Medical device trials enrol a few hundred people, and that number rarely resembles the millions who eventually receive the implant. A University of Leeds spinout has now banked 10 million pounds to test devices on simulated populations before surgeons touch a real patient.
adsilico raised the Series A from PXN Ventures, with existing backers Northern Gritstone and Parkwalk returning. The company builds AI simulation software that models how a heart valve or a stent will perform across widely varied anatomy and physiology.
That gap is the product. A pivotal trial for a prosthetic valve may report outcomes in a few hundred patients who skew older, whiter and healthier than the eventual market. Devices can behave differently in a small aortic root, a calcified vessel or a younger patient.
The company, founded in 2024 by Sheena Macpherson, Dr Nishant Ravikumar and Dr Zeike Taylor, argues simulation can expand that sample without recruiting anyone. Regulators across the US and Europe have spent years signalling interest in in-silico evidence, and the UK’s Medicines and Healthcare products Regulatory Agency has pushed computational modelling as a route to cut trial cost and delay.
For manufacturers, cheaper pre-clinical iteration changes the economics of device design. For regulators, the harder question is whether a simulated cohort is auditable enough to stand in for a real one. adsilico’s next milestone is proof that it is.
