Adaptive Calibration of Gafchromic EBT3 Films Using Generalized Additive Neural Networks

Abstract In radiation therapy, achieving precise dose distribution is crucial to minimize damage to normal tissues. Gafchromic EBT3 film is commonly used for quality assurance of 2D dosimetry but requires frequent recalibrations due to changes in sensitivity over time. This study introduces a new calibration method using a Keras generalized additive neural network (GANN) to address film aging. EBT3 films from four lots were calibrated using a 6 MV photon beam and scanned with Epson scanners. The GANN method showed percentage differences between calibrated and delivered doses within 5%, comparable to traditional recalibration methods. It demonstrated better stability and accuracy, significantly reducing the need for frequent recalibrations and offering a robust solution for long-term film dosimetry.