A Systematic Review of Artificial Intelligence-Based Clinical Decision Support Systems in Prostate Cancer Management.

Area of Science:

• Oncology
• Medical Informatics
• Artificial Intelligence

Background:

• Prostate cancer (PCa) is a significant global health concern, impacting over 1.4 million individuals annually.
• Artificial intelligence-based clinical decision support systems (AI-CDSS) are increasingly vital for enhancing diagnostic and treatment strategies in PCa management.

Purpose of the Study:

• To systematically review the development, implementation, and limitations of AI-CDSS in prostate cancer diagnosis, staging, treatment, and management.
• To evaluate the effectiveness and challenges of AI applications in PCa care.

Main Methods:

• Systematic literature search adhering to PRISMA guidelines across PubMed, Scopus, Web of Science, and Embase.
• PICO framework used to formulate research questions on AI-CDSS development, models, standards, and clinical integration.
• Two independent reviewers screened 964 articles, with disagreements resolved by a third reviewer.

Main Results:

• Ten studies met inclusion criteria, evaluating AI-CDSS for PCa diagnosis, risk prediction, survival estimation, and radiotherapy planning.
• Common AI models included gradient boosting, neural networks, and deep reinforcement learning (DRL), with modern ML outperforming traditional methods in prediction tasks.
• AI-CDSS demonstrated potential, with one system reducing decision-making time by 60%, but limitations included data quality, generalizability, and interpretability issues.

Conclusions:

• AI-based CDSS offer potential benefits for PCa diagnosis, prognosis, and treatment planning, potentially improving efficiency.
• Wider adoption necessitates high-quality, multi-site data, external validation, interpretable models, and seamless EHR integration.
• Prospective clinical evaluations are crucial to confirm the patient-level impact of AI-CDSS in PCa management.