Integrative Machine Learning Model Leveraging DCE-MRI and PSA Values for Advanced Risk Stratification in Prostate...

Area of Science:

• Radiology and Oncology
• Artificial Intelligence in Medicine

Background:

• Accurate prostate cancer grading is essential for effective treatment planning and risk assessment.
• Identifying metastatic risk groups aids in stratifying patients for appropriate management.

Purpose of the Study:

• To develop a machine learning (ML) model for predicting ISUP grade metastatic risk groups in prostate cancer.
• To integrate Dynamic Contrast-Enhanced Magnetic Resonance Imaging (DCE-MRI) pharmacokinetic parameters with Prostate-Specific Antigen (PSA) values.

Main Methods:

• Retrospective analysis of 102 prostate cancer patients.
• Standardization of DCE-MRI pharmacokinetic parameters (Ktrans, Kep, Ve, CER, MaxSlope, IAUGC).
• Utilized Synthetic Minority Oversampling Technique for dataset balancing and evaluated ML models, including Random Forest, using AUC.

Main Results:

• Random Forest classifier achieved a high Area Under the Curve (AUC) of 0.92.
• Optimal threshold of 0.3 identified for balancing sensitivity and specificity in high-risk group detection.
• SHAP analysis revealed PSA, MaxSlope, and Kep as significant predictors.

Conclusions:

• Integrating DCE-MRI parameters and PSA values with ML enhances prediction of prostate cancer metastatic risk groups.
• This approach offers a reliable tool for metastasis screening and personalized prostate cancer treatment.