Evaluation of the hyperspectral multicomponent calibration for plastic scintillation dosimetry

Objective. Accurate and precise dose measurements are critical for the validation of radiotherapy treatments. Plastic scintillation dosimeters (PSDs), when carefully calibrated, offer a solution. This study provides a systematic investigation of different calibration methods for PSDs based on hyperspectral identification of scintillation. Furthermore, performance of these calibration methods is evaluated for a conventional linac and an MR-linac. Approach. For this study, seven calibration types were applied, each considering a different number of physical and optical components, including different combinations. The inclusion of a component was analysed by evaluating the residual, which provided information about the amount of unlabelled signal. The residuals and the measured doses were evaluated for reproducibility, output factors and angular dependence. The error of the dose measurement offers an indication of how effectively the scintillation component is identified. Main results. The addition of calibration components, such as Cherenkov (proximal and distal), fluorescence, and drift correction, resulted in a progressive reduction in the residual. Consequently, a smaller proportion of the signal remained unlabelled. The absolute dose differences across the measurements also decreased with the addition of components. The inclusion of the Cherenkov component yielded the most substantial improvement in accuracy. For example, the dose discrepancy decreased from more than 70% to less than 1% for large field sizes on the MR-linac. Significance. These findings demonstrate that the hyperspectral calibration significantly improves PSD accuracy and precision, particularly with the inclusion of the Cherenkov component. This improvement is observed consistently across different setups and on both the conventional linac and the MR-linac.