Related Experiment Video
Updated: Apr 30, 2026
06:08
A Cognitive Fusion-guided Prostate Biopsy Using Multiparametric Magnetic Resonance Imaging and Transrectal Ultrasound
Published on: March 21, 2025
2.1K
Evaluating Clinically Significant Prostate Cancer with Pathology-Registered Radiomics: A Multi-Reader Assessment
Omer Tarik Esengur1, Stephanie A Harmon1, David G Gelikman1
1Molecular Imaging Branch, National Cancer Institute, National Institutes of Health, Bethesda, MD (O.T.E., S.A.H., D.G.G., E.C.Y., H.Z., B.S., R.L., P.L.C., B.T.).
Academic Radiology
|April 28, 2026
Summary
A new radiomics signature aids in detecting clinically significant prostate cancer (csPCa). Combining it with prostate-specific antigen density (PSAD) improves detection accuracy, approaching PI-RADS scores.
Area of Science:
- Radiology
- Oncology
- Medical Imaging
Background:
- Prostate cancer detection relies on MRI and PI-RADS scoring.
- Developing accurate, pathology-derived imaging biomarkers is crucial for improved diagnostics.
Purpose of the Study:
- To create a pathology-derived radiomics signature for detecting clinically significant prostate cancer (csPCa).
- To assess the signature's performance using simplified, lesion diameter-based MRI segmentations.
Main Methods:
- Retrospective analysis of 175 participants with biparametric MRI and whole-mount histopathology.
- Development of a 10-feature radiomics signature using PyRadiomics and recursive feature elimination.
- Evaluation of models including the signature, prostate-specific antigen density (PSAD), and PI-RADS scores using logistic regression, random forest, and XGBoost.
Main Results:
- The radiomics signature alone achieved 0.66 AUC and 62% accuracy.
- Combining the signature with PSAD (PSAD+Signature) yielded 0.75 AUC and 68% accuracy.
- PI-RADS combined with PSAD (PI-RADS+PSAD) achieved the highest accuracy at 74% (0.77 AUC).
Conclusions:
- A histopathology-trained radiomics signature shows moderate standalone performance for csPCa detection.
- The PSAD+Signature model significantly improved diagnostic performance over the signature alone.
- The PSAD+Signature approach offers a simplified, localized method that complements PI-RADS assessment with low implementation complexity.
