Computed Tomography-Based Radiomics for Long-Term Prognostication of High-Risk Localized Prostate Cancer Patients...

Area of Science:

• Oncology
• Radiology
• Medical Imaging Analysis

Background:

• High-risk prostate cancer (PCa) has a significant mortality rate (>40%).
• Traditional prognostic markers for PCa lack reliability.
• Whole pelvic radiotherapy (WPRT) is a treatment for localized PCa.

Purpose of the Study:

• To investigate planning computed tomography (pCT)-based radiomics for long-term prognostication.
• To predict progression-free survival in high-risk localized PCa patients receiving WPRT.

Main Methods:

• Retrospective study of 64 high-risk PCa patients (45 training, 19 testing).
• Radiomics model development using pCT images, PyRadiomics feature extraction from CTVprostate, and feature selection.
• Model performance evaluated using AUC, accuracy, sensitivity, and specificity with three-fold cross-validation.

Main Results:

• pCT-based radiomics model predicted six-year progression-free survival with mean AUCs of 0.76 (training) and 0.71 (testing).
• Radiomics signature (two texture features) achieved high accuracy (0.842), sensitivity (0.867), and specificity (0.750) in the testing cohort.
• Performance comparable to MRI-based radiomics studies.

Conclusions:

• pCT-based radiomics is a reliable, non-invasive method for predicting WPRT outcomes in high-risk localized PCa.
• This approach can potentially be integrated into routine clinical practice for prognostic prediction.
• CT is more accessible than MRI, making pCT-based radiomics a practical tool.