# Prostate cancer screening in men aged 50-69 years (STHLM3): a prospective population-based diagnostic study

Source: https://onco.cc/key-papers/paper-gronberg-sthlm3-lancet-oncol-2015/  
OnCo record `paper-gronberg-sthlm3-lancet-oncol-2015` (Key paper). Data CC BY-NC 4.0, attribute "Data from OnCo (onco.cc)"; commercial use needs a licence.

## TL;DR

A Stockholm study that combined a PSA test with other blood proteins, 232 genetic markers and clinical details into one risk score. At the same sensitivity as PSA alone, the score would have avoided about a third of prostate biopsies.

## Summary

STHLM3 was a prospective, population-based, paired, screen-positive diagnostic study of men aged 50 to 69 without prostate cancer, invited at random by date of birth from the Swedish population register. Every participant had both PSA and the predefined STHLM3 model: plasma proteins (PSA, free PSA, intact PSA, hK2, MSMB, MIC1), 232 single-nucleotide polymorphisms, and age, family history, previous biopsy and prostate examination. The aim was higher specificity than PSA without lower sensitivity for Gleason 7 or higher cancer. It was funded by Stockholm County Council and registered as ISRCTN84445406.

## Fields

- Kind: Key paper
- Last checked: 2026-11-03
- Journal: Lancet Oncology
- Year: 2015
- DOI: 10.1016/S1470-2045(15)00361-7
- Authors: Grönberg H, Adolfsson J, Aly M, Nordström T, Wiklund P, Brandberg Y, Thompson J, Wiklund F, Lindberg J, Clements M, Egevad L, Eklund M
- Findings: Area under the curve for Gleason 7 or higher cancer was 0.74 (95 per cent CI 0.72 to 0.75) for the STHLM3 model against 0.56 (0.55 to 0.60) for PSA alone (p<0.0001).; At the same sensitivity as a PSA cutoff of 3 ng/mL or more, the model could reduce the number of biopsies by 32 per cent (95 per cent CI 24 to 39) and avoid 44 per cent (35 to 54) of benign biopsies.
- What it means: It showed in a general population, not a referred clinic population, that a multi-marker risk score could spare many men a biopsy without missing more of the cancers that matter. The same group then tested the score inside the STHLM3-MRI screening trial.
- Caveats: A diagnostic accuracy study, not a trial of mortality; men with a negative score were not all biopsied, so cancers missed by both tests cannot be counted.; Developed and validated in Stockholm men; performance in other populations needs its own validation.

## Sources

- Lancet Oncology 2015: https://doi.org/10.1016/S1470-2045(15)00361-7
- PubMed: https://pubmed.ncbi.nlm.nih.gov/26563502/

## Connected records

- trials: [STHLM3-MRI (Stockholm3 and MRI-targeted biopsy in prostate cancer screening)](https://onco.cc/trials/sthlm3-mri/)
- cancers: [Prostate cancer](https://onco.cc/cancers/prostate/)
- fronts: [Early Detection & Screening](https://onco.cc/fronts/early-detection/)
- institutions: [Karolinska University Hospital](https://onco.cc/institutions/karolinska/)
- terms: [Overdiagnosis](https://onco.cc/terms/overdiagnosis/), [Polygenic risk score (PRS)](https://onco.cc/terms/polygenic-risk-score/), [PSA (prostate-specific antigen)](https://onco.cc/terms/psa/)
- people: [Henrik Grönberg](https://onco.cc/people/henrik-gronberg/), [Martin Eklund](https://onco.cc/people/martin-eklund/)
- bottlenecks: [Biomarkers are not validated or standardised](https://onco.cc/bottlenecks/b-biomarker-validation/), [The hardest cancers are found late](https://onco.cc/bottlenecks/b-early-detection/)
- journals: [The Lancet Oncology](https://onco.cc/journals/lancet-oncology/)

---
JSON: https://onco.cc/api/v1/entities/paper-gronberg-sthlm3-lancet-oncol-2015.json