Ultra-high b-value diffusion-weighted MRI for the detection of prostate cancer with 3-T MRI

Purpose: To determine, with histopathological findings of radical prostatectomy as reference, whether diffusionweighted imaging (DWI) using b = 2000 s/mm(2) for 3-T magnetic resonance imaging (MRI) is superior to the use of b = 1000 s/mm(2) for prostate cancer detection. Materials and Methods: This study evaluated 73 patients with biopsy-proven prostate cancer. All patients underwent preoperative 3-T MRI using T2-weighted imaging (T2WI) and DWI (b = 0, 1000, 2000 s/mm(2)). The following three sets of images were evaluated separately by two radiologists: protocol A (T2WI alone), protocol B (T2WI and DWI with b = 1000 s/mm(2)), and protocol C (T2WI and DWI with b = 2000 s/mm(2)). For estimation of diagnostic capability, area under the receiver operating characteristic (ROC) curve (AUC) was calculated. Results: A total of 341 cancer foci were found in the prostectomy specimens of 73 patients. Reader 1 rated AUCs of the three sets as; A 0.66, B 0.77, C 0.80. ROC analysis showed significant differences among the three protocols (A vs. B vs. C: P < 0.0001, A vs. B: P < 0.0001, B vs. C: P < 0.0001). Conclusion: The use of b = 2000 s/mm(2) for DWI with 3-T MRI is diagnostically superior to that of b = 1000 s/mm(2) for prostate cancer detection.