Staging of prostate cancer: Results of Radiology Diagnostic Oncology Group project comparison of three MR imaging techniques

Clare M. Tempany, Xiao Zhou, Elias A. Zerhouni, Matthew D. Rifkin, Leslie E. Quint, Catherine W. Piccoli, James H. Ellis, Barbara J. McNeil

Research output: Contribution to journalArticlepeer-review

Abstract

PURPOSE: To assess accuracy of three different magnetic resonance (MR) imaging techniques, including the endorectal coil, in staging prostate cancer. MATERIALS AND METHODS: MR imaging was performed in 213 patients with prostate cancer with a conventional body coil, with fat suppression and a body coil, and with an endorectal coil. Radiologists identified tumor invasion into periprostatic tissues, neurovascular bundles, and seminal vesicles. Each technique was evaluated separately, and in a subset of 74 patients the three techniques were evaluated together. Images obtained with the two body-coil techniques were read in combination with images obtained with the endorectal coil (combination A) and alone (combination B). RESULTS: Overall accuracy for conventional body-coil, fat-suppressed body-coil, and endorectal-coil MR was 61%, 64%, and 54%, respectively. Overall group accuracy for combinations A and B was 57% and 61%. Considerable interreader variability was found for combination A. CONCLUSION: No technique was highly accurate for staging early prostate cancer. Individual radiologists did achieve a high degree of staging accuracy with the endorectal-coil and body- coil combination.

Original languageEnglish (US)
Pages (from-to)47-54
Number of pages8
JournalRADIOLOGY
Volume192
Issue number1
DOIs
StatePublished - Jul 1994

Keywords

  • Diagnostic radiology, observer performance
  • Prostate, MR
  • Prostate, neoplasms

ASJC Scopus subject areas

  • Radiology Nuclear Medicine and imaging

Fingerprint

Dive into the research topics of 'Staging of prostate cancer: Results of Radiology Diagnostic Oncology Group project comparison of three MR imaging techniques'. Together they form a unique fingerprint.

Cite this