Phase 1b clinical trial via Modality AI (2023) (www.vergegenomics.com)

Verge Genomics announced it will integrate Modality.AI’s clinically validated multimodal platform into the Phase 1b proof‑of‑concept trial of VRG50635, its lead PIKfyve small‑molecule candidate for ALS. Modality.AI will remotely capture objective speech and language changes—analyzing both what patients say (linguistic content) and how they say it (acoustic/prosodic features and visual cues) via a virtual assistant—using participants’ own devices. The move supplements Verge’s earlier plan to deploy Emerald Innovations’ wireless biosensors for continuous home monitoring of sleep, breathing, gait and mobility, and follows VRG50635’s completed Phase 1 single‑ and multiple‑ascending dose study in healthy volunteers. For the AI/ML community this is a clear example of ML-driven digital biomarkers being embedded into interventional trials to reduce subjective endpoints and increase sensitivity to meaningful clinical change. Modality’s platform—backed by peer‑reviewed validation and existing use across observational and interventional studies—enables high‑frequency, at‑home data capture without clinicians or special hardware, potentially boosting trial power, reducing bias, and accelerating go/no‑go decisions. The collaboration also highlights a convergent trend: combining ML for target discovery (Verge) with ML for endpoint measurement (Modality) to raise the overall probability of clinical success, while underscoring ongoing needs around regulatory acceptance, data privacy, and cross‑site model generalization.