Requirements Federated Learning and mUlti-party computation Techniques for prostatE cancer
0.1.0 - ci-build
Requirements Federated Learning and mUlti-party computation Techniques for prostatE cancer, published by HL7 Europe. This guide is not an authorized publication; it is the continuous build for version 0.1.0 built by the FHIR (HL7® FHIR® Standard) CI Build. This version is based on the current content of https://github.com/hl7-eu/flute-requirements/ and changes regularly. See the Directory of published versions
Prostate Cancer (PCa) can be aggressive or indolent. Indolent forms can exist for a long time without causing symptoms or death, while aggressive forms often present symptoms and lead to cancer-specific mortality. At the present time, no consensus has been reached regarding the indolent or aggressive characteristics of PCa. The standard indicators for suspicion of PCa are an increase of serum prostate specific antigen (PSA) and/or an abnormality of digital rectal examination (DRE). This has led to an excessive number of unnecessary Prostate Biopsies (PBx) and large overdetection rates of indolent tumors, causing overtreatment of PCa. This is the main reason why active surveillance programs have been extended, as currently recommended in up to 30% of newly diagnosed tumors. Since the PCa screening crisis, due to the absence of mortality reduction observed in an American screening trial, there have been some improvements in the early diagnostic process of PCa:
VHIR developed a methodology to diagnose csPCa, based on the lessons learned from the ESRPC (European Randomized Study of Screening for Prostate Cancer). The PLCO (Prostate, Lung, Colorectal and Ovarian) test and the justification of its initial results have also recently been properly interpreted in the context of time bias due to the high contamination rate of its control group . Finally, the increase in the diagnosis of advanced and metastatic tumors and subsequent mortality observed since the recommendation about not to screen for PCa has contributed to considering the PCa screening as a current and urgent need.
Once PCa is diagnosed, TNM staging is performed and, in the case of clinically localized carcinomas (T1-2 N0 M0), the risk group is established, usually through the criteria established by the NCCN (National Comprehensive Cancer Network) or the EAU (European Association of Urology). These criteria are based on clinical stage T, serum PSA level and ISUP grade. Thus, today we are still in a scenario in which the prediction of tumor aggressiveness is based on three classic clinical parameters, TNM, Gleason grade and PSA serum level.
As patients with the same histological and clinical PCa parameters can have strikingly different molecular profiles, it appears rather obvious that genomic PCa biomarkers, especially those that can outperform/complement clinical and pathological prognosis factors, should be included as one more parameter in the predictive tools and commercially available tools (Depicher, Oncotype DX and Prolaris) have shown to improve PCa risk stratification.
The need for personalization in early detection and diagnosis of prostate cancer: Given that PCa is a heterogeneous disease with a wide variability, it is essential to provide a personalized approach for early detection, disease status (indolent or aggressive) and prediction of treatment response. A precise risk assessment will help to reduce the burden of biopsies in men who are believed to be at high PCa risk. Beside diagnosis, biomarkers should be used in disease monitoring, to avoid treatments after considering PCa risk, general health and age. Moreover, there is an urgent need to increase research efforts to identify those castrate-resistant prostate cancer (CRPC).