A Randomized Trial of Early Detection of Clinically Significant Prostate Cancer (ProScreen) (ProScreen)

Helsinki University Hospital, Finland, Tampere University Hospital, Finland, Finnish Cancer Registry, Finland, University of Turku, Finland, Lund University, Fimlab Laboratories, Finland, Laboratory HUSLAB, Finland, University of Helsinki, Hospital District of Helsinki and Uusimaa, Clinical Research Institute HUCH Ltd, Finland

A population-based randomised trial of prostate cancer screening will be carried out. A total of approximately 117,200 men aged 50-63 in Helsinki and Tampere are randomised to intervention (screening) or control arm. A reduction in harms of screening in the form of overdiagnosis is sought, while retaining as much as possible of the mortality benefit (reduction in prostate cancer mortality). Novel methods that have been shown to increase specificity for clinically relevant prostate cancer but never tested in a randomised setting will be employed in screening and diagnostics. The main end-point is prostate cancer mortality at 10 and 15 years of follow-up.

Frequent adverse effects have so far tipped the balance of benefits and harms against prostate cancer screening, and therefore the investigators will focus on employing the best possible means for reducing them. The project introduces a novel concept for PC screening that minimises overdiagnosis and overtreatment, while retaining the mortality benefit to shift the balance of screening benefits and harms to a favourable net effect. The strategy for implementation as a randomised screening trial utilises three levels of risk assessment (PSA, kallikrein panel and MRI) before the diagnostic procedure (prostate biopsy), each aimed at eliminating detection of indolent disease. The study hypothesis is that by virtue of the novel three-tiered screening algorithm, the beneficial screening effect (prostate cancer mortality reduction) can be retained, while the overdiagnosis can be largely eliminated. The impact of an integrative approach has never been evaluated - each of the methods has only been assessed in isolation. The breakthrough potential of the proposal lies in combining the three novel approaches and taking them to the forefront of applied research through a randomised trial. The key impact of the study is in defining whether the overall balance of benefits and harms of prostate cancer screening can be reversed by applying the best possible methods to detect only clinically important disease. If the study hypothesis is affirmed, it opens the way to introduction of prostate cancer screening. If the balance of harms and benefits is still unfavourable, the problem of overdiagnosis in prostate cancer may be intractable.

Prostate cancer, Screening, Randomised trial, Kallikrein, Magnetic resonance imaging, Prostate-specific antigen

Of the approximately 117,200 men aged 50-63 years resident in Helsinki and Tampere regions, a fourth will be randomised to screening and the rest to control arm (after exclusion of prevalent cases).

Depending on each diagnostic test result the participants in the screening arm will undergo PSA-testing, 4Kscore determination, MRI, and MRI/US fusion biopsy only.

An intention to screen analysis will be performed, with all men in the groups defined by random allocation, regardless of compliance. Follow-up starts at randomisation, and ends at death. Cox regression will be used with prostate cancer death as the outcome.

A secondary analysis of prostate cancer mortality will be performed using the Cuzick method with correction for contamination and selection bias due to non-compliance.

Intermediate outcomes include cumulative incidence of advanced (T3-T4 or M1) prostate cancer (number of cases relative to population size, not using incidence density to avoid the lead-time bias due to early detection by screening).

Intermediate outcomes include cumulative incidence of low-risk cancer (Gleason<7) as an indicator of overdiagnosis.

Diagnostic performance of 4Kscore among men with PSA>3 in terms of predictive values, sensitivity, and specificity for clinically significant PrCa (defined as Gleason 7+ cancers - including those diagnosed within the next four years for test-negative non-biopsied men i.e. false negatives).

Diagnostic performance of MRI based on PI-RADS v2 scores among men with PSA>3 and positive 4Kscore in terms of predictive values, sensitivity, and specificity for clinically significant PrCa (defined as Gleason 7+ cancers - including those diagnosed within the next four years for test-negative non-biopsied men i.e. false negatives).

Health-related quality of life among men diagnosed with prostate cancer will be assessed using the EPIC 26 instrument enrolling screen-detected and interval cases, as well as those diagnosed among non-participants and in the control arm.

Anxiety among men diagnosed with prostate cancer will be assessed using the Memorial Anxiety Scale for Prostate Cancer (MAX-PC), enrolling screen-detected and interval cases, as well as those diagnosed among non-participants and in the control arm.

Adverse effects of biopsy are evaluated using a questionnaire on complications including assessment of bleeding, lower urinary tract symptoms (LUTS), erectile dysfunction (ED), pain, and antibiotic treatment immediately after the biopsy.

Adverse effects of biopsy are evaluated with a questionnaire on complications, including assessment of bleeding, lower urinary tract symptoms (LUTS), erectile dysfunction (ED), pain, and antibiotic treatment 30 days after the biopsy.

Economic evaluation will commence with cost analysis, and the final analysis of incremental cost effectiveness ratio (with a decision analysis) will be conducted once data on both long-term cost and real outcome data on both utilities (quality-adjusted life-years) and mortality are available.

Inclusion Criteria: 50-63-year-old men (age in 2018) residing in Tampere or Helsinki Exclusion Criteria: Prevalent prostate cancer

The data will be managed and made openly available with Finnish Social Science Data Archive (FSD) and publisher websites. The type of data includes numerical data, mainly counts, and continuous variables classified as categorical. The data will be uploaded in tabular form (as a data matrix). The data will be deidentified and supplied only as frequencies. The released data will be available under CC license for research purposes (accessible for registered users of FSD database). Citation of the original research is needed whenever used by third parties.