Download PDF (external access)

Radiography

Publication date: 2015-02-01
Volume: 21 Pages: E1 - E4
Publisher: W.B. Saunders Co. Ltd.

Author:

Decoster, Robin
Mol, Harrie ; Smits, Dirk

Keywords:

Image quality, radiography, VGA, digital detector, Science & Technology, Life Sciences & Biomedicine, Radiology, Nuclear Medicine & Medical Imaging, Radiography, Digital detectors, Post-processing, Image perception, DIGITAL CHEST RADIOGRAPHY, COMPUTED RADIOGRAPHY, DOSE REDUCTION, PERFORMANCE, OPTIMIZATION, SYSTEMS, 1103 Clinical Sciences, Nuclear Medicine & Medical Imaging, 3202 Clinical sciences

Abstract:

Background: The dynamic range of modern detectors tolerates a higher detector dose or Detector Air Kerma (DAK) without negative effect on image quality. Necessary image quality is closely related to the clinical question: in order to keep the patient dose as low as possible, image quality criteria are needed for each type of radiography. The purpose of this study was to investigate the relation between image quality, described as an overall judgment versus as the visibil-ity of well-defined structures, and DAK in clinically accepted radiographs. Methods and Materials: 168 anterior-posterior radiographs of the knee and 152 radiographs of the pelvis were collect-ed randomly in 19 radiologic centers. Six radiologists with at least five-years experience scored the overall image quali-ty and the visibility of seven different anatomic structures, in a controlled environment on a five-point scale, using a Visual Grading Analysis (VGA) . The relation between the DAK and the VGA Score (VGAS) was evaluated. Results: The VGAS was 3.92 for the knee and 3.71 for the pelvis. The VGAS for CR and DR were significantly differ-ent (p<0.01). Intra-observer variability was not significant and inter-observer correlations were high and significant. Only for the pelvis radiographs produced with computed radiography, a rather weak but significant correlation was found between DAK and VGAS. Conclusion: The VGA revealed an image quality higher than diagnostically necessary in both datasets, and a high in-ter-observer correlation. Based on the DAK-range, it could be hypothesized that below a certain noise level no further visible improvement of the image quality was reached.