Oxford Heartbeat and the EIC Accelerator: Stent Trial Unpublished

A flow diverter is a mesh stent a few millimetres wide that is pushed through the arteries into the brain to seal off an aneurysm, and once deployed it can end up a very different length from the one printed on its label. Oxford Heartbeat, a London company founded in 2015, sells software called PreSize Neurovascular that builds a 3D model of the patient's vessels from the pre-operative scan and lets the surgeon place virtual stents in it before choosing one. Its own website says a case can be rehearsed in about three minutes.

The EIC Accelerator selected the company's next project, PreSize Surgery, at its May 2026 cut-off for a grant only. CORDIS records an EU contribution of €2,499,031.52, an EC signature on 9 September 2026 and a project running from 1 October 2026 to 30 September 2028. The new project goes further than planning: it promises guidance during the operation itself. That ambition arrives before the company's main evidence for the existing product, a 103-patient prospective NHS trial that closed in July 2025, has been published.

Sizing a stent before the operation

The planning problem is documented in the company's peer-reviewed papers. In a 2022 Journal of Neurosurgery study of 138 past Pipeline stent cases from five UK centres, the authors found deployed lengths that differed from the labelled length "in some cases by more than double", with no correlation between device size and how much it stretched. The software predicted the deployed length with a mean accuracy of 95.61%, or 96.13% after four outlier cases were excluded. One of the six authors works for Oxford Heartbeat.

The company's CE marking release of 16 March 2020 reported an earlier figure of 96.75% from 65 patients at four UK centres, against a 90% expectation it attributed to interventional neuroradiologists. Its company history page lists the CE mark under 2021 instead; the release is dated March 2020 and names the British Standards Institution as the notified body.

Two further studies extend the picture. A Hamburg study of 98 Derivo devices, published online in October 2023 with three Oxford Heartbeat employees among its authors, found 94.54% accuracy, and the two neuroradiologists using the software chose stents about 4 mm shorter on average than those actually implanted. A 2024 comparison in the Journal of NeuroInterventional Surgery, with no Oxford Heartbeat author, gave eight experienced specialists 51 past cases and found their estimates 81% accurate against 95% for the software.

All three studies are retrospective: they test predictions against cases that had already happened. None shows whether using the software changes which stent a surgeon picks in a live case, or whether that changes outcomes. That was the job of the prospective trial.

The trial that has not reported

The ISRCTN registry entry describes a post-market, non-randomised study sponsored by Oxford Heartbeat and funded by the National Institute for Health and Care Research. Surgeons planned each case twice, once their usual way and once with PreSize. The primary outcome is the accuracy of the simulated stent length in real practice; secondary outcomes include differences in the devices chosen, planning time, procedure time, radiation dose and intra-operative corrections. The registry records 103 patients enrolled between 4 August 2022 and 31 July 2024 at nine hospitals in England, Scotland and Wales, and an overall completion date of 31 July 2025.

The recruitment deadline was moved three times, from the end of 2022 to July 2024, and the intended publication date to 31 July 2025. The company announced on 9 July 2025 that the European Society of Minimally Invasive Neurological Therapy had endorsed the trial, that the Imperial Clinical Trials Unit was analysing the data independently, and that results were "anticipated to be published later this year". On 4 November 2026 the registry listed only a 2023 research summary as an output, and a PubMed search for "PreSize Neurovascular" returned the three retrospective studies above and no trial paper.

Two figures in the company's own material also differ from the registry. Its trial announcements gave 100 patients in 2024 and 103 in 2025, and the registry records 103. Its homepage, checked on 4 November 2026, says ten UK hospitals take part in the trial; the registry lists nine.

What PreSize Surgery promises

The CORDIS objective moves from planning to the operating room. It describes combining the pre-operative 3D simulation with real-time intra-operative guidance "powered by generative AI", fusing patient imaging with live surgical data to detect deviations during stent deployment. The project's own figures: brain aneurysms in 5% of the population, a 50% fatality rate on rupture, stents costing about €10,000 each, a European economic burden above €4.7 billion a year, a target of "increasing surgical accuracy to 95%", and a market opportunity above €550 million a year. These are the applicant's numbers, not findings of the EIC.

1. Accurate pre-operative stent simulationPublished94.5% to 95.6% in three retrospective studies, two co-authored by company staff.
2. Clinical impact in live casesNot publishedTrial ISRCTN17947655 completed July 2025; no results output listed on 4 November 2026.
3. Real-time intra-operative guidanceToo earlyProject started 1 October 2026; CORDIS shows no report yet.
4. Fewer complications and wasted stentsClaimStated in the project objective; no outcome data in public.

A small company with public money behind it

The Companies House entry shows the company as active, with a confirmation statement dated 13 July 2026 and its next accounts due by 31 December 2026. The filing history records micro company accounts for the periods to 31 July 2021, 31 July 2023 and 31 December 2024, a format that does not publish turnover or a profit and loss account, so no revenue or staff figure is public. It also records share allotments in June 2024, January 2025 and January 2026, without the amounts raised being stated on the register summary.

No priced equity round has been announced by the company. Data aggregator CB Insights lists about $2.08 million raised across seven rounds, with backers including Creative Destruction Lab, Innovate UK and DigitalHealth.London. The company's history page credits several Innovate UK grants from 2016, a Royal Academy of Engineering fellowship in 2017 and the 2020 NHS AI in Health and Care Award, one of 42 winners from about 530 applicants, which funds the trial. Spranger told Sifted, as summarised by NeuronFeed in July 2026: "If public grant funding did not exist, Oxford Heartbeat would not exist in the same shape or form today." The same report gives more than 1,000 surgeries supported in 14 European countries, a figure that could not be checked against a second source.

A UK company in an EU programme

The UK is an associated country of Horizon Europe, and Innovate UK Business Connect states that UK legal entities are eligible to receive its funding; CORDIS records Oxford Heartbeat as a self-declared SME at signature. CORDIS tags the project 100% to artificial intelligence and 40% to the digital agenda. The software has been studied at the University Medical Center Hamburg-Eppendorf and by Finnish clinicians, and in June 2026 the European Innovation Council named Spranger winner of the 2026 EU Prize for Women Innovators, worth €100,000, an award the Horizon Magazine profile ties to her work at the company.

The 2020 CE release said the company planned to move from the old directive to the EU Medical Device Regulation. Whether PreSize Neurovascular now holds an MDR certificate could not be verified in the public sources reviewed, and an openFDA search found no 510(k) clearance for the product in the United States.

The verdict

Oxford Heartbeat has not been asked for comment for this article. The reporting is based on publicly available sources, listed below.