Abstract :
[en] Early detection of significant prostate cancer (sPCa) has notably improved with the widespread use of magnetic resonance imaging (MRI) and targeted biopsies. This advancement has contributed to the recommendation for PCa screening in the European Union. Nevertheless, PCa suspicion is still primarily based on a serum prostate-specific antigen (PSA) level greater than 3.0 ng/mL and/or a suspicious digital rectal examination (DRE). Prostate MRI is then requested to identify lesions suspected of harbouring sPCa, particularly when the Prostate Imaging-Reporting and Data System (PI-RADS) score is above 2. In such cases, both targeted biopsies of suspicious lesions and systematic biopsies are recommended, whereas prostate biopsy is typically avoided when the PI-RADS score is below 3. The European Association of Urology recommends the use of predictive models to reduce unnecessary prostate biopsies and the over-detection of insignificant prostate cancer (iPCa), especially in uncertain scenarios as PI-RADS 3 where the mean rate of sPCa detection is 20% and the overdetection of iPCa remains up to 50%.1 Accessible web- or smartphone-based risk calculators are essential for integrating predictive models into routine practice for assessing individual sPCa likelihood. Furthermore, validating these predictive models in populations different from those used in their development is crucial to ensure generalizability and clinical reliability.
Name of the research project :
Federated Learning and mUlti-party computation Techniques for prostatE cancer (FLUTE)
Scopus citations®
without self-citations
0