Novel Diffusion-Weighted MRI for High-Grade Prostate Cancer Detection
Abstract
Diffusion-weighted MRI with the calculated apparent diffusion coefficient (ADC) has been associated with prostate cancer (PCa) grade however there is opportunity to improve the accuracy. We therefore implemented a quantitative fractional order calculus (FROC) model for the first time for PCa detection. A retrospective cohort of 37 men undergoing MRI and confirmatory biopsy (n=31) and radical prostatectomy (n=6) were studied. The FROC model was compared to ADC and the stretched exponential model (SEM) of diffusion. FROC was found to have improved accuracy (AUC = 0.82) compared to ADC and to the clinical standard of care (PI-RADS, AUC = 0.74). Therefore, a prospective validation study of 50 men with no prior biopsy and suspected PCa was performed. A unique image acquisition protocol was designed and implemented for the study. The FROC model performed similarly as in the retrospective cohort and to ADC (both AUC = 0.80). FROC significantly outperformed PI-RADS (AUC = 0.63). We conclude that quantitative diffusion parameters improve PCa detection.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jan 01, 2021
- Accession Number
- AD1152630
Entities
People
- Andre Kajdacsy-balla
- Hong-Cai Zhou
- Meltem Uyanik
- Michael Abern
- Peter H Gann
- Richard Magin
- Virgilia Macias
Organizations
- University of Illinois at Chicago