Eliminating Setup Uncertainty In Small Field Machine Matching: An Epid-Based Validation of Dosimetric Equivalence
Abstract
Purpose
To establish baseline Electronic Portal Imaging Device (EPID) data for validating dosimetric equivalence across a fleet of Varian TrueBeam linear accelerators. The primary objective was to reduce the inter-user and setup variability associated with standard water-tank measurements by implementing a consistent, equipment-based verification standard.
Methods
A single standardized DICOM RT plan was distributed to a fleet of matched linear accelerators. The plan consisted of open fields (2x2cm to 20x20cm), MLC-shaped fields, and EDW fields. Measurements were acquired for 6, 10, and 15MV (WFF) and 6 and 10MV (FFF) energies using the Varian Portal Dosimetry system. Central axis (CAX) calibrated units (CU) were analyzed to quantify inter-machine consistency. Early validation testing indicated that the AS1000 panel series exhibited inconsistent response characteristics; consequently, these devices were excluded. These results were compared against baseline commissioning data measured with ionization chambers and appropriate small field detectors.
Results
Valid EPID measurements demonstrated excellent agreement across the fleet, detecting output variations with equal or greater sensitivity than standard detector measurements. CAX consistency was maintained even across machines utilizing differing calibration dose profiles, demonstrating the method's robustness to calibration variations. For 6MV open fields, the inter-machine standard deviation was 0.25%, comparable to the 0.19% baseline variation established during commissioning. For measuring challenging 2x2cm small fields, EPID proved superior to manual methods. While standard measurements for 15MV 2x2cm fields showed a variation of 1.2% (attributed to setup uncertainty), EPID measurements reduced this variability to 0.12%. EDW factors measured via EPID agreed with the fleet mean within 0.32%.
Conclusion
Portal Dosimetry measurements provide a rapid and strictly reproducible method for verifying that multiple linear accelerators are dosimetrically equivalent. This methodology is particularly valuable for small field verification, where it eliminates the setup uncertainties inherent in manual detector placement, serving as a reliable baseline for routine machine validation.