A Bland-Altman analysis of the bias between computed tomography and ultrasound prostate volume measurements

Med Dosim. 2008 Autumn;33(3):234-8. doi: 10.1016/j.meddos.2008.01.001. Epub 2008 Mar 31.

Abstract

This study assesses the agreement between computed tomography (CT) measurements of prostate volume and those obtained by ultrasound (US), a well-established non-invasive technique. Twenty-six patients aged between 58 and 74 years were evaluated for prostate seed implant therapy using both CT and US measurements. The level of agreement between these 2 methods, which were strongly correlated (R2 = 0.828; p < 0.0001), was determined through Bland-Altman analysis. The mean prostate volume (+/- one standard deviation) of the sample was 31.8 +/- 10.5 cc for the CT method and 27.0 +/- 8.2 cc for the US method. The prostate volumes obtained by CT were, on average, 17% larger than the corresponding volumes determined by US. The average bias between the 2 imaging methods is 4.80 cc or 15%, which is significantly larger than the clinically acceptable margin of 10%.

MeSH terms

  • Adenocarcinoma / diagnostic imaging
  • Aged
  • Artifacts*
  • Brachytherapy / methods
  • Contrast Media
  • Data Interpretation, Statistical
  • Humans
  • Imaging, Three-Dimensional / methods
  • Iodine Radioisotopes
  • Male
  • Middle Aged
  • Organ Size
  • Prostate / diagnostic imaging
  • Prostatic Neoplasms / diagnostic imaging*
  • Research Design
  • Subtraction Technique*
  • Tomography, X-Ray Computed* / methods
  • Ultrasonography* / methods

Substances

  • Contrast Media
  • Iodine Radioisotopes