Connect with us

NEWS

Adsilico’s £10 Million Bet on Hearts That Never Lived

PXN leads a £10 million Series A for adsilico virtual patients, a wager that synthetic hearts can fill gaps real implant trials still leave.

Published

on

University of Leeds spinout adsilico has raised £10 million in Series A funding to test heart and brain implants on virtual patients built with generative AI. PXN Ventures led the round on 26 October 2026, with Northern Gritstone and Parkwalk joining again.

The cheque is a wager that synthetic anatomy can answer questions a trial of a few hundred people still cannot, before a device reaches a body.

A £10 Million Series A for Virtual Hearts

adsilico is based at Nexus, the university’s innovation hub, and sells in-silico studies to makers of heart valves, stents and related implants. Sheena Macpherson is chief executive and co-founder, working with Dr Nishant Ravikumar, head of virtual populations, and Dr Zeike Taylor, head of simulations. The company’s own site also lists Professor Alejandro Frangi among the 2024 founders.

Northern Gritstone, which was launched by the universities of Leeds, Manchester and Sheffield, put £5 million into this round, matching the headline on its own 26 October announcement. Parkwalk, a backer of university spinouts, returned as well. The new money is meant to deepen the virtual heart models, move into neurovascular devices such as flow diverters used on brain aneurysms, and grow the team from 14 to 34 people.

ADSILICO FUNDING ROUNDS

Round Date Amount Who put the money in
Seed February 2024 £3.5 million Northern Gritstone £2 million, Parkwalk £1.5 million
Series A October 2026 £10 million PXN Ventures lead; Northern Gritstone £5 million; Parkwalk returns
Equity total £13.5 million Seed plus Series A

That combined £3.5 million seed round was sold as an end-to-end in-silico trial stack, built on work funded by the Royal Academy of Engineering at Leeds. The Series A is the same thesis at commercial scale: device teams pay to see how an implant behaves across anatomies they will not recruit in time.

Implant Trials Still Enrol Women as a Minority

Heart valves, stents and similar devices still reach the market on evidence from a few hundred participants, and those groups skew male. In 2024, physician Nitzan Karny Epstein of Meir Hospital in Kfar Saba, Israel, and colleagues reviewed 195 randomised trials of high-risk implants published from January 2016 to May 2022. The median share of women was 33%, with an interquartile range of 24% to 48% and a spread from 7% to 100%.

Cardiovascular studies were thinner still. The median share of women was 29% in that field, 27% in stent trials, 24% in pump trials and 44% in valve trials. Orthopedic implant studies sat at 46%. The reviewers found no clear rise in women’s share by start year, including trials that began after the US Food and Drug Administration’s 2014 sex-specific device guidance.

WOMEN IN HIGH-RISK IMPLANT TRIALS

Group Median share of women
All 195 trials 33%
Cardiovascular trials 29%
Stent trials 27%
Valve trials 44%
Pump trials 24%
Orthopedic trials 46%

Sex-related covariates such as body size, hormones, other illness and pregnancy history can change how a device sits and wears, the authors wrote, so a male-heavy sample leaves clinicians short of data on the people who will actually receive the implant. Obstacles they list include under-referral, criteria that favour men, and a long-running fear about pregnancy that their own age data undercuts: mean age was over 50 in 175 of the 195 trials.

adsilico’s pitch is that a virtual cohort can be specified on purpose. The company says a maker can ask how a valve or stent behaves in, for example, women aged 55 to 75 of South Asian heritage, then use that readout to pick sizes, flag higher-risk anatomies and shape a later human study.

How Virtual Chimera Tests a Heart Valve

The product is called Virtual Chimera. It starts from anonymised clinical scans, then builds fully synthetic virtual patients the firm says never existed, so no real patient is exposed. Physics-based simulation then runs the device inside each anatomy under clinical oversight, with each stage documented.

HOW A VIRTUAL CHIMERA STUDY RUNS

  • The question: The team agrees the decision, the relevant anatomy and the outputs a device maker needs, often around feasibility or a pivotal trial.
  • The population: Generative models span sex, age, ethnicity, anatomy and disease severity, including rare shapes trials rarely catch.
  • The readout: Simulations, reviewed with clinicians, show which sizes and groups perform well, where risk concentrates and where uncertainty remains.

Two worked examples sit on the company site. In a mitral repair study, simulated leaflet seal, known as coaptation length, increased by 3.2 mm after the device was placed. In a neurovascular example, mean blood-flow velocity inside an aneurysm was 66% lower after implant placement than in the pre-implant model. Those figures are the firm’s own illustrations, not independent trial results, and they are the closest public specimen of what a buyer actually receives.

The sequence is the product. A physical implant can now meet a digital twin, and a whole synthetic neighbourhood of twins, before it meets a person. That is the claim device makers are being asked to buy.

Patients miss out on the chance of a life-saving device, or face uncertainty about whether it will work for them, because people like them weren’t represented when it was designed and tested. We give medical device manufacturers early insight into the impact of anatomical diversity when it can still shape the design and help avoid costly mistakes or harm to patients.

Sheena Macpherson, CEO and co-founder, adsilico

Dr Anthony Holmes is chief product officer and Beckie Bown is chief financial officer. Clinical advisors named on the site include Dr Stuart Grant in cardiothoracic surgery at Manchester University NHS Foundation Trust, Professor Dan Blackman in interventional cardiology at Leeds Teaching Hospitals, and Dr Hamed Nejadhamzeegilani in interventional neuroradiology at the same Leeds trust. The company says virtual populations are checked against real patient data, and that clinical advisors review whether a result makes clinical sense.

Investors Pay for a Conversation With Regulators

Dr Elizabeth Young, investment manager at PXN Ventures, said the Leeds team could affect device makers, the pre-clinical process and, “most important of all, patients,” and that PXN was joining as the firm “sets new standards for in silico testing.” Duncan Johnson, chief executive of Northern Gritstone, tied the round to groups trials have underserved.

Medical device manufacturers need to know that their products will work for the people who receive them, including populations that have been traditionally underserved. adsilico’s use of generative AI to build realistic virtual hearts and blood vessels provides vital information to help understand how these products will work much earlier, spotting potential problems before committing to costly human trials.

Duncan Johnson, Chief Executive, Northern Gritstone

Dr Anne Dobrée, investment director at Parkwalk, called Virtual Chimera a response to “a lack of diverse, representative data in clinical trials” and said the new capital should turn the science into partnerships at scale. Parkwalk has been in since the earliest days, after its £2 million seed investment in 2024 sat beside Northern Gritstone’s first cheque.

None of those lines claim a regulator has blessed a virtual cohort as a stand-alone proof. They claim an earlier, cheaper look at failure modes, and a file a sponsor can take into a conversation about design, sizing and who to enrol. That is the bet the £10 million is actually on.

Flow Diverters and a Headcount of 34

Work so far has centred on virtual hearts, with completed projects for heart-valve and neurovascular device makers, none of them named. Cardiac research is due to be presented at TCT, the Transcatheter Cardiovascular Therapeutics meeting, which runs from 31 October to 3 November 2026 at the San Diego Convention Center. That meeting is where structural-heart device teams compare notes on valves, clips and the evidence behind them.

The Series A is meant to carry the same method into the brain, starting with flow diverters for aneurysms, a class the company says has some of the smallest trial numbers. Hiring 20 people at Nexus is the practical half of that plan: more simulation staff, more clinical review, more therapy areas.

FROM SEED TO SERIES A

  1. February 26, 2024: Receives a £3.5 million seed from Northern Gritstone and Parkwalk to build the in-silico stack and the first team.
  2. 2024: Spins out of the University of Leeds, with Macpherson, Ravikumar, Taylor and Frangi as founders, after more than 15 years of academic research.
  3. October 26, 2026: Closes a £10 million Series A led by PXN Ventures, with Northern Gritstone adding £5 million, and plans to grow from 14 to 34 people.
  4. October 31, 2026: TCT opens in San Diego, where adsilico said its cardiac work would be presented.

Chris Richardson chairs the board, with more than 35 years in medical devices and exits that include a sale to Abbott. Frangi, named on the company site as a founder, is Bicentennial Turing Chair of Computational Medicine at the University of Manchester. At seed he argued that devices still go through a long evaluation with only a tiny share of the scientific evidence coming from computer modelling, and that the cost keeps delaying benefits for patients.

ASME V&V 40 and the Credibility File

adsilico is careful on the point investors like to blur. Model credibility, it says, is judged for a stated question: a model that is credible for sizing is not automatically credible for every other use. The firm assesses that work against ASME V&V 40, the risk-informed standard on which the FDA’s 2023 guidance on computational modeling builds. That file, the company writes, “supports a regulatory conversation; it is not regulatory approval.”

The 2023 FDA document tells sponsors how to argue that a computational model is credible for a premarket submission, including 510(k), PMA, De Novo and IDE files. It treats in-silico device testing and in-silico clinical trials as real uses of modelling, and it asks for final simulation results plus the credibility evidence behind them. It also says in-silico trials are still an emerging method, and that the guidance does not set specific rules for how to generate virtual cohorts.

In the UK, the Medicines and Healthcare products Regulatory Agency said on 25 March 2026 that it wants developers of medicines to submit non-animal data, including in-silico methods, with notes on robustness and human relevance, in parallel with conventional animal studies if those still exist. That note is about medicines applications, not implant files, and the agency said it does not expect a one-for-one swap of a named animal study for a named alternative. The two documents still describe the climate in which adsilico is selling: regulators will look at computer evidence, and they will not treat it as a rubber stamp.

WHAT WE KNOW

  • The round: £10 million Series A closed on 26 October 2026, taking equity funding to £13.5 million, with a planned rise from 14 to 34 staff at Nexus.
  • The method: Virtual Chimera builds synthetic hearts and vessels from anonymised scans, then runs physics-based implant simulations under clinical review.
  • The rulebook: FDA’s 2023 modelling guidance and ASME V&V 40 are the credibility frame the company cites; MHRA’s March 2026 note covers in-silico data in medicines files.

WHAT IS UNCONFIRMED

  • Named customers: Heart-valve and neurovascular projects are described, but the device makers are not.
  • TCT contents: Cardiac work is due in San Diego; the slides and any peer-reviewed paper are not yet public.
  • Approval path: No public file shows an implant licensed on adsilico cohorts alone, and the company does not claim that outcome.

The near work is concrete. Twenty hires, a brain-device library, and a TCT audience of valve and aneurysm teams will show whether those files get bought as a habit. The models support a conversation. They are not a stamp.

Frequently Asked Questions

Who Founded Adsilico and Where Did the Science Start?

Sheena Macpherson, Dr Nishant Ravikumar, Dr Zeike Taylor and Professor Alejandro Frangi founded the company in 2024 as a University of Leeds spinout, after more than 15 years of academic work that the seed announcement tied to the Royal Academy of Engineering and to Leeds’ schools of medicine and computing. Frangi holds the Bicentennial Turing Chair of Computational Medicine at the University of Manchester and has also been described as director of the Christabel Pankhurst Institute for Health Technology Research; earlier write-ups of the seed round placed the underlying technique at the Centre for Computational Imaging and Simulation Technologies in Biomedicine at Leeds.

Does a Virtual Chimera Study Replace Animal Tests or a Human Trial?

The company frames the simulations as an extra layer beside laboratory, animal and clinical studies, and says human-relevant evidence can be gathered before animal work begins, not that those studies vanish. FDA’s 2023 modelling guidance treats in-silico trials as an emerging use and, in the same document, says it does not provide specific recommendations for generating virtual cohorts or running an in-silico trial, which is why adsilico sells a documented credibility file rather than a substitute approval pack.

How Far Below Population Share Were Women in Those Implant Trials?

Karny Epstein’s team also computed a participation-to-prevalence ratio, the share of women in a trial divided by the share of women among people who receive that intervention. Ratios below 0.8 are read as under-representation: 0.77 for stents, 0.76 for valves, 0.69 for orthopedic prostheses and 0.56 for pumps, with pacemakers at 1.00. They reported no significant trend in women’s share by publication year or by trial start year across the period they studied.

What Is a Flow Diverter in This Expansion Plan?

A flow diverter is an implant used to treat brain aneurysms by changing blood flow in the parent vessel so the aneurysm can thrombose. adsilico named that class as the first neurovascular target for the Series A because, it said, trial numbers there are among the smallest, which is the same recruitment bind its heart models are sold to ease.

How Does the Company Say It Documents Trust in a Model?

Each study, according to the company, records data provenance, assumptions, validation and uncertainty, and clinical advisors review whether the output makes sense for the patients they treat. Credibility is scored against ASME V&V 40 for a single stated question, such as sizing, and the firm is explicit that a pass on that question is not a general licence to use the same model for every other decision.

Twenty new posts at Nexus and a San Diego meeting will test whether valve and aneurysm teams pay for those files after the slides are shown. The models are for the questions a few hundred recruits still leave open.

Disclaimer: This article is news reporting on a completed funding round and on published research and regulatory guidance; it is for information only. It is not medical advice, not a recommendation to use or avoid any implant, and not investment advice about adsilico, PXN Ventures, Northern Gritstone, Parkwalk or any other firm. Readers who are patients should speak with a qualified clinician before making treatment decisions, and readers who are considering an investment should speak with a qualified financial adviser. Figures, headcount plans, conference appearances and regulatory positions reflect the company statements and documents cited around the 26 October 2026 round and may change.

Harry is the editor of IAQABA, an independent publication he owns and runs. A decade in journalism, beginning as a reporter and now as the editor of his own titles, has left him with a clear test for what deserves a story: it has to change what a reader knows or decides, and it has to rest on something he can point to. That rules out recycled press releases, forecasts with no data behind them and rumours that no document supports. It leaves room for a great deal, and the site covers news, business, science and technology alongside sports, entertainment and lifestyle, with travel, auto and gaming given the same standard rather than lighter treatment. Sources are primary wherever possible: the regulator's filing, the company's own statement, the transcript, the dataset, or the product on Harry's desk. Figures are checked before they are published and rechecked if a reader questions them. Mistakes are corrected under a published policy. Readers across the world can reach him directly at support@iaqaba.com.

Click to comment

Leave a Reply

Your email address will not be published. Required fields are marked *

Trending