Netramark finds AI defined responder groups in A4 trial

Executive Summary

• Netramark Holdings Inc. presented new findings at the AD/PD 2026 conference demonstrating how its proprietary NetraAI platform identified clinically meaningful responder subgroups in the landmark A4 Alzheimer's trial.
• Despite the original phase 3 trial showing no overall statistically significant benefit for solanezumab, NetraAI's explainable machine learning approach revealed two distinct patient subgroups with large treatment effects (Cohen's d up to 1.52).
• The findings suggest that preserved neural reserve (higher regional brain volume and stronger baseline cognitive performance) may be a key determinant of anti-amyloid treatment response, highlighting AI's potential to improve precision trial design and reduce development costs.

Key Details

• AI Platform & Methodology: NetraAI uses a dynamical-systems-based explainable machine learning approach to analyze multimodal baseline variables including imaging, cognitive assessments, demographics, and biomarkers.
• Trial Context: Re-analysis of the phase 3 Anti-Amyloid Treatment in Asymptomatic Alzheimer's Disease (A4) trial, which originally showed no statistically significant overall benefit for solanezumab.
• Key Findings: Identified two biologically interpretable responder subgroups characterized by higher regional brain volume and stronger baseline cognitive performance.
• Treatment Effects: Large treatment effects within subgroups, with effect sizes reaching Cohen's d up to 1.52.
• Biomarkers/Variables: Treatment-responsive participants showed greater baseline limbic and temporal network integrity, including higher right amygdala or right superior temporal cortex volume, alongside stronger psychomotor speed and attention scores on the Digit Symbol Substitution Test.
• Industry Implications: Potential to improve trial design via patient enrichment, enable retrospective reanalysis of historical trials, and reduce development risk/cost through data-driven patient stratification.
• Conference Presentation: Poster titled "Decoding Heterogeneity in A4: Explainable ML Identifies Solanezumab-Responsive Subgroups in Preclinical AD" presented at the AD/PD 2026 International Conference in Copenhagen, Denmark (March 17–21, 2026).
• Technology Differentiation: NetraAI uses focus mechanisms to separate datasets into explainable and unexplainable subsets, preventing overfitting and improving clinical trial success likelihood compared to traditional AI methods.

Notable Quotes

• "These findings suggest that patient heterogeneity may be masking treatment effects in Alzheimer's trials, underscoring the need for approaches such as NetraAI that may identify interpretable patient subpopulations most likely to benefit from emerging therapies," said Dr. Joseph Geraci, chief technical officer and founder of Netramark. "Technologies capable of identifying biologically meaningful responder subgroups could fundamentally reshape how Alzheimer's clinical trials are designed."