Related Experiment Video
Updated: May 30, 2025
06:08
A Cognitive Fusion-guided Prostate Biopsy Using Multiparametric Magnetic Resonance Imaging and Transrectal Ultrasound
Published on: March 21, 2025
128
An Artificial Intelligence Model Using Diffusion Basis Spectrum Imaging Metrics Accurately Predicts Clinically
Eric H Kim1,2, Huaping Jing3, Kainen L Utt3
1Division of Urology, Department of Surgery, University of Nevada Reno School of Medicine, Reno, Nevada.
The Journal of Urology
|January 27, 2025
Summary
Diffusion basis spectrum imaging (DBSI) with AI accurately predicts clinically significant prostate cancer (csPCa). Combining DBSI and PI-RADS may reduce unnecessary biopsies for prostate cancer.
Area of Science:
- Radiology
- Artificial Intelligence
- Oncology
Background:
- Conventional prostate MRI has limitations in accurately detecting clinically significant prostate cancer (csPCa).
- Diffusion basis spectrum imaging (DBSI) offers advanced metrics for prostate cancer assessment.
- Biopsy remains the gold standard but carries risks and costs.
Purpose of the Study:
- To evaluate the efficacy of an artificial intelligence (AI) model utilizing DBSI metrics for predicting csPCa.
- To compare the diagnostic performance of the DBSI-AI model against established biomarkers like PSA density (PSAD) and PI-RADS.
- To assess the potential of the DBSI-AI model to reduce unnecessary prostate biopsies.
Main Methods:
- 241 patients underwent MRI with conventional and DBSI sequences prior to biopsy.
- AI models were trained using DBSI metrics, with biopsy pathology as the ground truth.
- The DBSI-AI model's performance was compared with PSAD and PI-RADS for csPCa risk discrimination (Gleason score >= 7).
Main Results:
- The DBSI-AI model independently predicted csPCa (OR 2.04, P < .01).
- DBSI-AI model alone showed performance similar to PSAD + PI-RADS (AUC 0.863 vs 0.859).
- The combination of DBSI-AI model + PI-RADS achieved the highest risk discrimination (AUC 0.894, P < .01).
- A strategy using DBSI-AI for PI-RADS 1-3 could reduce biopsies by 27% while missing 2% of csPCa.
Conclusions:
- An AI model based on DBSI accurately predicts csPCa.
- Combining the DBSI-AI model with PI-RADS can enhance risk stratification.
- This approach holds promise for reducing unnecessary prostate biopsies.
