Room: AAPM ePoster Library
Purpose: In vivo EPID measurements are able to detect changes in patient anatomy such as weight loss. However, shifts in the patient position after IGRT escape detection using current approaches. We propose an analysis method which shows promise for detecting and quantifying the magnitude of in-vivo patient shifts.
Methods: An anthropomorphic phantom with material equivalent to water, bone and liver was shifted by 5, 10, 15, and 20mm in all directions with in-vivo EPID images collected for each position. An analysis of the difference of two images in the high dose region was performed. Profiles of the image difference were investigated to extract information of the shifts. Additionally, images of patient treatments with suspected shifts were investigated.
Results: Comparing EPID images of the phantom with relative shifts, profiles of the image difference produces peaks where the full width half maximum estimates the magnitude of the shift, and the separation of the peaks estimates the object size. For the interface of bone and water an estimate of the shift with an error of 3 mm was found. For the interface of water and liver, the profiles contained more noise and the error increased to 5 mm. Only a projection of the shift perpendicular to the IMRT beam can be measured, more than one IMRT beam is required to get a 3D shift. For patient treatments, shifts of bony anatomy were observed and estimated to be on the order of 8 mm.
Conclusions: For IMRT beams shifts in bony anatomy perpendicular to the beam can be measured with an error of 3 mm and shifts in liver can be measured with an error of 5 mm. With knowledge of the magnitude of the shift, this technique can be used to estimate the dosimetric impact of patient shifts.
Funding Support, Disclosures, and Conflict of Interest: Supported by grant funding from Varian Medical Systems.