Keywords:
Dosimetric comparison, Radiation safety, Equipment, Dosimetry, Digital radiography, Radioprotection / Radiation dose, Radiographers
Authors:
E. Higgins, K. Matthews; Dublin/IE
DOI:
10.26044/ecr2019/C-1982
Methods and materials
Design
An anthropomorphic PIXY phantom [6] was exposed for AP lumbar spine and pelvis examinations using a digital radiography system.
AP pelvis and lumbar spine projections are amongst the most frequently performed projection radiography examinations where AEC is used [7].
Investigation of these two projections allowed all three AEC chambers to be incorporated into the research [8].
Exposure Conditions
Source to image distance (SID),
centring and collimation were optimal [9] and consistent throughout.
A stationary,
transparent,
115cm grid was used as per clinical routine. DAP was recorded from the inherent DAP meter,
which was appropriately calibrated.
Delivered mAs and reached exposure values (EXI) were recorded from the examination protocol displayed post exposure.
AEC Manipulation
The AEC system permitted the selection of nine exposure correction factor (ECF) settings (-2 to +2 in 0.5 increments) and five detector sensitivities (DS) (200,280,400,560,800).
Selection of a lower ECF results in earlier termination of exposure and thus relatively lower dose delivery.
Selection of a lower DS results in later termination of exposure and thus relatively higher dose delivery [10].
All possible combinations of ECF and DS were applied resulting in forty-five exposure combinations for each of AP lumbar spine and pelvis examinations. Each combination was exposed three times in order to minimise random error in the average DAP values recorded [11].
The images obtained with the AEC settings prevailing in clinical practice (ECF=0,
DS=400) were used as the reference point for DAP and image quality.
Image Quality Analysis
Only images associated with a DAP dose lower than the reference point were analysed by two methods.
Identical regions of interest were applied in identical positions over an anatomically homogenous area in each pelvis and lumbar spine image (Fig. 1 and Fig. 2).
Standard Deviation of pixel intensity values provided a simple indicator of image noise in each image.
Three observers then independently appraised each image.
Three criteria evaluating image detail in areas of high and low attenuation,
and image noise in an homogenous region were applied (Table 1).
All judgements were made in comparison with the reference image (Table 2).
In this Visual Grading Analysis (VGA),
scores <3 indicated image quality was less than diagnostic,
scores ≥3<6 indicated image quality was diagnostic but not as good as the reference image,
while scores of ≥6 indicated an image had the same or better quality than the reference image.
Inter-observer score analysis showed strong correlation between the three observers thus validating the objectivity of the grading criteria (Table 3).