3D Dose Reconstruction of Pretreatment Verification Plans Using Multiple 2D Planes from the OCTAVIUS/Seven29 Phantom Array

Abstract

The purpose of this study is to evaluate 3D dose reconstruction of pretreatment verification plans using multiple 2D planes acquired from the OCTAVIUS phantom and the Seven29 detector array. Eight VMAT patient treatment plans of different sites were delivered onto the OCTAVIUS phantom. The plans span a variety of tumor site locations from low to high plan complexity. A patient specific quality assurance (QA) plan was created and delivered for each of the 8 patients using the OCTAVIUS phantom in which the Seven29 detector array was placed. Each plan was delivered four times by rotating the phantom in 45° increments along its longitudinal axis. The treatment plans were delivered using a Novalis Tx with the HD120 MLC. Each of the four corresponding planar doses was exported as a text file for further analysis. An in-house MATLAB code was used to process the planar dose information. A cylindrical geometry-based, linear interpolation method was utilized to generate the measured 3D dose reconstruction. The TPS calculated volumetric dose was exported and compared against the measured reconstructed volumetric dose. Dose difference, dose area histograms (DAH), isodose lines, profiles, 2D and 3D gamma were used for evaluation. The interpolation method shows good agreement (<2%) between the planned dose distributions in the high dose region but shows discrepancies in the low dose region. Horizontal profiles, dose area histograms and isodose lines show good agreement for the sagittal and coronal planes but demonstrate slight discrepancies in the transverse plane. The 3D gamma index average was 92.4% for all patients when a 5%/5 mm gamma passing rate criteria was employed but dropped to <80.1% on average when parameters were reduced to 2%/2 mm. A simple cylindrical geometry-based, linear interpolation method is able to predict good agreement in the high dose region between the reconstructed volumetric dose and the planned volumetric dose. It is important to mention that the interpolation algorithm introduces dose discrepancies in small regions within the high dose gradients due to the interpolation itself. However, the work presented serves as a good starting point to establish a benchmark for the level of manipulation necessary to obtain 3D dose delivery quality assurance using current technology.