Rejection and redistribution of scattered radiation in scan equalization digital radiography (SEDR): Simulation with spot images

Abstract

The anti-scatter grid has been widely used to reject scatter and increase the perceptibility of a low-contrast object in chest radiography; however, it also attenuated the primary x-rays, resulting in a substantial loss of information and an increased relative noise level in heavily attenuated regions. A more dose efficient approach to scatter rejection is the slot-scan imaging technique. Another problem in chest radiography is the low transmitted x-ray intensity in heavily attenuating regions. It results in a higher relative noise level, thus limiting the contrast sensitivity. A solution to this problem is through the exposure equalization technique, with which the incident x-ray intensity is regionally modulated to compensate for the differences of x-ray attenuation due to the anatomic variation. We are in the process of implementing the scan equalization digital radiography (SEDR) technique, which combines the advantages of slot-scan imaging and exposure equalization. However, associated with the use of exposure equalization is a redistribution of scattered radiation at the detector, which may impact on the benefit of using exposure equalization in conjunction with the slot-scan imaging geometry. In order to understand the scatter properties and their impact in SEDR, we have used spot collimated digital radiographic images to synthesize simulated SEDR images with which scatter components, primary signals, and scatter-to-primary ratios (SPRs) were measured. It was shown that the anti-scatter grid rejected approximately 70% and 80% of scattered radiation in lightly and heavily attenuated regions, respectively, while the slot-scan method can reject as high as 95% (with 1 cm slot width) of scattered radiation without attenuation of the primary x-rays. Using a simple model for scatter effects, we have also estimated and compared the contrast-to-noise ratio degradation factors (CNRDFs, i.e., the fraction by which CNR is reduced). It was found that for quantum limited situations, the slot-scan technique has resulted in a substantial improvement of the image quality, as indicated by higher estimated CNRDFs (less scatter). An estimated improvement of 40%-50% in the lungs, 50%-90% in the mediastinum, and 60%-110% in the subdiaphragm was achieved with the slot-scan over the anti-scatter grid method. Compared to slot-scan imaging, SEDR resulted in higher SPRs in the lungs and lower SPRs in the mediastinum. In the subdiaphragmatic regions, the SPRs remain about the same. This corresponds to lower CNRDFs in the lungs, higher CNRDFs in the mediastinum, and about the same CNRDFs in the subdiaphragmatic regions. It was shown that although SEDR has resulted in minimum improvement over slot-scan imaging in reducing the SPRs, it could improve the contrast sensitivity by raising the primary signal levels in heavily attenuating regions. This advantage needs to be further investigated in our continuing study of the SEDR technique.