Author: Sarah Aubert, Leigh Conroy, Tony Tadic 👨🔬
Affiliation: Princess Margaret Cancer Centre 🌍
Purpose: Cardiac radioablation is an emerging non-invasive treatment option for patients with refractory or life-threatening ventricular tachycardia (VT), however it is difficult to deliver highly conformal doses to targets that experience complex respiratory and cardiac motion. We propose a workflow for the characterization of respiratory motion on dose delivered during cardiac SBRT using deformable image registration (DIR) and dose accumulation.
Methods: Using a retrospective dataset of lung cancer patients, we developed a DIR workflow between maximum inhale and exhale 4DCT image sets. An open-source tool was used to contour cardiac substructures and proximal luminal organs-at-risk (OARs) on the images. The accuracy of the registration and auto-contoured structures was evaluated on five patient datasets using a visual assessment scoring tool. The impact of motion between max inhale and exhale datasets was evaluated by quantifying the mean deformable vector field (DVF) for the left and right ventricles, left atrium, whole heart, stomach and esophagus.
Results: The superior-inferior (SI) motion of the dome of the liver between inhale and exhale images was used as a benchmark for respiratory motion, with a median of 1.0cm [0.6-1.7cm]. The left ventricle, which contains the target, had a median DVF of 0.7cm [0.4-1.5cm]. The surrounding cardiac structures, left atrium and right ventricle, had DVF of 0.8cm [0.5-1.4cm] and 0.5cm [0.2-0.8cm], while nearby luminal OARs had DVF of 0.6cm [0.2-1.0cm] and 0.5cm [0.4-0.9cm]. Consistent with previous literature, the greatest motion for all evaluated substructures was in the SI direction and the smallest motion was left-right.
Conclusion: The results from our work will enable recommendations for planning and motion management techniques for cardiac radioablation. Future work will assess the dosimetric impact of the motion on a subset of cardiac substructures, in addition to the application of the pipeline to more patients and the incorporation of models of cardiac motion.