Meeting Banner
Abstract #0094

Computed Diffusion-Weighted Imaging of the Prostate at 3T: Impact on Image Quality and Tumor Detection

Andrew B. Rosenkrantz1, Nicole Hindman1, Hersh Chandarana1, Fang-ming Deng2, James S. Babb1, Samir S. Taneja3, Christian Geppert4

1Radiology, NYU Langone Medical Center, New York, United States; 2Pathology, NYU Langone Medical Center, New York, United States; 3Urologic Oncology, NYU Langone Medical Center, New York, United States; 4Siemens Medical Systems, New York, United States

49 prostate cancer patients underwent 3T MRI including DWI with b-values 50 and 1000 mm2/sec. Computed DW-images with b-value 1500 were generated from the lower b-value images without additional scan time. Directly acquired b1500 images were also obtained in 39 patients. For two independent readers, using prostatectomy as reference, computed b-1,500 images had higher sensitivity and PPV for tumor than direct b-1,000 images and no significant difference than direct b-1,500 images. Scores for diagnostic confidence, tumor conspicuity, distortion, and benign prostate tissue suppression, were also significantly better for computed b-1,500 images than for direct b-1,000 images for both readers.

Keywords

achieved achieving acquisition additional allowed array arrow artifacts aside attribute audience avoiding aware background benign better blinded cancer challenged clinical coil comparable computed confidence conspicuous contrast critical decreased described details detection diagnosis diffusion direct directly distortion dominant either extrapolated fang findings generate ghosting grade greater highest impact improved improves included incomplete independently interpretation investigate involved leach lesion limitations localization location logistic mathematical mathematically measure medical ming newly note offering oncology overall pathologic pathology patients pelvic performance peripheral physicists placed potentially prostate providing quality radiologist radiologists radiology reader readers recent record recorded regression requires requiring scale scored sensitivity separate sets significantly simple solutions suffering suppression table target thereby third tissue transition trio tumor tumors undergoing utility visible visual zone