Predicting Prostate Cancer Risk and Its Associated Factors Using Machine Learning Techniques: A Retrospective Study

Area of Science:

• Medical Informatics
• Oncology
• Machine Learning

Background:

• Prostate cancer (PCa) is a significant global health concern.
• Artificial intelligence (AI) and machine learning (ML) are increasingly utilized for disease prediction and diagnosis.
• Identifying PCa risk factors is crucial for early detection and management.

Purpose of the Study:

• To apply ML techniques for predicting PCa.
• To identify the most significant risk factors associated with PCa development.
• To evaluate the performance of various ML algorithms in PCa prediction.

Main Methods:

• A retrospective study involving 597 patient records from Shahid Beheshti Hospital, Iran.
• Literature search across major databases (Web of Science, Scopus, PubMed) from 2000-2024 for risk factors.
• Development and evaluation of ML models including Logistic Regression, Gradient Boosting, Random Forest, XGBoost, Support Vector Machine, and Neural Networks.
• Performance assessment using accuracy, precision, recall, and F1-score.

Main Results:

• The XGBoost model demonstrated the highest predictive performance with 77.5% accuracy, 74.5% sensitivity, and 79.7% specificity.
• Significant predictors for PCa included total and free prostate-specific antigen (PSA) levels, hemoglobin, Body Mass Index (BMI), and fish consumption.
• The study identified 49 potential risk factors for PCa.

Conclusions:

• ML models can enhance the personalized identification of PCa risks.
• Further research is needed to refine ML algorithms and mitigate data biases for improved PCa prediction.
• The findings highlight the potential of ML in understanding and managing PCa.