Analyzing Cardiorespiratory Motion and Its Dosimetric Effect on Stereotactic Arrhythmia Radio-Ablation: A STOPSTORM.eu Consortium Study

PURPOSE: Stereotactic arrhythmia radio-ablation (STAR) is increasingly used to treat refractory ventricular tachycardia. The importance of cardiorespiratory motion during STAR is still poorly understood. We collected multicenter STOPSTORM.eu data to estimate cardiorespiratory motion and its dosimetric impact on STAR-deliveries.

METHODS AND MATERIALS: Data were collected from 61 ventricular tachycardia patients. Treatments were performed in free-breathing (FB) with an internal-target volume (FB:ITV) (25 patients), in FB with respiratory beam tracking (FB:TR) (11 patients), using abdominal compression (AC) (10 patients), or in deep inspiration breath-hold (BH) (15 patients). Reference computed tomography images were registered to all breathing/cardiac phases, and the resulting deformation vector fields were used to extract motion for target structures and auto-contoured cardiorespiratory structures (CSs). Additionally, CS were auto-contoured on all breathing/cardiac phases to extract motion independently. Inverse deformation vector fields were used to accumulate motion-adjusted dose distributions. Dose differences were estimated using the conformity index: (V PTV25Gy) 2/(V PTV · V 25Gy), for planning target volume, and maximum and mean dose differences for CS.

RESULTS: The median 95th percentile planning target volume motion was comparable between patients, with 6.0 mm (FB), 5.6 mm (AC), and 4.9 mm (BH). The left ventricle had the highest median 95th percentile motion out of all CS, with 6.8 mm (FB), 5.5 mm (AC), and 5.9 mm (BH). Estimated motion values were comparable between deformable image registration and auto-contouring (FB, 2.9-7.6 vs 3.1-8.5 mm; AC, 2.4-15.1 vs 3.0-9.0 mm; BH, 1.0-7.3 vs 2.4-6.7 mm). Median conformity index was reduced in the accumulated dose compared to the planned dose by: FB:ITV (-7.2%), AC (-14.1%), BH (-4.3%), and FB:TR (-1.3%). Mean CS dose increased by up to 2 Gy, and maximum dose increased by up to 1.2 Gy (excluding outliers). The largest dose differences were found for small CS.

CONCLUSIONS: Motion and its dosimetric effect have similar magnitudes between treatment setups. Motion findings were consistent between deformable image registration and auto-contouring. Only tracking significantly limited the dosimetric impact of motion.