Commissioning of a grid-based Boltzmann solver for cervical cancer brachytherapy treatment planning with shielded colpostats.

Abstract Purpose We sought to commission a gynecologic shielded colpostat analytic model provided from a treatment planning system (TPS) library. We have reported retrospectively the dosimetric impact of this applicator model in a cohort of patients. Methods and Materials A commercial TPS with a grid-based Boltzmann solver (GBBS) was commissioned for 192 Ir high-dose-rate (HDR) brachytherapy for cervical cancer with stainless steel–shielded colpostats. Verification of the colpostat analytic model was verified using a radiograph and vendor schematics. MCNPX v2.6 Monte Carlo simulations were performed to compare dose distributions around the applicator in water with the TPS GBBS dose predictions. Retrospectively, the dosimetric impact was assessed over 24 cervical cancer patients’ HDR plans. Results Applicator (TPS ID #AL13122005) shield dimensions were within 0.4 mm of the independent shield dimensions verification. GBBS profiles in planes bisecting the cap around the applicator agreed with Monte Carlo simulations within 2% at most locations; differing screw representations resulted in differences of up to 9%. For the retrospective study, the GBBS doses differed from TG-43 as follows (mean value ± standard deviation [min, max]): International Commission on Radiation units [ICRU] rectum (−8.4 ± 2.5% [−14.1, −4.1%]), ICRU bladder (−7.2 ± 3.6% [−15.7, −2.1%]), D 2cc-rectum (−6.2 ± 2.6% [−11.9, −0.8%]), D 2cc-sigmoid (−5.6 ± 2.6% [−9.3, −2.0%]), and D 2cc-bladder (−3.4 ± 1.9% [−7.2, −1.1%]). Conclusions As brachytherapy TPSs implement advanced model-based dose calculations, the analytic applicator models stored in TPSs should be independently validated before clinical use. For this cohort, clinically meaningful differences (>5%) from TG-43 were observed. Accurate dosimetric modeling of shielded applicators may help to refine organ toxicity studies.