3D delineation of prostate, rectum and bladder on MR images

Comput Med Imaging Graph. 2008 Oct;32(7):622-30. doi: 10.1016/j.compmedimag.2008.07.002. Epub 2008 Aug 9.

Abstract

In radiotherapy planning, target volumes and organs at risk delineation are a tedious and time-consuming task. In this paper we address a method to assist the radiologist in this task. We developed a 3D deformable model for prostate segmentation and used a seeded region growing algorithm for bladder and rectum delineation on MR images. Evaluation of the methods is made by comparison of the results to manual delineation in 24 patients. The following parameters were measured: volume ratio (V R) (automatic/manual), volume overlap (V O) (ratio of the volume of intersection to the volume of union, optimal value=1), and correctly delineated volume (V C) (percent ratio of the volume of intersection to the manual defined volume, optimal value=100). For prostate the V R, V O and V C were 1.13 (+/-0.1), 0.78 (+/-0.05) and 94.75 (+/-3.3), respectively. For rectum, the V R, V O and V C were 0.97 (+/-0.1), 0.78 (+/-0.06) and 86.52 (+/-5), respectively. V R, V O and V C were 0.95 (+/-0.03), 0.88 (+/-0.03) and 91.29 (+/-3.1) for bladder, respectively.

Publication types

  • Retracted Publication

MeSH terms

  • Algorithms
  • Artificial Intelligence
  • Cluster Analysis
  • Humans
  • Image Enhancement / methods
  • Image Interpretation, Computer-Assisted / methods*
  • Imaging, Three-Dimensional / methods*
  • Magnetic Resonance Imaging / methods*
  • Male
  • Pattern Recognition, Automated / methods
  • Prostate / pathology*
  • Prostatic Neoplasms / diagnosis*
  • Rectum / pathology*
  • Reproducibility of Results
  • Sensitivity and Specificity
  • Urinary Bladder / pathology*