A blood test spared one in five surveillance biopsies without missing a significant upgrade

Design

prospective multicentre diagnostic accuracy trial

Population

199 men with ISUP grade group 1 prostate cancer on active surveillance in Sweden, Denmark and Norway, 2019-2022

Primary outcome

upgrading to ISUP grade group 2 or higher at surveillance biopsy

Effect

36% upgraded; Stockholm3 AUC 0.71 (95% CI 0.64-0.79); at threshold ≥15, sensitivity 0.93, NPV 0.87, 19% of biopsies spared, 7% of upgrades missed with none ISUP ≥3

Active surveillance for grade group 1 prostate cancer depends on repeated biopsy, which is the part patients dislike and drop out over. STHLM3-AS NorDCaP enrolled 199 men on surveillance across Sweden, Denmark and Norway between 2019 and 2022, all with grade group 1 disease, and gave each a Stockholm3 blood test before their surveillance biopsy. The outcome was upgrading to grade group 2 or above.

Thirty-six percent (72 of 199) were upgraded. Stockholm3 scores were higher in those upgraded (median 35, IQR 26-50) than in those not (23, IQR 14-34, P<0.01), with an area under the curve of 0.71 (95% CI 0.64-0.79). Using a threshold of 15, sensitivity was 0.93 and negative predictive value 0.87, sparing 19% of biopsies and missing five upgrades - 7% of them - none of which was grade group 3 or above. Dropping the threshold to 11 raised sensitivity to 0.99 but spared almost nothing. For comparison in the same men, PSA density ≥0.15 had sensitivity 0.60 and PI-RADS ≥3 had sensitivity 0.92, with areas under the curve of 0.71 and 0.72 - statistically indistinguishable from the blood test.

That comparison is the honest framing, and the authors give it: Stockholm3 did not beat MRI or PSA density, it matched them. Its value is that it is a blood test, which in a surveillance programme constrained by MRI access - as most Indian programmes are - is a different kind of usefulness from being more accurate. One in five biopsies avoided is a real gain for a man facing a decade of surveillance, and missing 7% of upgrades, all of them grade group 2, is a trade-off worth putting to him rather than deciding for him.

• Consider a blood-based risk score as part of a multimodal surveillance strategy where MRI access is the constraint
• Present the trade-off explicitly: about one biopsy in five avoided, about 7% of upgrades missed, none of them grade group 3 or above
• Do not drop MRI or PSA density in favour of it - they performed equivalently in the same men
• Note that 36% of men on surveillance were upgraded at biopsy, which is the more sobering number in this trial
• Check availability and cost locally before raising the test with a patient

Why it matters

It offers a way to reduce biopsies in men who will be on surveillance for years, using a blood sample rather than a scanner slot.

Don't overread it

One hundred and ninety-nine men, a single biopsy timepoint, and performance no better than PSA density or MRI in the same cohort.

The statistics, in plain English

An area under the curve of 0.71 is modest discrimination - better than a coin toss, well short of a decisive test - and it is the same figure PSA density and MRI achieved here, so no one test is winning. A negative predictive value of 0.87 sounds reassuring but depends entirely on the 36% upgrade rate in this cohort; in a population where fewer men were upgraded, the same test would look better, and where more were, worse. The five missed upgrades are a small enough number that 7% could easily be 3% or 13% in the next cohort.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.