A blood-based screening test that combines prostate-specific antigen with protein biomarkers, genetic risk markers, and clinical factors detected more clinically significant prostate cancers within 2 years than prostate-specific antigen testing alone, according to a secondary analysis of the STHLM3-MRI trial.
Researchers evaluated 12,670 men aged 50 to 74 years who underwent both prostate-specific antigen testing and Stockholm3 screening between 2018 and 2020 in the Stockholm region of Sweden. Men were excluded from the trial if they had prior prostate cancer, a biopsy within 60 days, magnetic resonance imaging contraindication, or severe comorbidity.
The primary outcome was clinically significant prostate cancer, defined as grade group 2 or higher, diagnosed within 2 years of the screening blood test. Outcomes were assessed through linkage to the Swedish National Cancer Register and the Stockholm Prostate Cancer Diagnostics Register.
In the original trial design, men with either a prostate-specific antigen level of 3 ng/mL or greater or a Stockholm3 score of 11 or greater were referred for further workup and randomly assigned to systematic biopsy or magnetic resonance imaging with systematic and targeted biopsies for lesions with a Prostate Imaging Reporting and Data System score of 3 or higher. Men with negative screening results did not routinely undergo baseline biopsy, and the investigators used 2-year registry follow-up to identify clinically significant cancers diagnosed following the baseline screening test.
During follow-up, 443 men were diagnosed with clinically significant prostate cancer. Stockholm3, using the prespecified threshold of 11 or greater, detected 400 of those cancers, yielding a sensitivity of 90%. Prostate-specific antigen testing at a threshold of 3 ng/mL or greater detected 327 clinically significant cancers, corresponding to a sensitivity of 74%.
Stockholm3 missed 43 clinically significant cancers, compared with 116 missed by prostate-specific antigen testing. The false-negative rate was 10% with Stockholm3 and 26% with prostate-specific antigen testing. Specificity was similar between the 2 screening approaches, at 89% and 90%, respectively. False-positive rates were also similar, at 11% and 10%.
Per 1,000 men screened, Stockholm3 detected about 32 clinically significant cancers compared with about 26 detected by prostate-specific antigen testing. Stockholm3 produced about 102 false-positive results per 1,000 men screened, compared with about 95 false-positive results per 1,000 men screened with prostate-specific antigen testing.
Decision curve analysis showed greater clinical net benefit for Stockholm3 compared with prostate-specific antigen testing across biopsy decision thresholds up to approximately 25% predicted risk. The researchers interpreted this higher net benefit as reflecting fewer missed clinically significant cancers and fewer unnecessary biopsies across those thresholds.
The researchers also reported a modest increase in detection of low-risk disease. Stockholm3 identified 164 grade group 1 cancers, compared with 145 detected by prostate-specific antigen testing. They noted that this increase may reflect a tradeoff between greater detection of clinically significant disease and additional downstream testing or surveillance for low-risk diagnoses.
Sensitivity analyses using alternative thresholds showed a similar pattern. At a Stockholm3 threshold of 15 or greater, sensitivity for clinically significant prostate cancer was 75% and specificity was 95%. At a prostate-specific antigen threshold of 4 ng/mL or greater, sensitivity was 52% and specificity was 95%. With age-adjusted prostate-specific antigen thresholds, sensitivity was 42% and specificity was 96%.
Several limitations should be considered. Participation among invited men was approximately 25%, which may limit generalizability. Baseline magnetic resonance imaging and biopsy were performed only among men with positive screening results, and only 9.1% of participants underwent biopsy. As a result, the design could not provide an unbiased estimate of baseline diagnostic accuracy across the entire cohort. The study relied on registry linkage to identify clinically significant cancers during follow-up, but some undetected cancers may have been missed.
Follow-up was limited to 2 years, preventing assessment of disease progression, prostate cancer mortality, overall survival, and long-term cost-effectiveness. The cohort was also predominantly Swedish or European, although the researchers noted that a large multiethnic study including Asian, Black, and Hispanic men found similar Stockholm3 performance.
Implementation may also depend on cost and health system context, according to the researchers. They reported that US pricing has not been established and that European pricing has ranged from $550 to $750. They noted that whether reduced missed clinically significant cancers or fewer unnecessary biopsies offset the added cost will depend on downstream resource use and health system priorities.
The researchers concluded that Stockholm3 may offer a more favorable short-term balance between detection of clinically significant prostate cancer and additional diagnostic evaluation than prostate-specific antigen testing alone in a screening setting. They noted that longer follow-up is needed to determine whether improved short-term detection translates into better long-term clinical outcomes or cost-effective screening implementation.
“These findings support the use of Stockholm3 in the context of a risk-adapted screening approach, enabling more precise identification of men at higher risk for clinically significant disease while reducing unnecessary biopsies,” wrote lead study author Thorgerdur Palsdottir, PhD, of Karolinska Institutet in Stockholm, Sweden, and colleagues.
Disclosures: The study was funded by the Swedish Research Council, Swedish Prostate Cancer Society, Stockholm Region, and the Swedish Cancer Society. The funders had no role in the study design or the decision to submit the study for publication. Disclosure forms are available with the article online.
Source: Annals of Internal Medicine
