The Longevity Imaging Brief · Issue 02

The Longevity Imaging Brief — Issue 02

The scan you already had is a biomarker panel you never read

a panel you never read

One considered briefing a month on the economics, evidence and governance of preventative imaging — for insurers, health systems, policy teams and investors who need the signal without the sales pitch.

The signal this month

The debate about preventative imaging is stuck on the wrong scan. It argues about whether healthy people should book a whole-body MRI. Meanwhile, the more consequential opportunity is already sitting on hospital servers: the tens of millions of CT scans ordered every year for an unrelated clinical reason, each one carrying a quantified read-out of a person's cardiovascular, metabolic and skeletal risk that no one is being paid to extract.

A single abdominal CT already contains bone mineral density, aortic calcium, visceral fat, muscle mass and liver fat. Automated tools can now read all five, and in combination they match or exceed the best clinical risk models we currently use — on an image that has already been acquired, already been paid for, and already been reported for something else [S11]. That is the difference between ordering more imaging and counting what the imaging already told you. Only one of those adds cost.

Three findings worth your attention

A calcium score is one of the most validated prognostic numbers in medicine — and it can fall out of a scan done for another reason. Coronary artery calcium scoring reliably stratifies cardiovascular risk in people with no symptoms, predicts ten-year events, and reclassifies patients beyond traditional risk calculators; a score of zero identifies a genuinely low-risk group [S1][S3]. Its negative predictive value is high enough to rule out clinically relevant disease [S2]. The honest counter-note: a zero today is not a zero forever — it warrants reassessment every five to ten years, not reassurance for life [S2].

Bone is the quiet biomarker. In the UK alone there were an estimated 520,000 fragility fractures in 2017, costing £4.5 billion, and the number is projected to rise 30% by 2030 [S8]. Yet the vertebral fracture — the archetypal osteoporotic fracture — is the hardest to diagnose clinically, and the spine is often the part of a CT the radiologist was not asked to look at [S8]. Automated bone-density measures pulled from routine CT have out-performed conventional DXA at identifying the people who have actually fractured (AUC 0.885 versus 0.668) [S10]. The technology is validated; what is missing is the workflow and the payment model to use it [S9].

Liver fat and visceral fat are cardiometabolic risk, made visible. Visceral adipose tissue measured on CT or MRI is an independent marker of cardiovascular and metabolic mortality [S13]; a new radiomic index reads early hepatic steatosis off a low-dose calcium scan with an AUC of 0.91, validated against both MR spectroscopy and biopsy [S12]. These are not exotic research signals. They are on scans your members and patients have already had.

The counter-case, stated plainly

If opportunistic biomarkers are the strong side of the ledger, whole-body screening of well people is where the caution belongs — though the problem is Bayesian, not a harm. A meta-analysis of 5,373 asymptomatic people found critical or indeterminate incidental findings in 32.1% of whole-body MRI studies, with a pooled false-positive proportion of 16% — and not one study verified its negatives beyond five years [S4]. A separate meta-analysis of over 9,000 people put the confirmed-cancer detection rate at just 1.57% [S5]; at a base rate that low the positive predictive value of a flagged finding is poor, so most positives are not the disease being sought. The disciplined response is to govern the modality, not abandon or oversell it — structured reporting through ONCO-RADS, explicit guardrails for who is scanned, and the performance metrics and ROC curves we already hold breast screening to, so a whole-body programme can finally be measured and audited [S17]. That governance is what belongs on the ledger.

The distinction that matters for anyone allocating capital: opportunistic biomarkers add information to a scan you were always going to do, at near-zero marginal imaging cost and radiation. Elective whole-body screening adds a scan, a cost, and an incidental-finding burden. They are not the same investment case, and they should not be assessed with the same number.

What I'd watch next

The bottleneck is no longer the algorithm — most of these measures are automatable and independently validated [S9][S11]. It is three unglamorous things: a workflow that surfaces the second-read without slowing the first, a reimbursement line that pays for extracting value already latent in the image, and a governance framework that decides who is accountable when an opportunistic finding is flagged and missed. The organisations that solve the plumbing, not the model, will own this.

References

Written by Dr Lisa Sorger — consultant radiologist, healthcare executive, medical administrator, company director and founder of myradiologist.ai. Every figure in this issue is sourced to the public record; the source register is available on request.

Next issue: Value in radiology — why the specialty's worth is measured by the wrong yardstick.

All views expressed here are my own personal opinions and are not medical advice. General information only — not clinical or financial advice. myradiologist.ai · ABN 29 692 758 115 · ACN 692 758 115