Development of artificial intelligence-based multiplex network for individualized risk stratification of prostate cancer

Background Prostate cancer represents the most common neoplasia among men in Western countries, having a substantial impact on patients' lives and on healthcare systems. Its wide range of clinical aggressiveness requires an accurate disease risk stratification to tailor the treatments. Current classification tools' accuracy is still sub-optimal, especially considering low- to intermediate-risk patients, potentially resulting in over- and under-treatment. To date, single-biomarker or oligo-genic approaches have not shown a substantial predictive benefit and are not part of the routine practice. A promising prognostic and predictive role of artificial intelligence (AI) - based integrated platforms has emerged for other malignancies (notably, the ARIADNE platform for the triple-negative breast cancer profiling). To date, the role of these approaches in prostate cancer management is unexplored. Hypothesis We hypothesize that an artificial intelligence-based platform, integrating clinical, pathologic, imaging, genomic and transcriptomic profile of prostate cancer would outperform currently available risk-stratification tools, leading to a better definition of cancer progression and recurrence risk.