Using radiomics model for predicting extraprostatic extension with PSMA PET/CT studies: a comparative study with the...

Area of Science:

• Oncology
• Radiology
• Medical Imaging

Background:

• Extraprostatic extension (EPE) is a critical factor in prostate cancer staging and treatment planning.
• Accurate preoperative assessment of EPE is essential for guiding surgical and therapeutic decisions.

Purpose of the Study:

• To evaluate a radiomics model for predicting EPE in prostate cancer using PSMA PET/CT imaging.
• To compare the predictive performance of the radiomics model against the Mehralivand Grading System, an MRI-based assessment tool.

Main Methods:

• Radiomics features were extracted from PSMA PET/CT images of 206 patients undergoing radical prostatectomy.
• Predictive models were developed using Support Vector Machine and Random Forest algorithms.
• Performance was compared to the Mehralivand Grading System in 63 patients with both PSMA PET/CT and mpMRI.

Main Results:

• The PSMA PET/CT radiomics model achieved an AUC of 76.8%, outperforming the Mehralivand Grading System (AUCs ranging from 60.2% to 66.8%).
• The radiomics model demonstrated superior sensitivity (72.0%) and specificity (81.5%) compared to the grading system.
• Statistical analysis confirmed the radiomics model's significant outperformance over all readers of the Mehralivand Grading System (p < 0.013).

Conclusions:

• The radiomics model derived from PSMA PET/CT shows significant promise for improving preoperative EPE assessment in prostate cancer.
• This AI-driven approach may enhance the accuracy of predicting EPE compared to current MRI-based methods.
• Further validation in larger, independent cohorts is recommended to establish clinical utility.