Owkin enters into agreement to build innovative methods with the potential to enhance AI-enabled external control arms

Owkin, a startup pioneering Federated Learning and AI technologies for medical research and clinical development, announces an agreement with Actelion Pharmaceuticals Ltd., one of the Janssen Pharmaceutical Companies of Johnson & Johnson, to augment clinical trials with advanced machine learning methodologies. The aim of this initial project with Janssen’s R&D Data Science team is to investigate innovative machine learning-based methods for the estimation of treatment effect in clinical trials involving real-world data sources.

Owkin’s expertise with machine learning and multimodal real-world data powers innovations that can be leveraged to support decision-making in Drug Research and Development, biomarker identification, and clinical development processes.

Owkin and Janssen’s R&D Data Science team will focus on innovative double/debiased machine learning-based approaches that enable adjustment for high-dimensional confounders to overcome important challenges of standard methods, such as bias and confounding. Detecting efficacy with small trials and external control cohorts, which is often the case for rare diseases, is a challenge. Double/debiased machine learning, a method developed originally in the context of econometrics with contributions from Nobel Prize recipient Esther Duflo, maybe a way of achieving sufficient statistical power in this particular setting.

This first project with Janssen focuses on Pulmonary Arterial Hypertension (PAH), a rare, progressive disease where the pressure in the blood vessels of the lungs is elevated, resulting in stress on the heart. Despite recent advances, PAH still has no cure and remains a severely debilitating condition that leads to heart disease and early death. PAH is difficult to diagnose, but early diagnosis and treatment are critical to helping improve life expectancy.

Gilles Wainrib, Owkin Co-Founder and Chief Science Officer

We’re thrilled to embark on this project to demonstrate how imperative machine learning methodologies are to improve clinical trial design and evaluation. Ultimately this has potential to help bring safe and effective drugs to patients faster.

The results from this project could potentially support regulatory submissions to health authorities, bringing much needed methodological innovations into practice. The methodologies deployed with this project are disease area-agnostic and have the potential to be used in multiple other applications throughout the discovery and development pipeline.