Describing and characterising variability in ALS disease progression.

Area of Science:

• Neurology
• Biostatistics
• Machine Learning in Medicine

Background:

• The revised Amyotrophic Lateral Sclerosis Functional Rating Scale (ALSFRS-R) score is the standard for measuring ALS progression.
• However, the ALSFRS-R score does not fully capture the diverse nature of ALS.
• A novel measure of variability in ALSFRS-R scores is introduced for better disease characterization.

Purpose of the Study:

• To introduce and validate a new measure of variability in ALSFRS-R scores.
• To explore the relationship between ALSFRS-R score variability and disease span.
• To develop a machine learning model for predicting ALS variability.

Main Methods:

• Utilized data from 5030 ALS clinical trial patients from the Pooled Resource Open-Access ALS Clinical Trials database.
• Calculated disease progression variability using a novel measure and correlated it with disease span.
• Developed and validated a machine learning model using clinical, laboratory, and demographic data to predict variability class.

Main Results:

• Increased variability in ALSFRS-R scores correlated with a longer disease span.
• The machine learning model achieved 60.1-72.7% accuracy in predicting variability class.
• Identified key clinical, laboratory, and demographic predictors of ALS variability.

Conclusions:

• ALSFRS-R score variability is a significant factor in disease span, particularly for rapidly progressing ALS.
• Identified predictors may offer insights into the pathophysiology of ALS variability.
• Future clinical trials could aim to increase variability in fast-progressing ALS patients.