Accuracy assessment of the superposition principle for evaluating dose distributions of elongated and curved 103 Pd and 192 Ir brachytherapy sources Elizabeth A.Bannon, Yun Yang, and Mark J. Rivard a) Department of Radiation Oncology, Tufts University School of Medicine, Boston, Massachusetts 02111 (Received 24 January 2011; revised 21 April 2011; accepted for publication 22 April 2011; published 26 May 2011) Purpose: The current brachytherapy dose calculation formalism determines clinical dose distribu- tions using the superposition principle. However, this approach cannot account for intersource attenuation for extended brachytherapy sources such as elongated and curved sources. The purpose of this study is to determine the line segment length required for optimal accuracy of dose calcula- tions in the vicinity of elongated and curved 103 Pd and 192 Ir sources using the superposition tech- nique intrinsic to the AAPM TG-43 formalism. Methods: Monte Carlo (MC) simulations were performed in water for 103 Pd and 192 Ir sources with linear, toroidal, and hairpin geometries. Dose distributions for 0.1–8.0 cm line segments from MC simulations were entered into treatment planning system (TPS) using the TG-43 approach with 0.05 cm spatial resolution. Line source dose distributions were benchmarked using MC-to-MC comparisons with the superposition principle, TPS-to-TPS fluence map comparisons, and MC-to- TPS gamma-index comparisons. Toroidal and hairpin geometries were constructed using line seg- ments in the TPS, then TPS-calculated dose distributions were compared to the MC-simulation results. Gamma-index comparisons were performed in the iba Dosimetry Omni-Pro I’mRT software using 2 mm distance-to-agreement Dd and 2% dose error DD criteria, with a passing rate of 98% of pixels meeting the c  1:00 tolerance deemed acceptable. Results: For the MC-to-MC superposition check for line source, the average ratio of the superposi- tion to the solid source length was 1.051 for 103 Pd and 1.009 for 192 Ir through the whole volume with maximum ratios of 1.34 and 1.32, respectively. TPS-to-TPS comparisons between a solid line source and multiple line segments also provided good agreement. The MC-to-TPS benchmarking indicated where the gamma-index comparison failed were inside the source and within 0.25 cm of the source long-axis. Excluding these regions, 99.6% and 99.9% of the 57 600 in-plane pixels satis- fied the gamma-index criteria for 103 Pd and 192 Ir, respectively. The optimal line segment length for both 103 Pd and 192 Ir toroidal sources was about 0.5 cm or one-fifth of the torus diameter, whichever was smaller. For all toroidal geometries and all line segment lengths examined, at least 98.9% and 100.0% pixels met the gamma-index criteria for 103 Pd and 192 Ir, respectively. For both 103 Pd and 192 Ir hairpin source, all geometric variations had passing rates exceeding 99.2%. However, the best results were obtained from 0.4 cm line segments on the curved part for 103 Pd while it was independ- ent of line segment length for 192 Ir. Conclusions: A method for using a conventional TPS for brachytherapy treatment planning of elongated and curvilinear brachytherapy sources was developed and benchmarked. This method was evaluated using a gamma-index comparison technique, where appropriate pass-rate criteria were identified. Using a variety of subsegment lengths, the total length for a straight-line source was generally reproduced with accuracy improving as subsegment length increased, approaching the total straight-line length. Toroidal sources were similarly modeled with line segments, and an accuracy tradeoff was found between geometric errors and simulating dose anisotropy along the long-axes. The gamma-index comparison method to analyze the results was shown to be more powerful than point-wise comparison methods, and more versatile than dose ratios on the same plane. V C 2011 American Association of Physicists in Medicine. [DOI: 10.1118/1.3590380] Key words: brachytherapy, Monte Carlo, TPS, superposition, 103 Pd and 192 Ir, elongated source, curved source I. INTRODUCTION The current approach for brachytherapy dose calculations relies on the superposition of single-source dose distribu- tions, and is specified in the series of American Association of Physicists in Medicine (AAPM) TG-43 reports (i.e., Nath et al. 1995, Rivard et al. 2004, Rivard et al. 2007). 1–3 The source dose distribution is assumed to be cylindrically sym- metric and obtained in a liquid water phantom of a fixed vol- ume for full radiation scattering. Total dose is calculated by combing dose from each single source. This model works well for conventional brachytherapy seeds with active 2957 Med. Phys. 38 (6), June 2011 0094-2405/2011/38(6)/2957/7/$30.00 V C 2011 Am. Assoc. Phys. Med. 2957