SU‐E‐T‐50: A Multi‐Institutional Study of Independent Dose Verification Software Program for Lung SBRT

Abstract

Purpose:The accuracy of dose distribution depends on treatment planning system especially in heterogeneity‐region. The tolerance level (TL) of the secondary check using the independent dose verification may be variable in lung SBRT plans. We conducted a multi‐institutional study to evaluate the tolerance level of lung SBRT plans shown in the AAPM TG114.Methods:Five institutes in Japan participated in this study. All of the institutes used a same independent dose verification software program (Simple MU Analysis: SMU, Triangle Product, Ishikawa, JP), which is Clarkson‐based and CT images were used to compute radiological path length. Analytical Anisotropic Algorithm (AAA), Pencil Beam Convolution with modified Batho‐method (PBC‐B) and Adaptive Convolve (AC) were used for lung SBRT planning. A measurement using an ion‐chamber was performed in a heterogeneous phantom to compare doses from the three different algorithms and the SMU to the measured dose. In addition to it, a retrospective analysis using clinical lung SBRT plans (547 beams from 77 patients) was conducted to evaluate the confidence limit (CL, Average±2SD) in dose between the three algorithms and the SMU.Results:Compared to the measurement, the AAA showed the larger systematic dose error of 2.9±3.2% than PBC‐B and AC. The Clarkson‐based SMU showed larger error of 5.8±3.8%. The CLs for clinical plans were 7.7±6.0 % (AAA), 5.3±3.3 % (AC), 5.7±3.4 % (PBC ‐B), respectively.Conclusion:The TLs from the CLs were evaluated. A Clarkson‐based system shows a large systematic variation because of inhomogeneous correction. The AAA showed a significant variation. Thus, we must consider the difference of inhomogeneous correction as well as the dependence of dose calculation engine.