Diffusion tensor magnetic resonance imaging of prostate cancer

Purpose: To explore the feasibility of 3T magnetic resonance (MR) diffusion tensor imaging (DTI) and fiber tracking (FT) in patients with prostate cancer. Materials and methods: Thirty consecutive patients (mean age, 62.5 years) with biopsy proven prostate cancer underwent 3T-MR imaging (MRI) and DTI using a 6-channel external phased-array coil before radical prostatectomy. Regions of interest of 14 pixels were defined in tumors and nonaffected areas in the peripheral zone (PZ) and central gland (CG), according to histopatology after radical prostatectomy. Apparent diffusion coefficient (ADC) and fractional anisotropy (FA) values were determined. Differences in mean ADC and FA values among prostate cancer, normal PZ and CG were compared by 2-sided Student t test. The predominant diffusion direction of the prostate anisotropy was color coded on a directionally encoded color (DEC) map. A 3D reconstruction of fiber tract orientations of the whole prostate was determined using the continuous tracking method. The overall image quality for tumor localization and local staging was assessed in retrospective matching with whole-mount section histopathology images. Nodules detected at MRI were classified as matched lesions if tumor presence and extension were evidenced at histopathology. Results: For all the patients, the DTI sequence images were suitable for the evaluation of the zonal anatomy of the prostate gland and the tumor localization. Quantitative evaluation of the regions of interest (ROIs) showed a mean ADC value significantly lower in the peripheral neoplastic area (1.06 +/- 0.37 X 10(-3) mm(2)/s) than in the normal peripheral portion (1.95 +/- 0.38 X 10(-3) mm(2)/s) (P < 0.05). The mean FA values calculated in the normal peripheral (0.47 +/- 0.04) and central area (0.41 +/- 0.08) were very similar (P > 0.05). The mean FA values in the neoplastic lesion (0.27 +/- 0.05) were significantly lower (P < 0.05) than in the normal peripheral area and in the normal central and adenomyomatous area. DEC map showed a top-bottom type preferential direction in the peripheral but not in the central area, with the tumor lesions reducing the diffusion coding direction represented as color zones tending toward gray. Tractographic analysis permitted good delineation of the prostate anatomy (capsule outline, peripheral and central area borders) and neoplastic lesion extension and capsule infiltration compared with histopathology. Conclusions: Three Tesla DTI of the prostate gland is feasible and has the potential for providing improved diagnostic information.