University of Oxford-led research using cardiovascular artificial intelligence technology developed by Caristo Diagnostics, an Oxford spin-out, will be presented at the European Society of Cardiology (ESC) Congress 2026 in Munich, Germany, from 28 to 31 August 2026. The studies span cardioembolic stroke prediction from routine cardiac CT, the behaviour of Caristo's inflammation measure on non-contrast cardiac CT, and arterial inflammation analysed across more than 80,000 patients in the UK and the United States. All of it is research: these are potential applications under investigation, distinct from the products Caristo currently markets.
Caristo develops AI tools designed to analyse cardiac CT imaging for biological signals associated with coronary inflammation and plaque. Its FAI-Score™ is the company's measure for assessing inflammatory signals in the tissue surrounding the coronary arteries from cardiac CT images. The material point for investors is that the ESC programme examines where those signals might be read next — not a widening of what has been authorised for clinical use.
Oxford-led research at ESC
The headline item is a late-breaking presentation, "Using AI to Detect Atrial Myopathy and Predict Cardioembolic Stroke from Routine Cardiac CT: The ORFAN-MAESTRIA Study", to be delivered by Prof. Charalambos Antoniades on 31 August 2026. The study examines whether AI applied to cardiac CT scans already acquired in routine practice can identify atrial myopathy and inform prediction of cardioembolic stroke. The findings have not yet been presented, and no clinical conclusion should be drawn ahead of that.
A second strand looks at coronary inflammation quantified on both contrast and non-contrast cardiac CT, and whether the FAI-Score behaves usefully on non-contrast scans — a question with practical weight, given how many non-contrast studies are performed for calcium scoring. Related work explores whether inflammatory signals can flag higher-risk patients whose calcium score is zero, a group that current risk models can under-recognise. A further study examines coronary inflammation and cardiovascular risk in patients without standard modifiable risk factors.
A separate presentation describes a "virtual biopsy" of arterial cytokine activation and its relationship with long-term cardiac mortality across more than 80,000 patients from the UK and the US. The phrase is a research descriptor for an imaging-derived inference, not a clinically approved substitute for tissue biopsy. Oxford researchers will also take part in wider ESC sessions covering AI in cardiovascular prevention, AI-supported risk stratification and advanced cardiac imaging.
Expanding the potential applications of cardiac AI
The distinction between research and authorised product runs through the whole programme, and it is the part most easily lost in coverage. Caristo's currently available portfolio consists of CaRi-Heart, which holds U.S. FDA De Novo marketing authorization, and CaRi-Plaque, which holds U.S. FDA clearance. These are different regulatory routes with different evidentiary bases, and neither is accurately summarised as "FDA approved".
Everything being presented in Munich sits outside that authorised perimeter. Caristo chief executive Steve Deitsch has framed the ESC research as an effort to explore applications beyond coronary artery disease, while drawing a line between those research programmes and the company's authorized products. UKFOS makes no assessment of the technology's clinical performance; the studies will speak for themselves once presented and published.
From Oxford research to clinical technology
Caristo was founded as a University of Oxford spin-out, and the research being presented remains Oxford-led — a translational pathway that runs from academic imaging science through clinical validation to regulated product. According to the supplied source, the company was named a World Economic Forum Technology Pioneer in 2026.
That trajectory is the reason the story sits within UKFOS themes rather than outside them. Oxford's cluster continues to produce companies whose value is created in the gap between a research finding and a regulated, reimbursed clinical product — a gap measured in years, cohorts and regulatory submissions rather than in software releases.
Why it matters for investors
Three observations, offered as context rather than advice. First, AI is moving deeper into clinical imaging and risk stratification, where the input data — cardiac CT scans acquired for other reasons — already exists at scale, which changes the economics of deployment relative to technologies requiring new hardware or new patient pathways.
Second, university spin-outs are increasingly the vehicle through which medical research becomes a commercial diagnostic platform, making the quality of the underlying academic programme a durable part of the investment case rather than a footnote to it.
Third, and most practically: regulatory authorisation of an existing product and research into new applications are different stages of technology development, and they carry different risk. Investors assessing companies in this category should look at which claims are authorised, which are under investigation, and which are neither — a discipline that applies well beyond cardiovascular imaging.
UKFOS editorial · published 25 August 2026