AI and multimodal patient data for better, faster clinical trials - Bristol Myers Squibb - Our stories

AI and multimodal patient data for better, faster clinical trials

As artificial intelligence (AI) programs achieve exponential gains in power and sophistication, Bristol Myers Squibb is exploring ways to leverage these technologies at many stages of the drug development process. In May 2022, BMS entered into a multi-year, strategic collaboration with Owkin, initially focusing on cardiovascular diseases.

“Owkin will use AI to enhance the design and execution of our trials, helping us optimize endpoint definitions, patient subgroups, and treatment effect estimation, with covariate adjustment and external control arms,” said Venkat Sethuraman, senior vice president, Global Biometrics and Data Sciences. 

“In the future, this methodology could make trials for rare diseases much more efficient than current approaches. It could even be used to avoid having placebo arms in some trials,” Sethuraman said.

What differentiates Owkin’s patient data access

Owkin leverages multimodal patient data through partnerships with renowned academic centers in the U.S. and Europe. This high-quality patient data, reflecting the evolving standard of care, fuels their AI. They also collaborate through initiatives such as MOSAIC to incorporate innovative modalities like spatial omics to expand their data portfolio. By using privacy-enhancing technologies like federated learning, Owkin grows its global data network while ensuring data security and academic partner trust for research collaborations.

With a unique combination of data access, AI capabilities, and biological expertise, the Owkin partnership enhances the chances of success in drug discovery and development. Here’s a look at their approach for optimizing drug development.

The goal: optimized drug development 

Owkin aims to optimize each phase of clinical development by applying novel machine learning methodologies. They have three offerings to tackle expensive and lengthy clinical trials. 

The first is Owkin’s AI external control arms, which provide early estimates of treatment efficacy for single-arm Phase 1/2 trials to inform phase transition decisions. 

Secondly, they offer inclusion criteria models to inform trial recruitment for better defined patient populations in Phase 2 and 3 trials. 

Finally, Owkin offers data-driven covariate adjustment by developing prognostic biomarker models to identify key covariates linked to outcome from external data analysis. These covariates are controlled for during the statistical analysis of Phase 3 trials to increase statistical power, allow broader eligibility criteria and decrease sample sizes without sacrificing power. 

Other benefits from this partnership

Owkin actively engages with regulators, securing support for their data-driven approaches, including a letter of support from the European Medicines Agency (EMA) and a CE mark for their AI diagnostic tool MSIntuit CRC. The company also has a robust and continuously updated source of multimodal external data, capturing the latest standard of care and disease heterogeneity. Additionally, they have a leading team with renowned expertise in AI and biostatistics, with 42 publications in top journals.

BMS’ multi-year, strategic collaboration with Owkin will initially focus on cardiovascular diseases. “This is an area of significant unmet need, and one where we have the opportunity to use technology in new ways to accelerate potential breakthroughs for patients who are waiting for medicines,” says Sethuraman. 

“Our hope is that AI will eventually help us reduce the cycle time for drug development in all our therapeutic areas,” he said.