RARPKB: a knowledge-guide decision support platform for personalized robot-assisted surgery in prostate cancer

Area of Science:

• Urology
• Surgical Oncology
• Medical Informatics

Background:

• Robot-assisted radical prostatectomy (RARP) is crucial for prostate cancer (PCa) treatment.
• Clinical complexity and PCa heterogeneity challenge RARP decision-making.
• Physician expertise limitations impact rational RARP planning.

Purpose of the Study:

• To organize knowledge from complex RARP cases.
• To establish an online platform, the RARP knowledge base (RARPKB).
• To provide evidence for personalized RARP treatment plans.

Main Methods:

• Conducted PubMed searches over two decades for RARP publications.
• Collected, classified, and structured surgical details, patient data, and outcomes.
• Developed a knowledge-guided decision-support tool using MySQL, DataTable, ECharts, and JavaScript.
• Validated RARPKB using ChatGPT-4 and assessment scales.

Main Results:

• The RARPKB platform integrates data from 583 studies, 1589 cohorts, and over 1.9 million patients.
• The decision-support tool provides personalized surgical plan and complication recommendations.
• RARPKB demonstrated superior performance over ChatGPT-4 in authenticity, matching, and personalized recommendations.
• High usability scores (System Usability Scale: 88.88) and Net Promoter Score (85) indicate user satisfaction.

Conclusions:

• RARPKB is the first knowledge base for robot-assisted surgery, focusing on PCa.
• It aids in personalized and complex surgical planning for PCa, potentially improving efficacy.
• RARPKB serves as a reference for future AI applications in clinical practice.