University of Leeds spinout adsilico has secured £10m in Series A investment to expand technology that creates virtual patient populations for medical device testing. The financing was led by PXN Ventures, with existing backers Northern Gritstone and Parkwalk participating. The company intends to broaden its simulation models beyond the heart and increase its workforce from 14 to 34 people.
Medical device developers face substantial costs in establishing whether an implant will perform safely across patients who differ in anatomy, age and medical condition. Clinical trials can be constrained by recruitment costs and the difficulty of enrolling a sufficiently varied group. adsilico’s Virtual Chimera technology uses generative AI to create virtual hearts and blood vessels from anonymised clinical scans, then subjects them to physics-based simulations of device performance. The purpose is to test design questions before committing to later development and clinical activity.
The distinction between simulated evidence and clinical evidence is central. A virtual model can explore a wider range of anatomical scenarios, but it does not itself demonstrate safety or replace all laboratory, clinical and regulatory requirements. Model accuracy, assumptions, validation and the relevance of input scans affect the reliability of the results. The company positions the system as an additional evidence source alongside established testing methods, rather than a licence to bypass them.
The Leeds team cites longstanding concerns about under-representation in medical device studies. It refers to research of 195 trials of high-risk implants in which women accounted for 33% of participants, falling to 29% for cardiovascular trials. Where particular groups are under-represented, manufacturers can have less evidence about device performance in those populations before products are widely deployed. Simulations can potentially help engineers investigate anatomical variability before a final design is fixed or an expensive trial protocol is agreed.
For manufacturers, these models may have particular value when deciding which device sizes to develop, which anatomical features could make implantation more difficult and where additional evidence is needed. A cardiovascular implant designed around a narrow set of vessel sizes may require further design work to accommodate different patient groups. Identifying those constraints earlier could influence tooling, production planning and the choice of patients for later studies. Such benefits remain contingent on the quality of the simulations and their acceptance by regulatory reviewers.
Chief executive and co-founder Sheena Macpherson said the investment would extend the company’s work into other therapy areas, including devices used to treat brain aneurysms. Neurovascular interventions can involve particularly small patient populations and difficult anatomical conditions. The business has already worked on projects involving heart valve and neurovascular device manufacturers, but has not quantified how many eventual product approvals its technology has accelerated.
The underlying technology combines two different technical disciplines. Generative techniques construct anatomical representations from historical imaging data, while mechanical simulations attempt to predict how devices interact with those structures. Each stage can introduce uncertainty: a plausible virtual organ is not automatically representative of all patients, and an accurate physical model still requires reliable material and boundary assumptions. The commercial service therefore depends on validation that customers and regulators can inspect and understand. Such assurance may become a competitive differentiator as the field grows. It is also why the company describes the evidence as informing, rather than replacing, real-world device development and clinical evaluation.
Founded in 2024 by Sheena Macpherson, Dr Nishant Ravikumar and Dr Zeike Taylor, adsilico operates from Nexus, the University of Leeds innovation hub. Its next technical work includes extending virtual patient models beyond cardiovascular implants to neurovascular devices such as flow diverters. The company has completed projects with medical device manufacturers, but computer-generated evidence still has to be evaluated against the standards applicable to each device and intended clinical use. The £10m round finances further model development and recruitment, rather than a change to those regulatory requirements.
You must be logged in to post a comment.