Related Experiment Video
Updated: Jul 31, 2025
Measurements of Motor Function and Other Clinical Outcome Parameters in Ambulant Children with Duchenne Muscular Dystrophy
Published on: January 12, 2019
A machine learning analysis to evaluate the outcome measures in inflammatory myopathies
Maria Giovanna Danieli1, Alberto Paladini2, Eleonora Longhi3
1SOS Immunologia delle Malattie Rare e dei Trapianti, AOU delle Marche & Dipartimento di Scienze Cliniche e Molecolari, Università Politecnica delle Marche, via Tronto 10/A, 60126 Torrette di Ancona, Italy; Postgraduate School of Allergy and Clinical Immunology, Università Politecnica delle Marche, via Tronto 10/A, 60126 Ancona, Italy.
Artificial intelligence (AI) identified key predictors of long-term outcomes in Idiopathic Inflammatory Myopathies (IIM). Machine learning models accurately forecast muscle strength and disease activity, aiding future treatment strategies for IIM patients.
Area of Science:
- Rheumatology and Immunology
- Computational Biology and Bioinformatics
- Artificial Intelligence in Medicine
Background:
- Idiopathic Inflammatory Myopathies (IIM) are rare systemic autoimmune diseases.
- IIM involve multisystemic organ damage beyond the musculoskeletal system.
- Machine Learning (ML) excels at analyzing complex datasets for pattern recognition and prediction.
Purpose of the Study:
- To evaluate long-term outcomes in Idiopathic Inflammatory Myopathies (IIM) using Artificial Intelligence (AI).
- To identify key predictors of disease activity and damage in IIM patients.
- To explore the utility of ML in assessing disease progression and treatment response.
Main Methods:
- Analysis of 103 IIM patients diagnosed using 2017 EULAR/ACR criteria.
- Inclusion of clinical data: organ involvement, treatments, creatine kinase, MMT8, MITAX, HAQ-DI, MDI, PGA.
- Application of supervised ML algorithms (lasso, ridge, elastic net, CART, random forest, SVM) for predictive modeling.
Main Results:
- CART regression tree algorithm accurately predicted muscle strength (MMT8) at follow-up.
- Clinical features like RP-ILD and skin involvement predicted disease activity (MITAX).
- ML demonstrated good predictive capacity for damage scores (MDI, HAQ-DI).
Conclusions:
- AI algorithms effectively identified crucial factors correlating with IIM disease outcomes.
- ML offers potential for refining disease activity and damage assessments in IIM.
- Future ML applications may aid in validating and implementing new classification and outcome criteria for IIM.
