(Tuesday, 7/16/2019) 10:00 AM - 10:30 AM
Room: Exhibit Hall | Forum 4
Purpose: To determine the pixel sensitivity map (PSM) for an amorphous silicon electronic portal imaging device (EPID) using a single flood field signal.
Methods: A raw acquired EPID signal may be decomposed into the incident particle fluence signal, the inherent pixels gain, and the background signal. Particle fluence varies slowly and is locally spatially-constant. Pixels response is a fast and abrupt varying signal which perturbs the local fluence response. The background signal is due to the EPID panel electronics, and is determined during radiation absence. To determine the PSM, we first correct for the background signal. Then we apply a regression model that captures the underlying slowly varying features of the corrected EPID signal. The captured fluence signal is then used to decouple the PSM signal from the corrected EPID signal. To validate the generated PSM, we confirm that an open field output is EPID-position independent and compare our results to other published PSM generation methods.
Results: The EPID pixel gain values are normally distributed with mean value of 1.0 and standard deviation of 0.01 for 6MV beams with and without the flattening filter. Model validation shows that a PSM, generated with this method, alters an open field output value by <1.0%. Flattening-filter free beams are found to generate PSMs which are well-described by regression models. Post PSM application, the processed signal is EPID-position independent.
Conclusion: We present an approach to generate the PSM for linac MV EPIDs. The method is based on the fact that the EPID signal may be decomposed into a smooth surface convolved with spatially-dependent pixels gain map. This method does not require shifting the EPID panel , enabling it PSM generation for linacs with fixed EPIDS, nor requires multiple EPID panel irradiations. This approach improves EPID transit dosimetry and enables automatic EPID signal calibration.
Funding Support, Disclosures, and Conflict of Interest: This work is partially supported by VARIAN Medical System.
Keywords
Not Applicable / None Entered.
Taxonomy
Not Applicable / None Entered.
Contact Email
