Skip to main content

SU-E-I-94: Automated Image Quality Assessment of Radiographic Systems Using An Anthropomorphic Phantom.

Publication ,  Conference
Wells, J; Wilson, J; Zhang, Y; Samei, E; Ravin, CE
Published in: Med Phys
June 2014

PURPOSE: In a large, academic medical center, consistent radiographic imaging performance is difficult to routinely monitor and maintain, especially for a fleet consisting of multiple vendors, models, software versions, and numerous imaging protocols. Thus, an automated image quality control methodology has been implemented using routine image quality assessment with a physical, stylized anthropomorphic chest phantom. METHODS: The "Duke" Phantom (Digital Phantom 07-646, Supertech, Elkhart, IN) was imaged twice on each of 13 radiographic units from a variety of vendors at 13 primary care clinics. The first acquisition used the clinical PA chest protocol to acquire the post-processed "FOR PRESENTATION" image. The second image was acquired without an antiscatter grid followed by collection of the "FOR PROCESSING" image. Manual CNR measurements were made from the largest and thickest contrast-detail inserts in the lung, heart, and abdominal regions of the phantom in each image. An automated image registration algorithm was used to estimate the CNR of the same insert using similar ROIs. Automated measurements were then compared to the manual measurements. RESULTS: Automatic and manual CNR measurements obtained from "FOR PRESENTATION" images had average percent differences of 0.42%±5.18%, -3.44%±4.85%, and 1.04%±3.15% in the lung, heart, and abdominal regions, respectively; measurements obtained from "FOR PROCESSING" images had average percent differences of -0.63%±6.66%, -0.97%±3.92%, and -0.53%±4.18%, respectively. The maximum absolute difference in CNR was 15.78%, 10.89%, and 8.73% in the respective regions. In addition to CNR assessment of the largest and thickest contrast-detail inserts, the automated method also provided CNR estimates for all 75 contrast-detail inserts in each phantom image. CONCLUSION: Automated analysis of a radiographic phantom has been shown to be a fast, robust, and objective means for assessing radiographic image quality. The method reduces the burden of manual measurements and provides a means by which to monitor and compare radiographic system performance.

Duke Scholars

Published In

Med Phys

DOI

ISSN

0094-2405

Publication Date

June 2014

Volume

41

Issue

6

Start / End Page

152

Location

United States

Related Subject Headings

  • Nuclear Medicine & Medical Imaging
  • 5105 Medical and biological physics
  • 4003 Biomedical engineering
  • 1112 Oncology and Carcinogenesis
  • 0903 Biomedical Engineering
  • 0299 Other Physical Sciences
 

Citation

APA
Chicago
ICMJE
MLA
NLM
Wells, J., Wilson, J., Zhang, Y., Samei, E., & Ravin, C. E. (2014). SU-E-I-94: Automated Image Quality Assessment of Radiographic Systems Using An Anthropomorphic Phantom. In Med Phys (Vol. 41, p. 152). United States. https://doi.org/10.1118/1.4888044
Wells, J., J. Wilson, Y. Zhang, E. Samei, and C. E. Ravin. “SU-E-I-94: Automated Image Quality Assessment of Radiographic Systems Using An Anthropomorphic Phantom.” In Med Phys, 41:152, 2014. https://doi.org/10.1118/1.4888044.
Wells J, Wilson J, Zhang Y, Samei E, Ravin CE. SU-E-I-94: Automated Image Quality Assessment of Radiographic Systems Using An Anthropomorphic Phantom. In: Med Phys. 2014. p. 152.
Wells, J., et al. “SU-E-I-94: Automated Image Quality Assessment of Radiographic Systems Using An Anthropomorphic Phantom.Med Phys, vol. 41, no. 6, 2014, p. 152. Pubmed, doi:10.1118/1.4888044.
Wells J, Wilson J, Zhang Y, Samei E, Ravin CE. SU-E-I-94: Automated Image Quality Assessment of Radiographic Systems Using An Anthropomorphic Phantom. Med Phys. 2014. p. 152.

Published In

Med Phys

DOI

ISSN

0094-2405

Publication Date

June 2014

Volume

41

Issue

6

Start / End Page

152

Location

United States

Related Subject Headings

  • Nuclear Medicine & Medical Imaging
  • 5105 Medical and biological physics
  • 4003 Biomedical engineering
  • 1112 Oncology and Carcinogenesis
  • 0903 Biomedical Engineering
  • 0299 Other Physical Sciences