Prostate Region-Wise Imaging Biomarker Profiles for Risk Stratification and Biochemical Recurrence Prediction.

Area of Science:

• Radiology
• Oncology
• Medical Imaging

Background:

• Accurate prognostication for prostate cancer (PCa) is vital for treatment selection.
• Identifying patients at high risk for biochemical recurrence (BCR) is a clinical challenge.

Purpose of the Study:

• To identify imaging biomarker profiles from MRI (perfusion/diffusion + radiomic features) capable of discriminating PCa patient risk and BCR occurrence.
• To evaluate the predictive value of these imaging biomarkers, with and without clinical data, for long-term outcomes.

Main Methods:

• Retrospective evaluation of localized PCa patients treated with neoadjuvant androgen deprivation therapy and radiotherapy.
• Extraction of imaging features from MRI, analyzed per prostate region and for the whole gland.
• Univariate and multivariate analyses to assess biomarker performance.

Main Results:

• Prostate region-specific imaging biomarker profiles, primarily radiomic features, differentiated risk groups and predicted BCR.
• Increased heterogeneity-related radiomic features correlated with worse prognosis and BCR.
• Imaging biomarkers showed good predictive ability (AUC > 0.725), significantly improving with combined clinical data (AUC 0.841-0.877 for BCR prediction).

Conclusions:

• Region-aware imaging profiles from MRI effectively identify PCa patients with poorer prognosis and higher BCR risk.
• Combining imaging biomarkers with clinical variables enhances predictive accuracy, particularly for BCR.