Tuesday Case of the Day Physics Authors: Mark P. Supanich, PhD; Phil Rauch, MS; Susan M. Lang, MS; Michael Flynn, PhD Department of Radiology, Henry Ford Health System Detroit, MI HISTORY: X-ray technicians suspect ghosting as the cause of artifacts overlaying anatomy on 2nd and 3rd images of a whole leg stitching series on a digital indirect detector system. However, Images have normal exposure index values of 400, 500 and 500 , which seems to obviate ghosting as the cause of the artifact. What is the best solution to mitigate the artifact? A. There is no artifact, patient is wearing compression garment(s) which are constricting the tissue in this image B. Reduce exposure to eliminate image lag C. Collimate to exclude areas exterior to the hips and long bones. D. Increase time between exposures to allow for complete erasure of detector residual signal Figure 1: Three image whole leg series with 8 second delays between acquisitions on an a-Si indirect detector. FINDINGS: Physicists compared each image of the 3-image series and found that the outline of the patient’s anatomy in the first image reappeared at the same detector locations in subsequent images. For the second and third images, the detector signal was greater in the regions to the right of the outline. 85 kVp 195 mAs 66 kVp 6.5 mAs 66 kVp 3.4 mAs Figure 2: Images (cropped) which compose the stitched image in Figure 1. FINDINGS: The image of a ruler from another patient with the same artifact demonstrated that the artifact is due to increased detector signal rather than signal retention. In Figure 3 a previous exposure of an unattenuated region with a lead ruler overlay resulted in signal changes in subsequent images. The phenomenon shown is known as “bright burn” or “after-glow” and results in increased signal in the highly exposed region. The bright region on the number zero has less signal because this is where the detector was masked by the tick mark on the ruler in the previous image and thus the bright burn was not present. The rest of the number zero has increased signal due to bright burn. The right portion of the base of the numbers 2 and 4 have more signal than that rest of the number because of the bright burn phenomenon. Figure 3: Region of detector receiving high exposure demonstrates bright burn in the subsequent image Diagnosis: Correct answer is C: Collimate the exposures to exclude areas exterior to the hips and long bones. This approach will protect the unattenuated area of the detector from the high exposure and eliminate the manifestation of increased signal in subsequent images. DISCUSSION: The artifact noted by the technicians is due to extreme over-exposure of the unattenuated areas of the first image, resulting in increased signal in the same detector regions of the subsequent images. The fundamental cause of the artifact is known as bright burn, which occurs due to light yield enhancement in the CsI(Tl) scintillator. There are two additional mechanisms that will produce a subsequent increase in signal in regions subjected to high exposure. These are (1) photodiode residual signal, and (2) photodiode gain effect. Photodiode residual signal is usually corrected via the use of reset lights, so it is unlikely that this was a significant factor. However, both bright burn from the CsI(Tl) and the photodiode gain effect likely contributed to the production of the artifact. While the resultant Exposure Index (EI) values were indeed normal, the actual exposure technique required for the first image was quite high, causing the areas of the detector with no overlying tissue to be exposed to a x-ray fluence. EI is derived from pixel values in the segmented, anatomical portion of the image, and as this region contained normal pixel values, a normal EI was reported. It is important to keep in mind that a normal EI does not provide any information about areas outside of the segmented area. In this case, the detector area which was over-exposed was outside the segmented area and contributed no information to the EI. Answer A is incorrect because although the patient was wearing a constricting garment around the waist in the first image, there was no such garment on the anatomy in the second or third images. Answer B is not a good solution because reducing the exposure to the level required to eliminate the artifact would result in an inadequate image requiring re-exposure. Answer D is incorrect because bright burn has been known to manifest for hours or longer. The manufacturer has suggested the utilization of their “stitching filters”, which are a form of equalization filters designed for use in whole spine imaging. The filters are placed lateral to the neck to mediate high exposures outside the skin line. However, the filters did not match the anatomy of the patient in the current whole leg study. A final alternative might be to acquire the images in reverse order, saving the highest exposure image for last. Unfortunately, the system did not allow acquisition from the bottom up. References/Bibliography: • Wieczorek, H and Overdick, M. Afterglow and Hysteresis in CsI:Tl. Proc. 5th International Conf on Inorganic Scintillators and their Applications 2000; 385-390. •Yorkston, J. Recent Developments in Digital Radiography Detectors. Nuclear Instruments and Methods in Physics Research A 2007; 974-985. •Overdick, M et al. Temporal Artefacts in Flat Dynamic X-Ray Detectors. Proc of SPIE Vol 2001; 4320:47-58. •Chotas, HG et al. Principles of Digital Radiography with Large-Area, Electronically Readable Detectors: A Review of the Basics. Radiology 1999; 210:595-99. •McDermott, LN et al. Dose Response and Ghosting of an Amorphous Silicon Electronic Portal Imaging Device. Med Phys 2004; 31:285-95. •Bloomquist, AK et al. Lag and Ghosting in a Clinical Flat-Panel Selenium Digital Mammography System. Med Phys 2006; 33:2998-3005.