Can a commercially available EPID dosimetry system detect small daily patient setup errors for cranial IMRT/SRS?

Purpose: The purpose of this study was to determine if the Sun Nuclear PerFRACTION electronic portal imager device dosimetry software would be able to detect setup errors in a clinical setting and would be able to correctly identify the direction in which the setup error was introduced. Methods and materials: A 7-field intensity modulated radiation therapy (IMRT) treatment plan for a centrally located tumor was developed for 1 phantom and 5 canine cadaver heads. Systematic setup errors were introduced by manually moving the treatment couch by 1, 3, and 5 mm in each translational direction to assess stereotactic radiation surgery (SRS), IMRT, and 3-dimensional (3D) treatment tolerances after the initial alignment was performed. An angular setup error of 5 degrees yaw was also assessed. The delivered treatment fluence was automatically imported in the PerFRACTION software and compared with the baseline fluence. Results: In the canine phantom, a 5-mm shift was undetected by gamma analysis, and up to a 2-cm shift had to be introduced for the gamma pass rate of 3%/3 mm to fall below a 95% pass rate criterion. The same 5-mm shift using 3% difference caused the pass rates for 2 fields to drop below the 95% tolerance. For each respective translational shift, the affected beam angles were consistent across the cadaver heads and correlated with the direction of translational shift. The best field pass rate, worst field pass rate, and average pass rate across all 7 fields was analyzed to develop clinical guidance on parameter settings for SRS, IMRT, and 3D tolerances. Conclusions: PerFRACTION 2-dimensional mode successfully detected setup errors outside the systematic error tolerance for SRS, IMRT, and 3D when an appropriate analysis metric and pass/fail criteria was implemented. Our data confirm that percent difference may be more sensitive in detecting plan failure than gamma analysis. (C) 2017 American Society for Radiation Oncology. Published by Elsevier Inc. All rights reserved.