An AI tool that reads ordinary heart CT scans for hidden arterial inflammation flagged possible cardiovascular risk in 42.9% of asymptomatic, low-risk adults — patients whom standard blood tests and conventional risk calculators could not distinguish from genuinely healthy ones, according to a study announced Wednesday by preventive health clinic Biograph.

The first US study of its kind by the company used Caristo Diagnostics' CaRi-Heart technology, which analyzes coronary CT angiography (CCTA) images to quantify inflammation in the tissue surrounding the heart's arteries. Biograph announced the results in an October 22 press release, and said the analysis was offered to clinically eligible members of its top-tier membership. As always with company-announced research, the findings deserve scrutiny rather than headlines — more on that below. For more on how AI is reshaping medicine and science, see our AI industry coverage.

What the Study Found

The research combined Biograph's high-resolution CCTA imaging and cardiometabolic profiling with Caristo's CaRi-Heart and CaRi-Plaque technologies to assess coronary inflammation in adults with no known heart disease. The headline findings, per the announcement:

  • 42.9% of asymptomatic, low-risk participants exhibited elevated coronary inflammation
  • 17.9% showed very high inflammation in at least one major heart artery
  • Inflammation frequently appeared without any plaque buildup or blockage — about 47% of participants with no detectable coronary plaque still had elevated coronary perivascular inflammation

The company argued in announcing the results that evaluating coronary inflammation alongside traditional imaging of plaque burden and stenosis yields "a more complete picture of cardiovascular risk." Biograph's position is that AI-detected inflammation may identify an earlier window for intervention than either plaque scans or bloodwork.

Standard Risk Tools Missed the Signal

Perhaps the most consequential claim in the study is negative: the usual suspects of cardiovascular screening failed. According to the release, routine bloodwork and clinical tools — including cholesterol markers such as ApoB and non-HDL cholesterol, blood sugar (HbA1c), systemic inflammation (hs-CRP) and the widely used AHA PREVENT-ASCVD risk calculator — "showed no correlation with the presence of coronary inflammation and failed to identify which patients had inflamed arteries."

Dr. Michael Doney, Biograph's executive medical director, said the findings suggest the technology may be "identifying an upstream phase of disease and a window for proactive intervention."

The Technology, and Its Regulatory Status

CaRi-Heart is not a research prototype. Caristo's technology has received FDA De Novo marketing authorization — the pathway for novel low-to-moderate risk devices — and is described in the release as the first and only solution cleared to quantify coronary inflammation from routine coronary CT angiography. Its companion tool, CaRi-Plaque, holds FDA clearance for AI-assisted plaque and stenosis analysis.

The caveats worth keeping in view

Three qualifications matter. First, the results arrive via press release from a company that sells the underlying diagnostics; the clinical manuscript has been posted as a preprint, meaning it has not been through formal peer review. Second, the release does not disclose the participant count or demographic breakdown, making it impossible to judge how robust the 42.9% figure is across ages and risk profiles. Third — and to Doney's credit, this is stated plainly — "more research is needed to validate this finding and establish the prognostic value and benefit of care directed at coronary perivascular inflammation." Detecting inflammation is one thing; proving that treating it changes outcomes is another.

Why It Matters

Cardiovascular disease remains the world's leading cause of death, and a large share of first heart attacks occur in people who looked low-risk on standard screening. If AI analysis of scans already being taken — CCTA is a routine, non-invasive test — can reliably surface that hidden risk years earlier, it would be a genuine shift in preventive cardiology, and rival approaches to AI heart imaging are advancing quickly.

The idea behind the technology is medically intuitive. The coronary arteries run through sleeves of fat tissue, and inflammation in that perivascular fat changes measurable signals on a CT scan. Where conventional angiography asks "is the pipe narrowed?", inflammation analysis asks "is the disease process active?" — a question that matters because most heart attacks are triggered by plaques that were never severely narrowed in the first place.

There is also a screening-population catch worth naming. Testing large numbers of healthy people for subtle signals inevitably surfaces findings whose clinical meaning is uncertain, and Biograph's own announcement concedes that no routine marker correlated with the AI flag. Whether an inflamed artery in an otherwise low-risk 45-year-old is a treatment target or a false alarm is exactly what longitudinal follow-up must establish.

The honest summary of Wednesday's announcement: promising, regulator-endorsed technology, an eye-catching prevalence figure, and an evidence base that is early, self-reported and awaiting peer review. The preprint and any follow-up studies will determine whether 42.9% becomes a number that changes screening guidelines — or a cautionary tale about reading too much into a vendor's own cohort.

---

Stay Ahead of AI

Get the latest AI news, analysis, and breakthroughs — all in one place.

Read more AI news →