A Machine Learning-Based Predictive Model for Prostate-Specific Antigen Screening: Insights from SHAP Values Analysis  
Author Yen Chun Huang

 
Co-Author(s) Yi-Ting Lin; Mingchih Chen; Him-Hua Ho

 
Abstract Localized prostate cancer is one of the most common cancers in men, leading to cancer-related mortality and a decline in quality of life [1, 2]. However, patients who undergo radical prostatectomy (RP) and those who receive radiation therapy (RT) show no significant differences in overall survival rates and long-term quality of life (QoL)[3]. Shared Decision-Making (SDM) has become integral to managing medical decisions for patients with localized prostate cancer. However, when choosing a treatment, patients often encounter dilemmas. The current SDM approaches provide statistical data on survival rates and complications [3, 4]. Currently, very few studies measure the impact of these psychological and physiological effects on decision-making[5]. Thus, healthcare professionals may be unable to offer the most accurate guidance tailored to the individual characteristics of each patient[6]. This study aims to develop a machine learning classifier that integrates past participants' experiences and decisionmaking tendencies to provide future participants with customized recommendations.

 
Keywords Prostate-specific antigen (PSA); Machine learning (ML); Random Forest (RF); prostate-specific antigen (PSA); SHapley Additive exPlanations (SHAP)
   
    Article #:  DSBFI25-82
 
Proceedings of 3rd ISSAT International Conference on Data Science in Business, Finance and Industry
January 6-8, 2025 - Da Nang, Vietnam