Diffusion-weighted signal models in healthy and cancerous peripheral prostate tissues: comparison of outcomes obtained at different b-values

Abstract

Purpose: To evaluate the dependence on the b-values adopted of apparent diffusion coefficient (ADC), perfusion fraction (PF), slow and fast diffusion coefficient (Dslow, Dfast), corrected diffusion coefficient (D) and kurtosis (K), in healthy peripheral (HP) and peripheral cancerous (PCa) prostate tissues.

Materials and methods: Patients who underwent multiparametric prostate MR examination were retrospectively evaluated for possible inclusion. ADC, PF, Dslow, Dfast, D, and K were estimated both in HP and PCa tissues, using three different ranges of b-values: 0-2300, 0-1800, 0-800 s/mm2 (group A, B and C, respectively). Analysis of variance (ANOVA) and receiver operating characteristic (ROC) analysis were performed, to establish differences among groups and to evaluate sensitivity and specificity of every parameter in distinguishing HP and PCa tissues when calculated with different b-values.

Results: In all, 57 patients were included. ANOVA showed significant differences of all parameters between group A-B vs. C, both in HP and PCa tissues. In ROC analysis K showed the best area under the curve (AUC) when calculated in groups A and B (0.87 and 0.86), while it was comparable with the ADC one in group C (both 0.82).

Conclusion: A significant dependence on the adopted b-values of DWI parameters is shown. The best performance in distinguishing HP from PCa tissues was obtained by K, calculated using a high b-value sequence.

Keywords: ADC; biexponential DWI model; kurtosis; prostate DWI.