Relapse Prediction in Multiple Sclerosis: a Supervised Learning Approach

Bibliographic Details
Main Author: Valente, Joana Filipa Barros
Publication Date: 2022
Format: Master thesis
Language: eng
Source: Repositórios Científicos de Acesso Aberto de Portugal (RCAAP)
Download full: http://hdl.handle.net/10451/56753
Summary: Tese de mestrado, Ciência de Dados, 2022, Universidade de Lisboa, Faculdade de Ciências
id RCAP_04483fc0383ed7f0b7014f3b27b66bd8
oai_identifier_str oai:repositorio.ulisboa.pt:10451/56753
network_acronym_str RCAP
network_name_str Repositórios Científicos de Acesso Aberto de Portugal (RCAAP)
repository_id_str https://opendoar.ac.uk/repository/7160
spelling Relapse Prediction in Multiple Sclerosis: a Supervised Learning ApproachEsclerose MúltiplaEsclerose Múltipla Recidivante RemitenteSurtosAprendizagem AutomáticaAnálise preditivaTeses de mestrado - 2023Departamento de InformáticaTese de mestrado, Ciência de Dados, 2022, Universidade de Lisboa, Faculdade de CiênciasMultiple sclerosis (MS) is an autoimmune disease of the central nervous system affecting, approximately, 2.8 million people around the world. Its onset may be abrupt or insidious and its evolution, clinical manifestation, and treatment response may vary widely across patients, making it difficult to predict and manage the disease outcomes. The majority of MS patients (around 85%-90%) are diagnosed with relapse-remitting MS (RRMS), a form of the disease that is characterised by alternating periods of relapses and of total or partial recovery (remission). These relapses last at least 24h, develop acutely or subacutely, are mostly monosymptomatic, and are documented to affect patients’ disability levels, at least in the early stages of the disease, and their social, working, and household activities. Despite all the existing research focused on understanding the characteristics and impact of relapses, these remain largely unpredictable both in time and location. Moreover, the fact that this disease manifestation affects most MS patients and that the anticipation of a future relapse may help clinicians to provide more timely and effective treatments to MS patients, emphasizes the need to develop models that are able to predict when a next relapse may occur. Using a dataset containing 3,679 relapses from 859 patients, this study proposes developing a prediction model for when a future relapse may occur based on the clinical profile of RRMS patients at the time of relapse. To this end, three classifiers were learnt (Logistic Regression, Decision Tree, and Random Forest), considering as target variables if a relapse happened within one or two years, and as a clinical profile the combination of demographic, symptoms, disability, treatment, and MRI data. The best-performing model was the Random Forest, with an AUC of 68%, when evaluating if a relapse may happen within two years from the previous relapse. The most relevant variables for predicting the timing of a future relapse include the time elapsed between the two most recent relapses, the number of relapses already endured by a patient, the number of different medicines the patient has tried since the onset of the disease, disease duration, the number of symptoms experienced in a relapse, and the age at onset. Overall, the findings reported in this study can help clinicians to assess the future prospects of the disease, for a given patient, and to tailor the healthcare provided accordingly.Tomás, Helena Isabel Aidos LopesMadeira, Sara Alexandra CordeiroRepositório da Universidade de LisboaValente, Joana Filipa Barros2024-11-30T01:30:41Z202320222023-01-01T00:00:00Zinfo:eu-repo/semantics/publishedVersioninfo:eu-repo/semantics/masterThesisapplication/pdfhttp://hdl.handle.net/10451/56753TID:203490290enginfo:eu-repo/semantics/openAccessreponame:Repositórios Científicos de Acesso Aberto de Portugal (RCAAP)instname:FCCN, serviços digitais da FCT – Fundação para a Ciência e a Tecnologiainstacron:RCAAP2025-03-17T14:55:22Zoai:repositorio.ulisboa.pt:10451/56753Portal AgregadorONGhttps://www.rcaap.pt/oai/openaireinfo@rcaap.ptopendoar:https://opendoar.ac.uk/repository/71602025-05-29T03:29:01.625338Repositórios Científicos de Acesso Aberto de Portugal (RCAAP) - FCCN, serviços digitais da FCT – Fundação para a Ciência e a Tecnologiafalse
dc.title.none.fl_str_mv Relapse Prediction in Multiple Sclerosis: a Supervised Learning Approach
title Relapse Prediction in Multiple Sclerosis: a Supervised Learning Approach
spellingShingle Relapse Prediction in Multiple Sclerosis: a Supervised Learning Approach
Valente, Joana Filipa Barros
Esclerose Múltipla
Esclerose Múltipla Recidivante Remitente
Surtos
Aprendizagem Automática
Análise preditiva
Teses de mestrado - 2023
Departamento de Informática
title_short Relapse Prediction in Multiple Sclerosis: a Supervised Learning Approach
title_full Relapse Prediction in Multiple Sclerosis: a Supervised Learning Approach
title_fullStr Relapse Prediction in Multiple Sclerosis: a Supervised Learning Approach
title_full_unstemmed Relapse Prediction in Multiple Sclerosis: a Supervised Learning Approach
title_sort Relapse Prediction in Multiple Sclerosis: a Supervised Learning Approach
author Valente, Joana Filipa Barros
author_facet Valente, Joana Filipa Barros
author_role author
dc.contributor.none.fl_str_mv Tomás, Helena Isabel Aidos Lopes
Madeira, Sara Alexandra Cordeiro
Repositório da Universidade de Lisboa
dc.contributor.author.fl_str_mv Valente, Joana Filipa Barros
dc.subject.por.fl_str_mv Esclerose Múltipla
Esclerose Múltipla Recidivante Remitente
Surtos
Aprendizagem Automática
Análise preditiva
Teses de mestrado - 2023
Departamento de Informática
topic Esclerose Múltipla
Esclerose Múltipla Recidivante Remitente
Surtos
Aprendizagem Automática
Análise preditiva
Teses de mestrado - 2023
Departamento de Informática
description Tese de mestrado, Ciência de Dados, 2022, Universidade de Lisboa, Faculdade de Ciências
publishDate 2022
dc.date.none.fl_str_mv 2022
2023
2023-01-01T00:00:00Z
2024-11-30T01:30:41Z
dc.type.status.fl_str_mv info:eu-repo/semantics/publishedVersion
dc.type.driver.fl_str_mv info:eu-repo/semantics/masterThesis
format masterThesis
status_str publishedVersion
dc.identifier.uri.fl_str_mv http://hdl.handle.net/10451/56753
TID:203490290
url http://hdl.handle.net/10451/56753
identifier_str_mv TID:203490290
dc.language.iso.fl_str_mv eng
language eng
dc.rights.driver.fl_str_mv info:eu-repo/semantics/openAccess
eu_rights_str_mv openAccess
dc.format.none.fl_str_mv application/pdf
dc.source.none.fl_str_mv reponame:Repositórios Científicos de Acesso Aberto de Portugal (RCAAP)
instname:FCCN, serviços digitais da FCT – Fundação para a Ciência e a Tecnologia
instacron:RCAAP
instname_str FCCN, serviços digitais da FCT – Fundação para a Ciência e a Tecnologia
instacron_str RCAAP
institution RCAAP
reponame_str Repositórios Científicos de Acesso Aberto de Portugal (RCAAP)
collection Repositórios Científicos de Acesso Aberto de Portugal (RCAAP)
repository.name.fl_str_mv Repositórios Científicos de Acesso Aberto de Portugal (RCAAP) - FCCN, serviços digitais da FCT – Fundação para a Ciência e a Tecnologia
repository.mail.fl_str_mv info@rcaap.pt
_version_ 1833601714367758336

Similar Items