Preliminary validation of a new methodology for estimating dose reduction protocols in neonatal chest computed radiographs

Steven Don, Bruce R. Whiting, Charles F. Hildebolt, W. James Sehnert, Jacquelyn S. Ellinwood, Karin Töpfer, Parinaz Masoumzadeh, Richard A. Kraus, Keith A. Kronemer, Thomas Herman, William H. McAlister

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

3 Scopus citations


The risk of radiation exposure is greatest for pediatric patients and, thus, there is a great incentive to reduce the radiation dose used in diagnostic procedures for children to "as low as reasonably achievable" (ALARA). Testing of low-dose protocols presents a dilemma, as it is unethical to repeatedly expose patients to ionizing radiation in order to determine optimum protocols. To overcome this problem, we have developed a computed-radiography (CR) dose-reduction simulation tool that takes existing images and adds synthetic noise to create realistic images that correspond to images generated with lower doses. The objective of our study was to determine the extent to which simulated, low-dose images corresponded with original (non-simulated) low-dose images. To make this determination, we created pneumothoraces of known volumes in five neonate cadavers and obtained images of the neonates at 10 mR, 1 mR and 0.1 mR (as measured at the cassette plate). The 10-mR exposures were considered "relatively-noise-free" images. We used these 10 mR-images and our simulation tool to create simulated 0.1- and 1-mR images. For the simulated and original images, we identified regions of interest (ROI) of the entire chest, free-in-air region, and liver. We compared the means and standard deviations of the ROI grey-scale values of the simulated and original images with paired t tests. We also had observers rate simulated and original images for image quality and for the presence or absence of pneumothoraces. There was no statistically significant difference in grey-scale-value means nor standard deviations between simulated and original entire chest ROI regions. The observer performance suggests that an exposure ≥0.2 mR is required to detect the presence or absence of pneumothoraces. These preliminary results indicate that the use of the simulation tool is promising for achieving ALARA exposures in children.

Original languageEnglish
Title of host publicationMedical Imaging 2006
Subtitle of host publicationPhysics of Medical Imaging
StatePublished - 2006
EventMedical Imaging 2006: Physics of Medical Imaging - San Diego, CA, United States
Duration: Feb 12 2006Feb 16 2006

Publication series

NameProgress in Biomedical Optics and Imaging - Proceedings of SPIE
Volume6142 II
ISSN (Print)1605-7422


ConferenceMedical Imaging 2006: Physics of Medical Imaging
Country/TerritoryUnited States
CitySan Diego, CA


  • Chest radiography
  • Computed radiography
  • Neonate
  • Noise simulation
  • X-ray dose reduction


Dive into the research topics of 'Preliminary validation of a new methodology for estimating dose reduction protocols in neonatal chest computed radiographs'. Together they form a unique fingerprint.

Cite this