Author: Piotr Pater, Nicholas G. Zacharopoulos 👨🔬
Affiliation: McGill University, Medical Physics Unit, Gerald Bronfman Department of Oncology, McGill University 🌍
Purpose: To develop a method for accurate relative film dosimetry that incorporates non-uniformity
corrections.
Methods: The accuracy of Devic’s [1] linearization approach was evaluated using simulations with mea-
sured dose-response curves on datasets with maximum dose values ranging from 0.2 to 10 Gy. Based on
the results of these simulations, the linearization approach was refined to incorporate parameter optimiza-
tion for increased accuracy, and simulations were repeated for comparison. The optimized linearization
was then combined with Micke’s [2] multichannel dosimetry (MCD) to account for non-uniformities. This
new relative optimized linearization (ROL) was then validated via a comparison against MCD on three
datasets: open field, wedge field, and a VMAT plan. To confirm the robustness of this new ROL method, the
VMAT plan was further evaluated with induced errors representative of real clinical scenarios, including
positional and dosimetric errors.
Results: The optimized linearization reduced maximum errors from up to 7% in the red channel for EBT3
and 5% in the green channel for EBT4 to below 2% across all dose ranges and channels for both film
types. ROL effectively accounted for non-uniformities and showed excellent agreement with MCD across
all three channels. In some cases, ROL outperformed MCD by producing dose values closer to those from
in-water ion-chamber scans. Additionally, ROL demonstrated robustness in the presence of positional
and dosimetric errors.
Conclusion: The proposed techniques provide a practical and efficient approach to implementing rela-
tive film dosimetry with non-uniformity corrections without the need for time-consuming batch-specific
dose-response curves.