Multiple Sclerosis Severity Estimation and Progression Prediction Based on Machine Learning Techniques

Plati, Daphni; Tripoliti, Evanthia E.; Zelilidou, Styliani; Vlachos, Kostas; Konitsiotis, Spyridon; Fotiadis, Dimitrios I.

doi:10.1109/embc48229.2022.9871213

Cited by 12 publications

(4 citation statements)

References 27 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Previous studies on prediction of disability (based on EDSS) are limited by a short follow-up time 36 , few included parameters, such as CSF lymphocyte count 37 and clinical measures 36 , or using expression levels of a limited number of proteins [38][39][40] . Sufficient follow-up time is necessary since it usually requires several years for pwMS to display changes in their disability scores 41 .…”

Section: Discussionmentioning

confidence: 99%

Proteomics reveal biomarkers for diagnosis, disease activity and long-term disability outcomes in multiple sclerosis

Åkesson,

Hojjati,

Hellberg

et al. 2023

Nat Commun

View full text Add to dashboard Cite

Sensitive and reliable protein biomarkers are needed to predict disease trajectory and personalize treatment strategies for multiple sclerosis (MS). Here, we use the highly sensitive proximity-extension assay combined with next-generation sequencing (Olink Explore) to quantify 1463 proteins in cerebrospinal fluid (CSF) and plasma from 143 people with early-stage MS and 43 healthy controls. With longitudinally followed discovery and replication cohorts, we identify CSF proteins that consistently predicted both short- and long-term disease progression. Lower levels of neurofilament light chain (NfL) in CSF is superior in predicting the absence of disease activity two years after sampling (replication AUC = 0.77) compared to all other tested proteins. Importantly, we also identify a combination of 11 CSF proteins (CXCL13, LTA, FCN2, ICAM3, LY9, SLAMF7, TYMP, CHI3L1, FYB1, TNFRSF1B and NfL) that predict the severity of disability worsening according to the normalized age-related MS severity score (replication AUC = 0.90). The identification of these proteins may help elucidate pathogenetic processes and might aid decisions on treatment strategies for persons with MS.

show abstract

Section: Discussionmentioning

confidence: 99%

Proteomics reveal biomarkers for diagnosis, disease activity and long-term disability outcomes in multiple sclerosis

Åkesson,

Hojjati,

Hellberg

et al. 2023

Nat Commun

View full text Add to dashboard Cite

show abstract

“…Previous studies on prediction of disability (based on EDSS) are limited by a short follow-up time 35 , few included parameters, such as CSF lymphocyte count 36 and clinical measures 35 , or using expression levels of a limited number of proteins [37][38][39] . Sufficient follow-up time is necessary since it usually requires several years for pwMS to display changes in their disability scores 40 .…”

Section: Discussionmentioning

confidence: 99%

Proteomics profiling reveals biomarkers for predicting diagnosis, disease activity and long-term disability outcome in multiple sclerosis

Åkesson

Hojjati

Hellberg

et al. 2023

Preprint

View full text Add to dashboard Cite

To personalize multiple sclerosis (MS) treatment strategies, sensitive and reliable protein biomarkers are needed to predict disease trajectory. We used protein profiling based on the highly sensitive proximity extension assay combined with next generation sequencing (Olink Explore) to measure 1463 proteins in cerebrospinal fluid (CSF) and plasma from 143 people with early-stage MS and 43 healthy controls. With longitudinally followed discovery and replication cohorts we identified CSF proteins that consistently predicted both short- and long-term disease progression. Neurofilament light chain (NfL) in CSF was confirmed as superior for predicting the absence of disease activity 2 years after sampling (replication AUC = 0.77). Importantly, we also identified a combination of 11 CSF proteins (CXCL13, LTA, FCN2, ICAM3, LY9, SLAMF7, TYMP, CHI3L1, FYB1, TNFRSF1B, NfL) that could predict the severity of disability worsening according to the normalized age-related MS severity score (replication AUC = 0.90). Identified proteins elucidate pathogenetic processes and can aid decisions of treatment strategies for people with MS.

show abstract

“…In a recent study by Plati et al [27], the authors addressed the MS severity estimation problem and predicted disease progression using Machine Learning (ML) techniques. Utilizing data from the ProMiSi project, encompassing demographic details, clinical data, test results, treatment, and comorbidities of 30 patients recorded at three time points, the ML methods achieved notable ac-curacy, with 94.87% for MS severity estimation and 83.33% for disease progression prediction.…”

Section: Literature Reviewmentioning

confidence: 99%

Machine Learning Algorithms for Early Prediction of Multiple Sclerosis Progression: A Comparative Study

Haouam,

Benmalek

2024

AAIML

View full text Add to dashboard Cite

Multiple Sclerosis (MS) is a chronic autoimmune disease characterized by central nervous system (CNS) degeneration, leading to diverse neurological symptoms. Managing MS poses a challenge due to its unpredictable progression. This study focuses on early prediction of MS progression using machine learning (ML) algorithms, comparing the effectiveness of Random Forest, XGBoost, Decision Tree, and Logistic Regression. Clinical, genetic, and environmental factors were analyzed in a cohort of Mexican mestizo patients recently diagnosed with Clinically Isolated Syndrome (CIS). Data preprocessing addressed missing values, and feature selection tailored to the population’s characteristics was applied. The dataset was split into training and testing sets, maintaining stratification for CDMS and nonCDMS cases. Machine learning models were trained with optimized hyperparameters. Performance evaluation metrics, including accuracy, precision, recall, F1-score, and AUC-ROC, were employed. The Random Forest model exhibited superior performance (AUC: 0.93, accuracy: 87%), outperforming other models. Variable importance analysis identified top predictors, including Periventricular MRI, Age, Infratentorial MRI, and Oligoclonal Bands. The study’s clinical implications highlight ML’s potential in enhancing early MS prognosis, aiding timely interventions. The Random Forest model, with its robust performance, emerges as a valuable tool for identifying patients at risk. While the study contributes to predictive analytics in neurological disorders, limitations include cohort specificity and retrospective data use. Prospective studies and further exploration of data sources are recommended for broader applicability. This research demonstrates the efficacy of ML in early MS progression prediction, providing clinicians with a promising tool for personalized patient care. The findings contribute to advancing predictive healthcare analytics and emphasize the significance of tailored interventions in neurological disorders.

show abstract

Multiple Sclerosis Severity Estimation and Progression Prediction Based on Machine Learning Techniques

Cited by 12 publications

References 27 publications

Proteomics reveal biomarkers for diagnosis, disease activity and long-term disability outcomes in multiple sclerosis

Proteomics reveal biomarkers for diagnosis, disease activity and long-term disability outcomes in multiple sclerosis

Proteomics profiling reveals biomarkers for predicting diagnosis, disease activity and long-term disability outcome in multiple sclerosis

Machine Learning Algorithms for Early Prediction of Multiple Sclerosis Progression: A Comparative Study

Contact Info

Product

Resources

About