Related Experiment Video
Updated: Nov 23, 2025
09:30
Quantitative Magnetic Resonance Imaging of Skeletal Muscle Disease
Published on: December 18, 2016
19.8K
Semi-automated PIRADS scoring via mpMRI analysis
Nikhil J Dhinagar1, William Speier1, Karthik V Sarma1
1University of California, Los Angeles, David Geffen School of Medicine, Department of Radiological Sciences, Los Angeles, California, United States.
Journal of Medical Imaging (Bellingham, Wash.)
|January 4, 2021
Summary
An AI tool assists in Prostate Cancer diagnosis by harmonizing Prostate Imaging-Reporting and Data System (PIRADS) scores for mpMRI. This artificial intelligence model improves accuracy in identifying aggressive tumors for early treatment.
Area of Science:
- Radiology
- Artificial Intelligence
- Oncology
Background:
- Prostate cancer (PCa) is a leading cause of cancer death in men.
- Multiparametric magnetic resonance imaging (mpMRI) aids in detecting clinically significant PCa (csPCa).
- Prostate Imaging-Reporting and Data System (PIRADS) scores standardize mpMRI interpretation, but suffer from inter-reader variability.
Purpose of the Study:
- To develop and validate a minimal input, semi-automated artificial intelligence (AI) system to harmonize PIRADS scoring.
- To reduce inter-reader variability in PIRADS score assignment for prostate cancer detection.
Main Methods:
- A deep learning model, termed the seed point model, was developed.
- The model uses a single-click seed point for lesion annotation on mpMRI, unlike traditional methods requiring full lesion annotation.
- The model was trained and validated on mpMRI data from 617 patients to classify lesions with PIRADS score ≥ 4.
Main Results:
- The seed point model achieved an average receiver-operator characteristic (ROC) area under the curve (ROC-AUC) of 0.704.
- This performance was significantly higher than previously published benchmarks.
- The model demonstrated robust performance in classifying PIRADS scores on mpMRI.
Conclusions:
- The proposed AI model can assist in PIRADS scoring of mpMRI, acting as a quality-promoting second read.
- It can provide expertise in settings lacking specialized radiologists for prostate mpMRI interpretation.
- The model aids in identifying higher PIRADS tumors for improved early clinical management and treatment of PCa.
