Anthropomorphic versus geometric chest phantoms: a comparison of scatter properties.


Journal Article

Previously, we have used an anthropomorphic chest phantom to study scatter reduction in digital chest radiography. Image metrics, such as scatter fractions, contrast, noise, and resolution, are not easily measured due to the anatomical structure in the phantom. A geometric chest phantom, recently developed for quality control purposes, offers the possibility of being used to calculate image quality measurements. Here, we compare the scatter properties of the two phantoms to determine if the geometric phantom can be used in our studies of scatter compensation techniques. A calibrated photostimulable phosphor system was used to acquire images of the two phantoms. An array of beam stops was placed in front of each phantom to calculate scatter fractions. Each phantom had approximately 2 in. of polystyrene material added to the posterior to increase scatter fractions to those normally seen in patients. Exposure parameters were 300 mA for 0.009 sec with a source to image distance of 100 cm. Energies were varied from 60 to 130 kVp. Scatter fractions were determined for different areas of anatomy for each energy and each phantom. For all energies examined, the two phantoms compare well for scatter fractions in each of six regions. For example, at 95 kVp, the geometric phantom had average scatter fractions of 0.72 and 0.88 in the lung and mediastinum regions, respectively. These values were 0.74 and 0.90 for the anatomic phantom. For comparison, measurements of scatter fractions in patients at these values have been reported as 0.65 and 0.90 for the lung and mediastinum regions. The geometric phantom is an excellent tool which can be used in place of the anthropomorphic phantom for studies involving scatter compensation. In addition to having a gray level histogram typical of a human chest, this phantom has uniform regions where image quality measurements can be calculated.

Full Text

Cited Authors

  • Baydush, AH; Ghem, WC; Floyd, CE

Published Date

  • May 2000

Published In

Volume / Issue

  • 27 / 5

Start / End Page

  • 894 - 897

PubMed ID

  • 10841391

Pubmed Central ID

  • 10841391

International Standard Serial Number (ISSN)

  • 0094-2405

Digital Object Identifier (DOI)

  • 10.1118/1.598954


  • eng

Conference Location

  • United States