Investigation of dose optimization by air gap method in spine and pelvic radiography

Abstract

Radiation safety to the patients in medical imaging is always of high importance. Spine and pelvic radiographic examinations that involve a relatively high radiation dose to the radiosensitive organs of the patients are of particularly concern. A major reason for the relatively high dose being delivered to patients in these examinations is caused by the application of an anti-scatter grid. To this end, the main theme of this study was to explore the feasibility of replacing the anti-scatter grid by an optimal air gap in spine and pelvic radiography to maximize the dose reduction effect in patients while retaining the resultant images of diagnostic quality in computed radiography (CR) and digital radiography (DR) systems using an anthropomorphic RANDO phantom. To gain a holistic picture on the mechanism of this dose reduction method, it was important to uncover the scientific basis on the dose and scatter distribution patterns in diagnostic radiography. One objective of this research project was to uncover the air kerma distribution patterns of the x-ray tube output and the subsequent scatter distribution within and around a homogenous scattering medium by generating the 3D dose maps. These were achieved by the physical measurement of air kerma at selected sites by an ionization chamber detector or thermoluminescent dosimeters and, coupled with Monte Carlo (MC) simulations on the measured data. Results on the investigation of air kerma distribution patterns among the three different brands of x-ray tube revealed the unique skewed shape 3D distribution maps from the output of the studied x-ray tubes. This finding filled the knowledge gap on the anode heel effect investigation as was reflected on the characteristics variation of radiation intensity output of the x-ray tube. Findings on the investigation of air kerma distribution patterns on the water-equivalent dry slabs phantom by MC simulation on the photon transport uncovered the non-uniform distribution of scatter inside and around the scattering medium, and were presented in three-dimensional maps. Results on the investigation of dose reduction by the air gap method using the RANDO phantom confirmed that significant dose reduction effect with retention of image quality at diagnostic level was confirmed in all of the tested radiographic examinations in both the CR and DR systems. These examinations included cervical spine, thoracic spine, lumbar spine, pelvis and full spine. Image Quality Score (IQS) and Visual Grading Analysis (VGA) tests were employed as the image evaluation methods for the assessors to determine the image quality of the images in according to the pre-set criteria. Among the non-grid images with different air gap thicknesses in place, those images being rated with the highest scores in both tests were considered as the best in image quality. Results on the IQS and VGA scores uncovered that the optimal air gap thickness for each of the tested radiographic examinations was in the range of 5cm to 20cm. By replacing the anti-scatter grid with an optimal air gap thickness, a substantial dose reduction effect ranging from 46.7% to 81.6% at the regions of radiosensitive organs of the tested RANDO phantom was evidenced. Effective doses simulated from the PCXMC 2.0 program further revealed that a substantial dose reduction effect ranging from 50.3% to 80% was found in the non-grid examinations with the optimal air gap thickness, which indicated that the risks on cancer induction could be reduced in the similar fashion. Of the same kind of radiographic examination, a further reduction in both the effective and organ doses by 3.8% to 18.2% were achieved while changing the imaging system from CR to DR. In conclusion, the use of an optimal air gap to replace the anti-scatter grid was proven to be an effective dose reduction method on high dose spine and pelvis radiographic examinations in both the CR and DR systems with image quality retained at the diagnostic level as was proven in the present study on a RANDO phantom model.