of animal models can contribute to improve early AD diagnosis in patients. Concretely, the combination of MRI images of TgF344-AD rats and radiomics could have a great potential in detecting quantitative imaging biomarkers of AD. The main objectives of this project was to evaluate the ability of radiomics to discriminate between control (CTR) and transgenic animals (Tg) and also to evaluate the disease progression considering the stages young-CTR, old-CTR, young-Tg, old-Tg in three different regions, which were hippocampus, amygdala and thalamus. Moreover, age feature was included in order to analyse if it contributed to a better discrimination between groups and stages. To accomplish this goal, radiomics and machine learning pipelines were developed in a Python implementation. The database was composed by 34 wildtype rats and 30 TgF344-AD animals, each of them evaluated at different ages between 3 and 18 months, resulting in 98 control images and 96 in Tg. Radiomics was applied to this database in order to obtain the corresponding texture, shape and intensity features for each region. Then, several machine learning algorithms were used to test the biomarker combination obtained from the feature selection. From here, it was possible to find the best model that optimized the accuracy of the classification, and therefore, to obtain highly accurate non-invasive biomarkers of the disease. In the binary classification of rat images between control and Tg animals, the highest accuracy reached in our method was 78.84% using hippocampus. Regarding the classification in the aforementioned 4 groups, an accuracy value of 76.92% was obtained in the hippocampus too. Radiomics showed promising results when applied in transgenic TgF334-AD rats. Both CTR and Tg as well as the different stages Young-CTR, Old-CTR, Young-Tg and Old-Tg were properly identified in hippocampus. However, age was considered as indicator of disease stage and a more accurate individual assessment could result in better classification.
dc.format.mimetype application/pdf
dc.language.iso eng
dc.rights © Tots els drets reservats
dc.title A novel approach to assess Alzheimer’s Disease by using radiomics in TgF344-AD rats: Development and application of radiomic features as imaging biomarkers to evaluate the different stages of Alzheimer’s disease
dc.type info:eu-repo/semantics/bachelorThesis
dc.subject.keyword Alzheimer’s Disease (AD)
dc.subject.keyword Machine learning
dc.subject.keyword Magnetic Resonance Imaging (MRI)
dc.subject.keyword Biomarkers
dc.subject.keyword Radiomic features
dc.rights.accessRights info:eu-repo/semantics/openAccess

Thumbnail

This item appears in the following Collection(s)

Show simple item record

Search DSpace


Advanced Search

Browse

My Account

Statistics

Compliant to Partaking