Author: Jack Neylon 👨🔬
Affiliation: Department of Radiation Oncology, University of California, Los Angeles 🌍
Purpose: Commissioning surface applicators is notoriously laborious. In addition, when commissioning Varian surface applicators with vertically positioned source, we found limited data and resources available to compare and validate our results. To facilitate a consistent and repeatable analysis, python scripts were developed and shared on a public github repository along with our clinical results.
Methods: Gafchromic films were exposed to acquire in-plane and cross-plane dose distributions for each applicator. Films were sandwiched between solid water slabs. For vertically oriented sources, it is not trivial to identify the central axis due to self-shielding. First, a cross profile was taken at 4mm depth beyond dose maximum, and central axis was assigned as the center point of the largest full width half maximum. Cross profile films were acquired at 3 and 5mm depth. Gaussian blur was applied, then masked at 10% dose maximum, and then either a disk or ellipse was fit to the image to find the center and radius. Once the center was found, minimum spanning disks were acquired at additional isodose levels of interest. Vertical and horizontal profiles were compared to assess centricity of the source and dose distribution. For elliptical applicators, major and minor axes were also identified during the fitting.
Results: Results were consistent with the data available from literature within 5% for PDD at depths greater than or equal to 2mm. Reasonable agreement was also found for 90% isodose diameters, given that Iftimia et al (2016) reported values at 4mm depth, and considering the likely variability in normalization methods.
Conclusion: The source code and full results from our analysis has been made available in a public github repository. We hope this data and methodology will prove useful to the field, and look forward to comparing results.