A two-gene urine test aims to cut needless prostate biopsies in the PSA grey zone

The prostate-specific antigen test has a well-known weakness: in the 4–10 ng/mL range that clinicians call the diagnostic grey zone, a raised PSA says little about whether a man actually has cancer, so many are sent for a biopsy that turns out to be unnecessary [s1]. A new study reports a urine test meant to narrow that grey zone — and reaches strong numbers in a prospective validation while carrying the usual caveats of an early diagnostic [s1].

What the test is

The assay, described in Molecular Biomedicine, measures DNA methylation — chemical marks on genes that differ between cancerous and healthy prostate tissue — in urine collected after a prostatic massage, so no biopsy is needed to run it [s1]. The developers started from tissue-based sequencing and large public methylome datasets to find prostate-cancer-specific methylation regions, narrowed the candidates with targeted sequencing, and converted the survivors into a simple quantitative-PCR readout [s1]. In a retrospective development cohort of 348 participants, stepwise selection and logistic-regression modelling produced a compact two-gene assay built on markers the authors label ADD3_2 and GSX2_2 [s1].

What it found

The model was then tested in an independent, prospective, multicentre cohort of 200 participants [s1]. It achieved an area under the curve of 0.86 in the development cohort and 0.91 in the validation cohort, while keeping specificity above 90% [s1]. Crucially, the authors report that performance held up in the PSA grey zone and for clinically significant cancer — the aggressive disease that actually needs treating, as opposed to the indolent kind that drives overtreatment [s1].

The practical claim comes from a decision-curve analysis, which the authors say indicates that using the tool could safely avert approximately 55% of unnecessary biopsies without compromising cancer detection [s1]. That is the number to hold on to, because reducing negative biopsies is precisely the job a grey-zone test exists to do. It is also the number that most needs independent confirmation, because a decision-curve estimate is a modelled projection, not an outcome counted in patients who were spared a biopsy and then followed.

Two design choices give the headline figures more weight than most early biomarker reports earn. Collecting urine after a prostatic massage concentrates shed prostate cells, and reading the two genes on a quantitative-PCR platform rather than full sequencing is the kind of simplification that makes a test cheap and reproducible enough to deploy widely [s1]. The specificity above 90% is the other half of the grey-zone problem: a test that rules cancer out has to do so without a flood of false positives of its own, or it simply trades one unnecessary biopsy for another [s1]. That the assay held its discrimination specifically for clinically significant cancer — rather than leaning on the detection of indolent disease to inflate its numbers — is the result most relevant to whether it could reduce overtreatment rather than add to it [s1].

How to read the result

Two framings keep this honest. First, the headline AUC of 0.91 comes from a validation cohort of 200 people, and the model that produced it was tuned on a separate, retrospective set of 348 [s1]. That is the right design — building and testing on different data — and prospective collection is a genuine strength over the retrospective datasets much biomarker work relies on. But 200 is a modest sample, and a single validation at the centres that developed a test tends to flatter it relative to how it will perform elsewhere.

Second, this is a crowded field, and a new entrant has to be judged against incumbents, not against PSA alone. A 2026 systematic review in BJUI Compass set out a framework to compare the five biomarkers already used to help men with a raised PSA decide on biopsy — the prostate health index, 4KScore, SelectMDx, ExoDx and MyProstate Score — and noted that only some have been validated in Black populations, a group at higher prostate-cancer risk [s2]. A test that matches or beats those on accuracy still has to prove it does so across the populations that will use it, and the new assay's validation does not yet speak to that [s1][s2]. The same discipline applies here as to other screening tools this site has covered: a number on a curated cohort is an upper bound on real-world performance, not a promise of it.

What to watch

Whether the assay is validated prospectively at centres that did not build it, and in diverse populations, before it is marketed as a biopsy-sparing tool. Whether the roughly 55% biopsy reduction survives contact with real clinical decisions, where a clinician's willingness to act on a negative test is its own variable [s1]. And whether, like so many promising diagnostics, it arrives with a worked-out pathway for the men it flags — because a test that averts biopsies is only a benefit if the cancers it is trusted to rule out really are being ruled out.

This article is informational and is not medical advice.

Sources

• [s1] "Development and validation of a two-gene urine DNA methylation assay for noninvasive prostate cancer detection: a prospective multicenter study." Molecular Biomedicine, published online 7 August 2026. https://doi.org/10.1186/s43556-026-00528-y
• [s2] "Comparing diagnostic prostate cancer biomarkers using a framework of clinically meaningful characteristics." BJUI Compass, published online 22 September 2026. https://doi.org/10.1002/bco2.70275