adsilico Raises £10M to Expand Virtual Patient Models

The big picture: adsilico, a University of Leeds spinout, has raised £10M in a Series A round led by PXN Ventures. The company uses generative AI to create virtual hearts and blood vessels from anonymized scans, aiming to address the under-representation of diverse patient groups in medical device trials.

Why it matters:

• Trial Diversity Gap: Medical device trials often under-represent women and other diverse patient groups, leading to devices that may not work effectively or safely for all. A JAMA Internal Medicine study found women comprised a median of 33% of participants in high-risk implant trials.
• High Development Costs: Late-stage mistakes in medical device development are costly, with the average cost of developing a new complex medical device around $522 million, including failures. Virtual trials can identify issues earlier.
• Regulatory Scrutiny: Regulators, including the MHRA in the UK, are increasingly scrutinizing bias in medical device design and testing, pushing for more robust guidance on diversity and equity in care.

How it works:

• Generative AI Models: adsilico uses generative AI to create virtual hearts and blood vessels from anonymized clinical scans, enabling the generation of diverse virtual patient populations.
• Physics-Based Simulation: These virtual models are then subjected to physics-based simulations to assess medical device performance across various patient demographics, including age, sex, ethnicity, and body size.
• Early Insight: The platform provides manufacturers with early insights into the impact of anatomical diversity, informing device design, identifying higher-risk patients, and helping to design more inclusive clinical trials.

The catch: While adsilico’s technology offers significant potential for improving medical device development and patient safety, the ultimate challenge lies in gaining full regulatory acceptance for virtual-patient data as definitive proof for real-world device efficacy. The FDA’s current guidance on computational modeling for devices does not yet fully cover standalone machine learning models, indicating a gap that needs to be addressed for broader adoption.

Key Facts

• Company: adsilico
• Amount: £10M
• Round: Series A
• Investors: PXN Ventures (lead), Northern Gritstone, Parkwalk
• Founders: Sheena Macpherson, Dr Nishant Ravikumar, Dr Zeike Taylor
• Sector: Medical Device AI Simulation
• Headquarters: Leeds, UK