Stijn W.T.P.J. Heijmink MD, Presenter: Nothing to Disclose

Jurgen J. Futterer MD, Abstract Co-Author: Nothing to Disclose

Tom W.J. Scheenen PhD, Abstract Co-Author: Nothing to Disclose

Christina A. Hulsbergen-van der Kaa, Abstract Co-Author: Nothing to Disclose

J. Alfred Witjes MD, PhD, Abstract Co-Author: Nothing to Disclose

Jelle O. Barentsz MD, Abstract Co-Author: Nothing to Disclose

To assess image quality, localization and staging accuracies between body array coil (BAC) and endorectal coil (ERC) MR imaging of prostate cancer (PC) at 3T.

In a prospective study, eighteen patients with biopsy-proven and clinically localized PC (mean age: 60 years, mean PSA level: 6.7 ng/ml, median biopsy Gleason score: 6) underwent MR imaging at 3T prior to radical prostatectomy. Both with a BAC and an ERC, T2-weighted images in three planes were obtained to assess PC localization and disease stage and axial T1-weighted series were used to exclude post-biopsy haemorrhage. The highest attainable resolution was used depending on coil setup. Prospectively, two radiologists (R1 and R2, with one and three years of experience, respectively) independently read all imaging. The readers had no knowledge of PSA levels or Gleason scores. All data was anonymized. BAC and ERC imaging sets were read separately and in random order. For both BAC and ERC imaging, image quality characteristics were scored on a five-point scale running from poor to excellent. Disease stage was scored on a five-point probability scale from definite T2 to definite T3 disease. For prostate cancer localization, the prostate was divided into 14 sections. Readers scored each section as cancer or healthy on a five-point scale. Whole-mount section histopathology was the standard of reference. The areas under the receiver operating characteristic curve (AUC) were calculated. P<0.05 was considered statistically significant.

The anatomical details significantly improved (all p<0.05) when using an ERC compared with BAC. Also, sensitivity for detection of stage T3 disease increased from 0% to 100% (3/3) for both readers, while maintaining high specificity (91%-100%). The AUC for R1 increased from 0.50 to 0.98 and from 0.80 to 1.00 for R2. The total number of detected tumors increased from 44% with BAC to 54% with ERC. Of the 18 index tumors, BAC imaging found 11 (R1) and 13 (R2) compared to 15 (R1) and 14 (R2) with ERC imaging.

Image quality and staging performance increased significantly with the use of an ERC. However, 3T BAC imaging may play a role in primary PC screening for detection of the index tumor.

Heijmink, S, Futterer, J, Scheenen, T, Hulsbergen-van der Kaa, C, Witjes, J, Barentsz, J, A Comparison of Image Quality, Localization, and Staging Accuracies of Prostate Cancer between Body Array and Endorectal Coil MR Imaging at 3T.  Radiological Society of North America 2005 Scientific Assembly and Annual Meeting, November 27 - December 2, 2005 ,Chicago IL. http://archive.rsna.org/2005/4412300.html