Advancing Prostate Cancer Diagnosis and Treatment Through AI-Driven Decision-making: A Comprehensive PRISMA-Based Systematic Review

Abstract

Decision-making during the diagnosis and treatment of prostate cancer often requires the evaluation of data from several sources. The purpose of this systematic was to evaluate how AI can enhance prostate cancer diagnosis and treatment. We search five different databases for available studies using a pre-specified search strategy. The study was conducted using PRISMA guidelines. A total of 1058 studies were found on five different databases (Scopus N=287, Web of Science N=204, PubMed/EMBASE N=306, Cochrane Library N=94, Google Scholar N=167). The articles were stored in the ENDNOTE library and all the duplicates were removed. We found 19 relevant studies that were included after a stringent assessment criterion based on the inclusion and exclusion criteria. Included studies were assessed for risk bias assessment using the Newcastle Ottawa Scale (NOS) in which one study was found to have high-risk bias assessment, seven were identified to have low-risk bias assessment while the rest of the studies were moderate. Clinicians can recognize complex correlations and handle massive data sets with the help of artificial intelligence. These tasks are extremely laborious and challenging for humans to complete. It is feasible to employ fewer resources while increasing overall effectiveness and precision in prostate cancer detection and treatment by utilizing AI algorithms and lowering the degree of subjectivity. This review provides a comprehensive comparison of AI algorithms in detection of prostate cancer.