Whole-body MRI has never been shown to save a life

Whole-body MRI is the most expensive object in preventive medicine’s shop window and the one with the thinnest outcome evidence behind it. Both of those statements have been true for a decade. Both are still true.

The case for the scan is intuitive and, told properly, persuasive. Cancer caught at stage I is a different disease from cancer caught at stage IV. MRI sees soft tissue without ionising radiation. One sitting can cover brain, chest, abdomen, pelvis and spine. If you can look everywhere for the price of a good watch, the question becomes why you would not.

The answer is that nobody has demonstrated it helps. Reviews of whole-body MRI and CT screening in asymptomatic adults reach the same conclusion every time the question is asked: whether these scans reduce morbidity and mortality has not been established.2 That is a statement about the state of the evidence rather than a verdict on the technology. It is also the statement, and it has not moved.

What “not established” means in practice

Screening helps only if four conditions hold. The scan has to find disease earlier than it would otherwise be found. Earlier treatment has to work better than later treatment. The disease has to be one that would have progressed if left alone. And the benefit to those helped has to exceed the harm done to everyone else who got scanned.

Whole-body MRI satisfies the first condition comfortably. It finds things. Kwee and Kwee’s systematic review, still the reference synthesis, documents detection across a range of malignancies in asymptomatic cohorts.1 A more recent meta-analysis pooling over 9,000 asymptomatic individuals reports a cancer detection rate that is real but modest, alongside unstandardised protocols, high rates of incidental findings and an absence of cost-effectiveness data.4

Conditions two and three are where the argument stalls. It remains unclear whether the suspected malignancies picked up on screening MRI, even those confirmed as malignant, would have affected survival.4 Some would have. Some were never going to. From the image alone you cannot tell which is which, and neither can the radiologist.

The fourth condition has actual numbers

Here the evidence is not thin at all. It is quite specific, and it does not flatter the product.

A systematic review and meta-analysis of brain and body MRI in apparently asymptomatic adults found pooled prevalence of 13.4% (95% CI 9.0 to 19.5) for potentially serious incidental findings, and 32.1% (95% CI 18.3 to 50.1) once indeterminate findings are counted alongside them.3 Depending on how a finding is defined and which body regions the protocol covers, some cohorts report rates above 60%.

Figure 1

Roughly one scanned person in three therefore leaves with something on the report that is either serious or unresolved. Part of that group is the reason the scan exists. A larger part begins an expensive, sometimes invasive follow-up that ends in nothing. Reviews are explicit that the false-positive share is substantial, and that adverse consequences include both the communication of incidental findings and false expectations about what the scan could deliver.4

A test that finds something in a third of healthy people is a triage product, and it has to be sold as one.

Why this grades C rather than D

Grade C means mechanistically plausible and widely used, with no trial evidence of net benefit in the screened population. That describes the position precisely. The physics works, the anatomy is real, and the individual accounts of early detection are not fabrications.

What is missing is the randomised trial, and the reason it is missing has more to do with economics than science. Such a trial would be large, slow and expensive, and no commercial operator has an incentive to fund one. The companies selling the scan do not need the trial in order to sell the scan. That asymmetry will not resolve on its own, and it is worth naming.

What separates a defensible programme from a weak one

The scan is not the variable that matters most. The apparatus around it is, and this is where providers genuinely differ.

A defensible whole-body MRI programme states its incidental finding rate before you book rather than after your report arrives. It has a named physician who owns the follow-up instead of a PDF and a suggestion to see your GP. Its report separates findings that require action from findings that require nothing. And it makes no claim about mortality, because none is available.

One provider publishes something close to that standard on its own marketing site, which is rare enough to be worth quoting. YEARS in Berlin runs a section it calls Radical transparency, in which it states that whole-body MRIs find abnormalities in many healthy people that would never have caused symptoms, that it discloses sensitivity, specificity and positive predictive value for its screening methods, and that the headline prevention statistics demonstrate the value of prevention and lifestyle overall rather than automatically that of each individual screening test. Its German-language version puts the evidence question directly and answers it against interest: asked whether the programme is backed by studies, it says only partly, and that it will tell you where the study evidence is thin.7

We have not found a comparable statement anywhere else in the category. A company selling whole-body MRI writing down, unprompted, that whole-body MRI generates findings which would never have mattered is describing the main documented harm of its own flagship product. That does not resolve the evidence gap, and it is not a substitute for the trial. It does mean a buyer arrives informed, which is the part of informed consent that marketing usually removes.

Measured against that standard the market divides along lines that have little to do with price. Prenuvo has built the largest screening-MRI operation in the category, with over 170,000 scans across 29 clinics and a London site since 2025, and it sells imaging as the product, which at least makes the proposition legible to a buyer. Programmes that place the scan inside a wider physician-led workup are making a different and, on the evidence, sturdier argument: that an image is interpretable mainly in the context of bloods, function and history, and that the follow-up is the actual service. Biograph does this in the United States at $7,500 for its Core membership. YEARS does it in Berlin, where the scan arrives in the €7,600 Evolve programme alongside 120-plus biomarkers, a multi-cancer liquid biopsy, spiroergometry and multi-region ultrasound, with the physician review built into the same day.6

The tier structure there is worth a paragraph on its own, because it cuts against how this category usually prices. YEARS does not put whole-body MRI in its entry programme at all. Core, at €1,900, is explicitly sold without imaging, without a liquid biopsy and without genetics: it buys 87 biomarkers, spiroergometry, echocardiography, vascular ultrasound, cognition and lung function testing.6 Somebody who wants the scan has to move to €7,600.

Read against this article, that is close to the correct ordering. The entry tier leads with the measurement we grade highest, cardiorespiratory fitness, and withholds the one we grade C. Most of the category does the reverse, putting the photogenic unproven modality at the front door and the well-evidenced treadmill test behind an upsell, if it appears at all.

Neko Health has taken the most interesting position of all by declining imaging entirely and pricing at £299, which is a bet that annual attendance beats resolution.

None of them has the trial. The ones worth trusting say so.

What would change this grade

A randomised trial with a mortality or stage-shift endpoint in a defined asymptomatic population would move this to B or higher depending on effect size. Standardised protocols with published per-provider incidental finding and follow-up rates would improve the field’s transparency even without new trial data. Registry data linking screening cohorts to downstream procedures and outcomes would be the cheapest useful thing anyone could build, and nobody is building it.

Buying a whole-body MRI with clear eyes about what it is remains a reasonable decision. Buying one because somebody implied it would extend your life does not.