Improving the accuracy of radiochromic film dosimetry in lung- and soft tissue-equivalent media in a dynamic phantom through dose-to-medium correction factors

<i>Objective.</i>The accuracy of film dosimetry in heterogeneous phantoms is diminished by several factors: (i) the perturbation of the beam by the film in low-density media, (ii) the calibration of film in terms of dose-to-water versus dose-to-medium reported by treatment planning systems (TPSs), and (iii) the motion of the phantom as it is irradiated, which is not resolved in the TPS dose calculation. The Australian Clinical Dosimetry Service is currently developing a film-based end-to-end audit for motion management encompassing lung and liver treatment sites. The purpose of this work was to derive spatially-resolved correction factors to improve the accuracy of audit film measurements, and test the efficacy of these factors in practice.<i>Approach.</i>The correction factor was defined by:kmed=Dbiol-m,biol-mDdet,phant-m. The numerator represents the TPS calculated dose from a dose-to-medium (Type C) algorithm in which the radiation transport properties are adjusted based on the biological material assigned in each voxel. The denominator represents the dose measured by the film, which is embedded in the non-biological phantom media. The EGSnrc egs_chamber user code was used to compute 2D dose planes for both the numerator and the denominator, using a simplified version of the phantom geometry, resulting in 2D correction factor maps. A series of measurements matching the simulated beam geometry/respiratory characteristics were conducted in the actual audit phantom. Corrected and uncorrected measurements were compared to TPS calculations. Agreement was quantified by the distribution of dose differences in all voxels receiving >1 Gy (10% of target dose).<i>Main results.</i>In the lung, corrections reduced the range of dose differences from -9.0%-8.7% to -4.4%-1.8%. In the liver, corrections resulted in a systematic improvement (average difference reduced from 1.5% to 0.7%). Overall, dose differences were reduced from a range of -9.1%-7.2% to within -4.4%-3.4% across all tested combinations.<i>Significance.</i>Correction factors meaningfully improved agreement between measurements and calculations.