A method for assessing the dosimetric consistency of single phase 4DCT dose accumulation based on deforming image registration

Abstract

Introduction The interplay effect in FFF-VMAT SBRT deliveries can produce clinically relevant discrepancies between planned and delivered doses to moving targets. To assess the impact of motion on treatment delivery, dose distributions calculated on each phase of a 4DCT scan can be accumulated in a single phase using deforming image registration (DIR). The use of this technique as a treatment quality assurance tool requires dosimetric validation. Purpose To develop a method for assessing the accuracy of DIR-based dose accumulation on 4DCT scans. Materials and methods Software was developed to interface with Eclipse TPS to calculate dose distributions on moving targets. This software splits each VMAT arc into partial arcs synchronized with the 4DCT scan. Dose matrices calculated on each breathing phase are accumulated on a single reference phase using the DIR dose accumulation algorithm in MIM Maestro software. For treatments planned on a cylindrically symmetric phantom with the target moving parallel to the symmetry axis and perpendicular to the VMAT arc’s planes, dose distributions calculated with a moving X-ray source are equivalent to those calculated with a moving phantom. Using a moving source to evaluate dynamic deliveries does not rely on DIR dose accumulation and therefore provides a novel method for verifying DIR-based dose accumulation algorithms on 4DCT datasets. Both methodologies were used for calculating dose distributions on the 20% phase of a 4DCT of a QUASAR phantom fitted with a cylindrical cedar insert containing a 3 cm acrylic target. The insert was moved following a 14 BPM realistic breathing trace with total amplitudes of 2 cm and 1 cm. Planar dose distributions along the principal axis were compared using gamma analysis. Results Gamma analysis was performed using Δ D = 2%, Δ x = 2 mm (10% D max threshold). The minimum passing rates were 99.1% and 99.2%, corresponding to the sagittal planes of the 2 cm and 1 cm amplitude movement. Conclusion The methodology presented here provides an easy and reliable way of assessing the performance of 4DCT DIR dose accumulation. The tested DIR accumulation algorithm performed exceptionally well, considering the ill-posed registration problem.