Purpose: Ashland Inc. EBT3 film is a widely used dosimetry quality assurance tool. The film usually is recalibrated each quarter due to the ‘aging effect’, and calibration uncertainties always exist between individual films even in the same lot. Instead of performing recalibration, a new method is proposed using an adaptive power function with a deep neural network (DNN) model.
Methods: EBT3 film was calibrated several times at different dates. Following “one red-channel after three-channel method”, the dose-dependent optical density (OD) of the red channel is fitted to the delivered dose using a power function with a constant fitting parameter. Instead of recalibrations, film-dose can be calculated using all the fitting parameters of the first calibration and the fresh measured background OD. A bias τ is added to the input OD to improve the fitting results and is trained using a designed DNN through the Keras functional Application Program Interface (API). The given dose is compared with the film-doses calculated using the presented DNN API method, using all new recalibrated fitting parameters, using all first fitting parameters, and using the adaptive method without τ.
Results: The results demonstrate that the difference between the film-dose with the delivered dose using the adaptive DNN API method is within 3.5%. Without using τ, the differences may reach 7% for a delivered dose lower than 50 cGy. If using the fitting parameters of first calibration, the difference is even greater than 50% around a given 50 cGy.
Conclusion: With the DNN model, the film-dose can be calculated using the first calibrated form and using the new measured background OD to adapt the constant fitting parameter rather than recalibrating the film. To precisely calculate the film-dose, this method can be adapted especially for each individual film.
Funding Support, Disclosures, and Conflict of Interest: This work was supported in part by the Ministry of Science and Technology of Taiwan (MOST 108-2221-E-214-006 and MOST 109-2221-E-214-003)