By Steve Herne, CEO of Unlearn

Early-stage ALS trials carry enormous weight and uncertainty. With limited patient populations, no validated biomarkers for efficacy, and a reliance on open-label designs, it’s often difficult to know if a drug is working or how to interpret changes in disease progression. That’s why we’re proud to be working with ProJenX to support their hybrid Phase 1 trial, PRO-101, evaluating prosetin, a novel brain-penetrant MAP4K inhibitor in ALS. The study includes an open-label extension, a common design in early ALS research, but one that can make it harder to interpret efficacy signals. 

We recently co-presented this important work with ProJenX at ALS Drug Development Summit, where we shared how Unlearn’s advanced machine learning model for ALS (ALS DTG) is being used to generate digital twins of the study’s participants. These digital twins will help ProJenX to derive deeper, more data-driven insights from their trial. Here’s how.

‍

How Digital Twins Are Used in PRO-101

In PRO-101, participants with ALS receive prosetin, and our digital twins serve as a simulated placebo arm, offering comprehensive individualized forecasts of each participant’s clinical outcomes had they not received treatment. These digital twins help ProJenX:

• Use a rigorously simulated comparator arm without administering placebo in an open-label design, providing a clearer picture of prosetin’s effect on disease progression.
• Explore optimal inclusion/exclusion criteria and endpoints for future studies, even in a small, early-phase study.
• Identify treatment-responsive subgroups by flagging individuals whose outcomes are unlikely under standard care.

Together, these uses provide ProJenX with a smarter, faster path to trial planning and go/no-go decisions while maintaining a lean study design.

‍

Powered by Advanced AI Models

The ALS-DTG used in this trial is powered by over 13,600 patient records from clinical trials, observational studies, and our recent partnership with APST Research, a leading platform for ALS clinical research. This collaboration brought in high-quality clinical data, patient self-assessments, biomarker analyses, and common ALS clinical assessments such as ALSFRS-R, SVC, FVC, and neurofilament light chain (NfL) measurements. 

NfL is now a key prognostic biomarker in ALS, offering a meaningful signal in early trials. By incorporating NfL into our digital twins forecasts, we’re helping sponsors like ProJenX track motor decline more precisely, optimize endpoints, and design trials that are more statistically powerful—even when the data are noisy or the studies are small.

‍
Built for Biotechs

This collaboration reflects how Unlearn supports biotech teams: with flexible engagement models and secure technology that integrates directly with your workflows. Our digital twins are already helping sponsors:

• Reduce sample sizes in Phase 2 ALS trials by up to 18% while maintaining power
• Add power to underpowered exploratory endpoints and subgroups
• Design trials that are faster to execute and more likely to succeed

From early signal detection to powering pivotal studies, Unlearn’s AI solutions support more confident decision-making across your development program.