A Consensus Clustering Approach to Amyotrophic Lateral Sclerosis Phenotyping
Pilar M Ferraro1, Sara Narteni2, Marta Lenatti2
1IRCCS Ospedale Policlinico San Martino, 16152, Genoa, Italy.
Studies in Health Technology and Informatics
|May 23, 2026
Summary
This study introduces a data-driven approach to characterize Amyotrophic Lateral Sclerosis (ALS) phenotypes using clinical data. Findings reveal distinct patient clusters, aiding in a better understanding of ALS heterogeneity.
Area of Science:
- Neurology
- Data Science
- Biostatistics
Background:
- Amyotrophic Lateral Sclerosis (ALS) presents significant phenotyping challenges due to its heterogeneity and rarity.
- Accurate phenotyping is crucial for understanding disease progression and developing targeted therapies.
Purpose of the Study:
- To develop and validate a data-driven methodology for characterizing ALS phenotypes.
- To investigate the alignment of data-derived patient clusters with established clinical phenotypes and medical knowledge.
Main Methods:
- A consensus clustering approach was employed, analyzing clinical data from a comprehensive battery of examinations.
- The method involved creating stable clusters across multiple random data sub-samples to ensure robustness.
Main Results:
- Consistent patient profiles were identified for bulbar onset ALS, primarily defined by onset characteristics.
- Spinal onset ALS patients demonstrated greater heterogeneity within their identified phenotype clusters.
Conclusions:
- The proposed data-driven approach effectively supports ALS phenotyping by identifying distinct patient subgroups.
- Results highlight differential heterogeneity between bulbar and spinal onset ALS, offering insights for future research and clinical stratification.
