Andrew B. Rosenkrantz MD, Presenter: Nothing to Disclose

Naomi M. Campbell MBBCh, Abstract Co-Author: Nothing to Disclose

Byron Gaing MD, Abstract Co-Author: Nothing to Disclose

Sooah Kim MD, Abstract Co-Author: Nothing to Disclose

James S. Babb PhD, Abstract Co-Author: Nothing to Disclose

Fang-Ming Deng MD, PhD, Abstract Co-Author: Nothing to Disclose

Samir S. Taneja MD, Abstract Co-Author: Consultant, Eigen Consultant, GTx, Inc Consultant, Bayer AG Consultant, Healthtronics, Inc Speaker, Johnson & Johnson Investigator, STEBA Biotech NV Royalties, Reed Elsevier

Past studies using a variety of methodologies have obtained conflicting results regarding the added value of diffusion-weighed imaging (DWI) and dynamic contrast-enhanced (DCE) MRI for detection of transition zone (TZ) tumor. In this study, our aim is to systematically evaluate the impact of multi-parametric MRI, including DWI performed using different b-values, on the accuracy, sensitivity, and specificity for TZ tumor localization.

This retrospective HIPAA-compliant study received IRB approval with a waiver of written informed consent. 106 men with prostate cancer (mean age 62±7years) who underwent 3T MRI using a pelvic phased-array coil before radical prostatectomy were included. Three radiologists independently reviewed cases to record the likelihood of tumor in each of six TZ regions. These scores were initially assigned using T2WI alone, re-assigned following integration of DWI-b1000 images and the corresponding apparent diffusion coefficient (ADC) maps, re-assigned again following integration of DWI-b2000 images, and re-assigned a final time following integration of DCE images, including both raw dynamic and post-processed parametric perfusion maps. Generalized estimating equations based on binary logistic regression were used to compare sessions for TZ tumor detection, using prostatectomy findings as the reference standard.

9.7% (62/636) of TZ sextants contained tumor. All three readers had higher sensitivity for T2WI+b1000/ADC compared with T2WI alone (R1: 54.8% vs. 33.9%, R2: 53.2% vs. 22.6%, R3: 50.0% vs. 19.4%, p≤0.002); two readers had further increased sensitivity also incorporating b2000 (R1: 74.2%, R2: 62.9%; p≤0.011), and the remaining reader had further increased sensitivity also incorporating both b2000 and DCE (R3: 61.3%, p=0.013). DCE otherwise did not improve sensitivity (p≥0.054). Other measures of performance were similar across the four sessions (R1: specificity 97.4%-98.3%, accuracy 91.2%-95.9%; R2: specificity 95.8%-98.4%, accuracy 91.0%-92.6%; R3: specificity 90.9%-96.7%, accuracy 88.1%-89.2%).

DWI assists TZ tumor detection through significantly higher sensitivity, particularly when using a very high b-value; DCE lacks further additional benefit.

The optimal parameter combination for localizing TZ tumor using 3T MRI entails both T2WI and DWI, but not DCE; with this approach, DWI should include a very high b-value (>1,000 s/mm2).

Rosenkrantz, A, Campbell, N, Gaing, B, Kim, S, Babb, J, Deng, F, Taneja, S, Transition Zone Prostate Cancer: Revisiting the Role of Multiparametric MRI at 3T.  Radiological Society of North America 2014 Scientific Assembly and Annual Meeting, - ,Chicago IL. http://archive.rsna.org/2014/14001223.html