Biexponential Apparent Diffusion Coefficients Values in the Prostate: Comparison among Normal Tissue, Prostate Cancer, Benign Prostatic Hyperplasia and Prostatitis

Objective: To investigate the biexponential apparent diffusion parameters of diverse prostate tissues and compare them with monoexponential apparent diffusion coefficient (ADC) value in the efficacy to discriminate prostate cancer from benign lesions. Materials and Methods: Eleven healthy volunteers and 61 patients underwent a conventional (b-factors 0, 1000 $s/mm^2$) and a 10 b-factor (0 to 3000 $s/mm^2$) diffusion-weighted imaging (DWI). The monoexponential ADC value and biexponential parameters of fast ADC (ADCf), fraction of ADCf (f), slow ADC (ADCs) value for 29 prostate cancer, 28 benign prostatic hyperplasia (BPH), 24 prostatitis lesions and normal tissue were calculated and compared. Receiver operating characteristic analysis was performed to determine the sensitivity, specificity and optimal cut-off points. Results: Prostate cancer had lower ADC, ADCf, f, and ADCs than all other tissues (p < 0.01). Prostatitis exhibited a lower ADC, ADCf, ADCs and f than the peripheral zone tissue (p < 0.01), and BPH showed a lower ADC and ADCf than the central gland tissue (p < 0.01). The ADCf demonstrated a comparable accuracy with ADC in differentiating cancer from BPH [area under the curve (AUC) 0.93 vs. 0.92] and prostatitis AUC 0.98 vs. 0.99) (both p > 0.05), but the AUC of f and ADCs in differentiating cancer from BPH (0.73 and 0.81) and prostatitis (0.88 and 0.91) were significantly lower than ADC (all p < 0.05). Conclusion: The biexponential DWI appears to provide additional parameters for tissue characterization in prostate, and ADCf helps to yield comparable accuracy with ADC in differentiating cancer from benign lesions.