Diagnosis of Prostate Cancer with Noninvasive Estimation of Prostate Tissue Composition by Using Hybrid Multidimensional MR Imaging: A Feasibility Study

Purpose: To evaluate whether compartmental analysis by using hybrid multidimensional magnetic resonance (MR) imaging can be used to diagnose prostate cancer and determine its aggressiveness. Materials and Methods: Twenty-two patients with prostate cancer underwent preoperative 3.0-T MR imaging. Axial images were obtained with hybrid multidimensional MR imaging by using all combinations of echo times (47, 75, 100 msec) and b values of 0, 750, 1500 sec/mm(2), resulting in a 3 x 3 array of data associated with each voxel. Volumes of the tissue components stroma, epithelium, and lumen were calculated by fitting the hybrid data to a three-compartment signal model, with distinct, paired apparent diffusion coefficient (ADC) and T2 values associated with each compartment. Volume fractions and conventional ADC and T2 were measured for regions of interest in sites of prostatectomy-verified malignancy (n = 28) and normal tissue (n = 71). Receiver operating characteristic (ROC) analysis was used to evaluate the performance of various parameters in differentiating prostate cancer from benign tissue. Results: Compared with normal tissue, prostate cancer showed significantly increased fractional volumes of epithelium (23.2% +/- 7.1 vs 48.8% +/- 9.2, respectively) and reduced fractional volumes of lumen (26.4% +/- 14.1 vs 14.0% +/- 5.2) and stroma (50.5% +/- 15.7 vs 37.2% +/- 9.1) by using hybrid multidimensional MR imaging. The fractional volumes of tissue components show a significantly higher Spearman correlation coefficient with Gleason score (epithelium: rho = 0.652, P = .0001; stroma: rho = -0.439, P = .020; lumen: r = -0.390, P = .040) compared with traditional T2 values (rho = -0.292, P = .132) and ADCs (rho = -0.315, P = .102). The area under the ROC curve for differentiation of cancer from normal prostate was highest for fractional volume of epithelium (0.991), followed by fractional volumes of lumen (0.800) and stroma (0.789). Conclusion: Fractional volumes of prostatic lumen, stroma, and epithelium change significantly when cancer is present. These parameters can be measured noninvasively by using hybrid multidimensional MR imaging and have the potential to improve the diagnosis of prostate cancer and determine its aggressiveness. (C) RSNA, 2018.