Diffusion-Weighted Magnetic Resonance Imaging of the Prostate: Improved Robustness With Stretched Exponential Modeling
Y. Mazaheri,A. Afaq,D. Rowe,Yonggang Lu,A. Shukla-Dave,J. Grover
DOI: https://doi.org/10.1097/RCT.0b013e31826bdbbd
2012-11-01
Journal of Computer Assisted Tomography
Abstract:Purpose This study aimed to compare the intraclass correlation coefficients of parameters estimated with stretched exponential and biexponential diffusion models of in vivo diffusion-weighted magnetic resonance imaging (MRI) of the prostate. Methods After the institutional review board issued a waiver of informed consent for this Health Insurance Portability and Accountability Act–compliant study, 25 patients with biopsy-proven prostate cancer underwent 3T endorectal MRI and diffusion-weighted MRI of the prostate at 10 b values (0, 45, 75, 105, 150, 225, 300, 600, 900, and 1200 s/mm2). The full set of b values was collected twice within a single acquisition. Intraclass correlation coefficients were calculated for intra-acquisition variability. From the biexponential model, the quantitative parameters diffusion coefficient (D), perfusion coefficient (D*), and perfusion fraction (f) were estimated. From the stretched exponential model, the quantitative parameters Kohlrausch decay constant (DK) and alpha (&agr;) were estimated. Results For the 25 patient data sets, the average intraclass correlation coefficients for DK and &agr; were 95.8%, and 64.1%, respectively, whereas those for D, D*, and f were 84.4%, 25.3%, and 41.3%, respectively. Conclusions The stretched exponential diffusion model captures the nonlinear effects of intravoxel incoherent motion in the prostate. The parameters derived from this model are more reliable and reproducible than the parameters derived from the standard, widely used biexponential diffusion/perfusion model.