Reducing MRI Use in Organized Prostate Cancer Testing Using Blood-based Biomarkers: The OPT Stockholm3 Study.

BACKGROUND AND

OBJECTIVE

We assessed the feasibility of implementation, reduction in magnetic resonance imaging (MRI) use, and cost implications of integrating Stockholm3 as a reflex blood test in organized prostate cancer testing (OPT).

METHODS

This pragmatic, population-based implementation study included men aged 50-52 yr invited to OPT in 2023 (n = 17 801) and 2024 (n = 30 556) in Health Region Stockholm-Gotland. The experimental group (2024) had a Stockholm3 test if prostate-specific antigen (PSA) ≥2 ng/ml and an MRI if Stockholm3 ≥15%. The control group (2023) followed standard OPT (MRI if PSA ≥3 ng/ml). Outcomes were MRI use (primary), biopsies performed, detection of Grade Group (GG) 1 and GG ≥2 cancer, and Stockholm3 cost thresholds. Relative proportions (RPs) and decision-tree cost modeling were used.

KEY FINDINGS AND LIMITATIONS: MRI use decreased by 64%, from 2.7% (181/6663) in 2023 to 1.0% (70/7070) in 2024 (RP 0.36; 95% confidence intervals [CI], 0.28-0.48); biopsy rates decreased from 0.6% to 0.4% (RP 0.67; 95% CI, 0.41-1.07). Detection of GG 1 cancer decreased from 0.17% (11/6663) to 0.08% (6/7070) (RP 0.51; 95% CI, 0.19-1.39), and detection of GG ≥2 cancer decreased from 0.20% (13/6663) to 0.14% (10/7070) (RP 0.72; 95% CI, 0.32-1.65). A cost of €92 per Stockholm3 test would achieve cost neutrality. Limiting Stockholm3 testing to men with PSA ≥3 ng/ml would reduce MRI use and costs. Study limitations include the nonrandomized design, few cancer events, restricted age range, and short follow-up.

CONCLUSIONS AND CLINICAL IMPLICATIONS: Integrating Stockholm3 as a reflex test in the first round of an OPT program is feasible. Reflex testing at PSA ≥2 ng/ml and PSA ≥3 ng/ml was associated with reductions in the numbers of MRI scans and biopsies performed. No cancers were detected at PSA 2.0-2.9 ng/ml; however, the study was underpowered to assess cancer detection, and this finding should be considered hypothesis-generating.