Keywords:
Quality assurance, Diagnostic procedure, Digital radiography, Radiographers
Authors:
S. H. Al-Murshedi, P. Hogg, A. England; Manchester/UK
DOI:
10.1594/ecr2018/C-1353
Methods and materials
Image acquisition
The anthropomorphic adult chest Lungman phantom produced by ‘Kyoto Kagaku’ Kyoto Kagaku,
(with and without a ‘fat jacket’ to represent increased BMI) [4] (Figure 1) was used to simulate the normal size and obese adult patient respectively.
The phantom was imaged in 17 X-ray units located in eight different hospitals within the north west of England using their existing routine adult chest x-ray protocols,
and the images were collected from the hospitals in DICOM format.
Perceptual image quality evaluation
The general image quality (IQ) and the lesion visibility (LV) for the phantom images for both the normal size and obese patients were evaluated visually using a relative visual grading analysis (relative VGA) method [5] by six observers (qualified diagnostic radiographers).
Six criteria have been used for general image quality evaluation based on European guidelines [6] and a study by [7],
while three criteria used for lesion visibility evaluation were based on a study by [7].
The 17 images from the 17 X-ray units were evaluated against a reference image on two 5 mega-pixel monochrome liquid crystal display (LCD) monitors which were calibrated to the DICOM grey scale standard,
as shown in Figure 2.
Relative VGA allowed the observers to evaluate the image as worse,
equal to,
or better than that of the reference image.
Physical image quality evaluation
The signal to noise ratio (SNR) and contrast to noise ratio (CNR) were calculated as a physical method for assessing image quality using ImageJ analysing software for JAVA.
SNR was computed as ratio between signal (average pixel values of region of interest ) and noise (standard deviation of the background) (SNR = S /σb) which is considered a direct method for measuring SNR based on study by[8] ,
while CNR was computed as the ratio of the difference between signal and noise divided by the noise value (CNR = S-B/σb) [8].
In addition,
conspicuity index (CI) was measured physically for selected feature from Lungman phantom images using conspicuity software [9].
Radiation dose measurements
The Unfors solid state dosimeter (RaySafe X2) [10] was utilised for dosimetry measurements and it was located at the surface of the phantom at the point of intersection of the X-ray beam.
The entrance surface dose (ESD) was calculated after including the backscatter factor (BSF) to the incident air kerma (IAK) values resulting from the unfors dosimeter.