Predicting Amyotrophic Lateral Sclerosis (ALS) Progression: A 7-Protein Machine Learning Panel

Overview

Amyotrophic lateral sclerosis (ALS), a progressive neurodegenerative disease that attacks the motor neurons responsible for controlling voluntary muscles, displays marked clinical heterogeneity, complicating the prediction of functional decline and limiting effective patient stratification in clinical trials. This study asked whether baseline cerebrospinal fluid (CSF) protein expression patterns could predict ALS progression rates and distinguish fast from slow-progressing phenotypes. We measured 793 proteins in CSF samples from all 43 ALS patients and 25 healthy controls, integrating protein expression levels with clinical measures of disease progression quantified by the Clinical Progression Rate (monthly decline in ALS Functional Rating Scale–Revised score from symptom onset). Candidate proteins were identified using a dual-filter strategy requiring statistical significance by Mann–Whitney U testing and a large effect size, with associations assessed using Spearman's rank correlation. Proteins correlated with age or sex in controls were excluded, and predictive performance was evaluated using Random Forest classification with stratified five-fold cross-validation. Among the 793 proteins measured, 22 differed significantly between fast (n=10) and slow (n=22) progressors, with effect sizes of 0.82–1.08; four showed strong correlations with clinical progression rate; and a seven-protein panel achieved 81.2% cross-validated accuracy with an area under the curve of 0.895 ± 0.090. Fast progressors exhibited increased inflammatory markers and reduced structural proteins, demonstrating that baseline CSF proteomic signatures can effectively predict ALS progression rates. These findings support improved patient stratification in clinical trials and warrant validation in larger independent cohorts.

Competition history

• CSEF 2026 Medicine & Physiology (Track 2) (Senior Division) · Entry S-21-10

Related projects

Closest projects by meaning, across every fair and year in the corpus.

Browse more like this

• All Biomedical and Health Sciences projects
• CSEF Biomedical and Health Sciences projects
• Every CSEF 2026 project