Learning a Distance for the Clustering of Patients with Amyotrophic Lateral Sclerosis

2025-11-03

Symposium on Advances in Databases and Information Systems

AI Summary

ALS presents challenges including small and heterogeneous patient populations, sparse longitudinal clinical data, and lack of consensus on clinical subtyping. To address these, this paper proposes a weakly supervised distance learning framework that integrates medical prior knowledge to learn personalized distance functions tailored for disease progression—specifically, declarative functional scoring. The method jointly models multivariate longitudinal clinical biomarkers with domain-specific constraints, enabling compatibility with standard clustering algorithms. Applied to a cohort of 353 ALS patients, it yields clinically interpretable patient subgroups. Experimental results demonstrate superior discriminative performance in survival prediction (significantly improved C-index) compared to existing clustering approaches, while maintaining high clustering purity. Critically, the identified subgroups exhibit clear clinical relevance—e.g., distinct progression trajectories and symptom profiles—thereby establishing an interpretable, empirically verifiable paradigm for precision subtyping and individualized intervention in ALS.

Technology Category

• Machine Learning: Semi-Supervised Learning
• Search and Optimization: Distributed Search
• Multiagent Systems: Distributed Problem Solving

Application Category

• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• User Modeling, Personalization and Recommendation: Explainable and interpretable methods for personalization
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Amyotrophic lateral sclerosis (ALS) is a severe disease with a typical survival of 3-5 years after symptom onset. Current treatments offer only limited life extension, and the variability in patient responses highlights the need for personalized care. However, research is hindered by small, heterogeneous cohorts, sparse longitudinal data, and the lack of a clear definition for clinically meaningful patient clusters. Existing clustering methods remain limited in both scope and number. To address this, we propose a clustering approach that groups sequences using a disease progression declarative score. Our approach integrates medical expertise through multiple descriptive variables, investigating several distance measures combining such variables, both by reusing off-the-shelf distances and employing a weak-supervised learning method. We pair these distances with clustering methods and benchmark them against state-of-the-art techniques. The evaluation of our approach on a dataset of 353 ALS patients from the University Hospital of Tours, shows that our method outperforms state-of-the-art methods in survival analysis while achieving comparable silhouette scores. In addition, the learned distances enhance the relevance and interpretability of results for medical experts.

Problem

• Developing distance metrics for clustering ALS patient progression data
• Addressing heterogeneity in small patient cohorts with sparse longitudinal data
• Improving clinical relevance and interpretability of ALS patient stratification

Innovation

• Learning distance measures for patient clustering
• Combining multiple descriptive clinical variables
• Using weak-supervised learning with medical expertise