618 Korean men had a prostate MRI before any PSA test. Testing PSA first would have made 90 percent of those scans unnecessary

Scanning first and testing blood afterwards sounds like the more thorough order, and in trial settings in North America and Europe upfront MRI for prostate cancer screening has looked promising. This is what happened when a Korean centre did it outside a trial, on men who had walked in for a health check.

What the scans found#

The pyramid narrows sharply. 535 men were PI-RADS 1 or 2 and stopped there. 83 were PI-RADS 3 or above and were sent for a biopsy conversation; 66 went through with it; 9 had cancer that mattered. That leaves 57 men biopsied and found not to have significant cancer — more than six negative biopsies for every cancer detected.

The comparison that undoes the strategy#

Because PSA was drawn at the same visit, the authors could reconstruct what a conventional pathway would have done with the same 618 men: scan only those with a PSA above 3.0 ng/mL. That pathway would have skipped 556 of the 618 scans, cancelled 58 of the 83 biopsy consultations and prevented 40 of the 57 biopsies that found nothing — and missed one of the nine cancers.

The median PSA in this group was 0.98 ng/mL. That number explains the result. Most of the men being scanned had no biochemical reason to be scanned, so the MRI was working on a population where the prior probability of disease was very low — which is exactly the setting in which any test produces more false alarms than findings.

Why the Korean setting changes the answer#

The authors conclude that the upfront-MRI pathway is inefficient in Korean men undergoing general health evaluations and that conventional PSA-gating should continue. Their closing point is the more general one: population-level differences should be considered carefully before extrapolating trial results to other settings.

That caution has teeth here. Korean national data shows roughly half of prostate cancer is high-risk at diagnosis and the low-risk share has been falling — a detection profile unlike the Western screening populations the upfront-MRI trials were run in. A strategy designed to find more of the small disease that Korea is not finding much of, deployed in a check-up population with a median PSA under 1.0, is being asked to solve a problem this population does not have.

The design limits are worth keeping in view: a single centre, a self-selected group who chose and paid for a scan, and a retrospective comparison against a simulated pathway rather than a randomised one. What it cannot be dismissed as is theoretical. 618 men had the scan, and the arithmetic of what it cost them is on the record.