RSNA 2008 

Abstract Archives of the RSNA, 2008


LL-GU2236-R01

Localization and Staging of Prostate Cancer on MR Imaging: Does Diffusion-weighted MR Imaging Increase Interreader Agreement?

Scientific Posters

Presented on December 4, 2008
Presented as part of LL-GU-R: Genitourinary

Participants

Courtney A. Woodfield MD, Presenter: Nothing to Disclose
David J. Grand MD, Abstract Co-Author: Nothing to Disclose
John Albert Pezzullo MD, Abstract Co-Author: Nothing to Disclose
Jason Machan, Abstract Co-Author: Nothing to Disclose
Glenn A. Tung MD, Abstract Co-Author: Nothing to Disclose

PURPOSE

To determine if diffusion weighted imaging (DWI) improves interreader agreement for both prostate cancer identification and staging of extracapsular extension (ECE) and seminal vesicle invasion (SVI).

METHOD AND MATERIALS

We performed an IRB approved retrospective review of endorectal prostate MRI performed with T2-weighted imaging and DWI on a 1.5 T magnet on men with biopsy proven prostate cancer. Single-shot echo-planar imaging DWI technique with b-values of 0 and 1000 s/mm2 was used. T2-weighted images alone (T2) or in conjunction with DWI (T2+DWI) were independently and randomly interpreted by 3 radiologists blinded to clinical data and biopsy results. On a 3 point scale (increased, no change, decreased), the readers also rated any change in confidence for tumor localization and staging with the addition of DWI. Interobserver agreement was estimated using kappa.

RESULTS

Prostate MRI was performed on 30 men of mean age 65.3±5.4 years with average PSA 10.5±8.0 ng/ml; 80 sextant core biopsies were positive for prostate cancer with mean Gleason Score 7±1. For tumor identification, mean kappa for T2 alone was 0.34 (range, 0.27-0.39) and for T2+DWI, k=0.59 (0.44-0.81); for staging ECE, mean kappa for T2 alone was 0.49 (0.35-0.63) and for T2+DWI, k=0.63 (0.56-0.71); for staging SVI, mean kappa for T2 alone was 0.57 (0.48-0.71) and for T2+DWI, k=0.70 (0.52-0.79). Interreader agreement increased for both tumor identification and staging of ECE and SVI in 8 of 9 comparisons, but increases in kappa were not statistically significant (p>.05). DWI increased reader confidence for tumor localization in a mean of 20 (66.7%) cases [range 15-23 (50-76.7%)], for ECE in a mean of 7 (23.3%) [1-17 (3.3-56.7%)] cases, and for SVI in a mean of 2 (6.7%) [4-20 (13.3-66.7%)] cases.

CONCLUSION

Despite the kappa increases for prostate cancer detection and staging, no statistically significant increase in reader agreement with the addition of DWI to T2 was found for this study population. Reader confidence for both tumor localization and staging did increase with the addition of DWI.

CLINICAL RELEVANCE/APPLICATION

Interreader inconsistency contributes to the variable detection and staging of prostate cancer with MRI. DWI may advance interreader agreement as well as the clinical application of prostate MRI.

Cite This Abstract

Woodfield, C, Grand, D, Pezzullo, J, Machan, J, Tung, G, Localization and Staging of Prostate Cancer on MR Imaging: Does Diffusion-weighted MR Imaging Increase Interreader Agreement?.  Radiological Society of North America 2008 Scientific Assembly and Annual Meeting, February 18 - February 20, 2008 ,Chicago IL. http://archive.rsna.org/2008/6020996.html