SU-E-T-93: A CT Polymer Gel Dosimetry System for End-To-End Dosimetry

Abstract

Purpose: To design and test an x-ray CT polymer gel dosimetry system for 3D end-to-end dosimetry of brain and head and neck radiotherapy Methods: A head and neck phantom has been designed to undergo the entire RT process from immobilization and planning CT through to treatment. The phantom houses a 1L polymer gel in one of two locations: cranial and inferiorly in the neck region allowing for quality assurance of both brain and head and neck treatment processes. The gel dosimeter measures the delivered radiation dose in 3D, and following CT read-out, provides a dose record of the entire treatment process, i.e. “end-to-end” dosimetry. Enhanced dose-sensitivity N-isopropylacrylamide (NIPAM) based gels along with an optimized CT protocol were used to minimize uncertainty in image CT numbers (H). Tests of phantom set-up were performed to quantify image noise, uniformity and positioning reproducibility using locking bar (ideal conditions) and aquaplast mask (clinical conditions). The end-to-end dosimetry capability was tested by undertaking the full RT process (immobilization, planning CT, treatment planning, treatment set-up and delivery) and comparing delivered and planned doses for a simple “star-pattern” irradiation. Results: Removing the phantom head for gel read-out minimized noise and artifacts (63% noise reduction), produced uniform images and has minimal impact on phantom re-positioning (less than 0.5mm). Full phantom re-positioning reproducibility was also excellent: less than 0.9 mm for both locking bar and mask immobilization. Initial end-to-end dosimetry tests indicate accurate localization of treatment dose to within 1 mm. Conclusions: An x-ray CT polymer gel dosimetry system for performing 3D end-to-end dosimetry has been designed, tested and is demonstrated to provide accurate 3D localization of delivered radiation dose. Future work will assess clinical processes as undertaken by appropriate RT staff and look at other treatment sites.