AI analysis of CT scans finds 42.9% of patients have coronary inflammation

▲ AI analysis of CT scans finds 42.9% of patients have coronary inflammation(biograph.com)

28 points by brandonb an hour ago | 18 comments

▲ bonsai_spool an hour ago | parent | next [-]

I don't see how one can spin this into a positive finding - their preprint says that the only correlate they find with their new measurement is BMI. By my reading, not even male sex is associated with more inflammation. It sounds like they've identified a fancy BMI estimator, and not a very good one.

This is in asymptomatic patients, who already have a great estimator for 10-year heart disease mortality risk in the coronary calcium score which isn't even obliquely mentioned in their abstract. Edit: the paper shows no correlation with calcium, either

https://www.medrxiv.org/content/10.64898/2026.10.19.26364786...

▲ gavinray 8 minutes ago | parent | next [-]

  > who already have a great estimator for 10-year heart disease mortality risk in the coronary calcium score

Heart attack at age 29 survivor here.

CAC scores only show calcified plaque, not vulnerable plaque, which is the sort likely to rupture and cause an event.

You can go get a scan and receive a CAC score of 0, walk away thinking you're perfectly healthy, and then die tomorrow because the scan couldn't detect the massive amounts of non-calcified plaque lining your arteries.

The only non-invasive way to measure vulnerable/soft plaque is via CCTA

See, for example, "Atherosclerotic Plaque in Patients with Zero Calcium Score at Coronary Computed Tomography Angiography"

It measured nearly 1 in 10 patients with a 0 CAC score as having vulnerable plaque:

https://pmc.ncbi.nlm.nih.gov/articles/PMC5967134/

  > "The frequency of atherosclerotic plaque in the coronary arteries in 367 patients with zero CS was 9.3% (34 individuals)"

▲ abeppu an hour ago | parent | prev [-]

> Men demonstrated higher rates of elevated FAI-Score than women (52.8% vs. 25.8%, p=0.016; ≥90th percentile: 26.4% vs. 3.2%, p=0.007).

I am not saying this is good or interesting science (IIUC this should make us believe that fewer people with coronary inflammation go on to develop heart disease than we would have otherwise thought?) but their score is at least correlated with sex.

▲ bonsai_spool 11 minutes ago | parent [-]

They went on to say that the only signal that survives multiple hypothesis correction was BMI, after reporting their nominal discovery of an association with sex.

    "Men demonstrated higher rates of elevated FAI-Score than women (52.8% vs.
    25.8%, p=0.016; ≥90th percentile: 26.4% vs. 3.2%, p=0.007). [...] Body mass
    index was the sole secondary correlate surviving false discovery rate
    correction (AUC = 0.71, p=0.001, all-test FDR q=0.028). Regularized
    multivariable models showed modest out-of-sample discrimination for FAI-Score
    elevation (cross-validated AUC = 0.71)."

▲ throw20261022 27 minutes ago | parent | prev | next [-]

> Crucially, no routine blood biomarker, lipid metric, or clinical risk score could identify these individuals with inflamed coronary arteries.

I call bs on this. There are two common blood tests that are meant for it:

1) hs-CRP (High-sensitivity CRP)

2) Lp-PLA2: Lipoprotein-associated phospholipase A2 (Lp-PLA2), also known as platelet activating factor Acetylhydrolase, is an inflammatory enzyme that circulates bound mainly to low density lipoproteins and has been found to be localized and enriched in atherosclerotic plaques. In multiple clinical trials, Lp-PLA2 activity has been shown to be an independent predictor of coronary heart disease and stroke in the general population. Measurement of Lp-PLA2 may be used along with traditional cardiovascular risk factor measures for identifying individuals at higher risk of cardiovascular disease events.

I have routinely had both of these blood tests among others, although I paid myself to get them.

As for hs-CRP, it will likely show an elevation. If you use a binary classifier with a higher threshold of what's safe, e.g. 3.0 instead of 1.0, you will miss it! I am guessing this is why they missed it.

▲

tredre3 19 minutes ago | parent [-]

> Importantly, routine bloodwork and clinical tools – including cholesterol markers (ApoB, non-HDL-C), blood sugar (HbA1c), systemic inflammation (hs-CRP), and a widely used cardiovascular risk calculator (AHA PREVENT-ASCVD) – showed no correlation with the presence of coronary inflammation and failed to identify which patients had inflamed arteries.

The article specifically says those tests didn't correlate with the result of the scans. Why do you call BS on it?

	▲	throw20261022 12 minutes ago \| parent [-]
		As for hs-CRP, I already explained the basis in my comment. It depends on whether they mean correlation in its strict literal sense or only after binary classification. As for Lp-PLA2, they crucially seem to have ignored it altogether.

▲ brandonb an hour ago | parent | prev | next [-]

This particular study looked at inflammation through imaging (CT scans) as part of Biograph's $17,500 program. You can get at the underlying signal fairly cheaply through a simple blood test (hs-CRP).

It's fairly well-known that inflammation is an independent risk factor for heart disease, alongside LDL cholesterol (or ApoB) and blood pressure: https://www.empirical.health/blog/inflammation-and-heart-hea...

The American College of Cardiology looked at a bunch of inflammation biomarkers and standardized on hs-CRP: https://www.jacc.org/doi/10.1016/j.jacc.2025.08.047

▲

jp57 37 minutes ago | parent | next [-]

From the article:

Importantly, routine bloodwork and clinical tools – including ... hs-CRP [among other tests] ... – showed no correlation with the presence of coronary inflammation and failed to identify which patients had inflamed arteries.

	▲	brandonb 13 minutes ago \| parent \| next [-]
		This is one of the things that makes me a little skeptical of their score. If it's accurate, there should be a positive correlation with known biomarkers. (I see you made a similar observation elsewhere.) In their paper they actually find an inverse correlation with hs-CRP, which disappeared when removing two patients who were sick. So I wonder if the sample size being small is partly driving this.
	▲	phoronixrly 22 minutes ago \| parent \| prev [-]
		Is this a positive thing? I read it as the AI results not correlating with test data...

▲

yokaze 38 minutes ago | parent | prev [-]

Maybe you didn't get then the latest news: https://www.tctmd.com/news/zeus-trial-ziltivekimab-fails-red...

Inflammation is likely a marker, but not the cause of heart failures. That doesn't contradict it being an independent risk factor, but up to recently, it was assumed be causal driver.

	▲	brandonb 20 minutes ago \| parent [-]
		Same thing happened with the Lp(a) drugs recently. Practically speaking, statins reduce LDL cholesterol (~50% at the strongest dose) and hs-CRP (~37%) and have cardiovascular outcomes data, so that's likely the route a patient would go based on the medications that are approved today.

▲ jp57 34 minutes ago | parent | prev | next [-]

If all the other biomarkers of inflammation "showed no correlation with the presence of coronary inflammation" as predicted by this scan, then how exactly do they assess the positive predictive value of the scan?

▲ esalman an hour ago | parent | prev | next [-]

Access to the study requires $17,500 per year, just fyi.

	▲	pixl97 an hour ago \| parent [-]
		That or OpenAI needs to turn back on their highly persistent model and we ask it about coronary inflammation.

▲ HPsquared an hour ago | parent | prev | next [-]

Before or after 2020?

▲ Theodores an hour ago | parent | prev [-]

Recently I learned that only those with over-active immune systems survived the 14th century plagues. Therefore, inflammation is to be taken seriously, or at least understood in this historical context, depending on your ancestry.

As for the article, the findings come as no surprise given what we know about the risk factors in the post-smoking world. Nowadays nutrition, specifically saturated fats in combination with lots of sugar and salt, as found in today's artificial foods is showing up in the research (not paid for by 'big beef').

There is no need to over-think it, just stay off the corporate 'food' and cook from fresh ingredients, where you don't get 'HFSS' (UK government term for high fat/salt/sugar) held together with weird chemicals. Sure, nobody has time for that, but this is the frontier of public health, a.k.a. the 'obesity crisis', with a simple solution - actually bother to cook!