Six-degree-of-freedom respiratory motion modelling significantly reduced 3D localization errors compared to no compensation (0.8 mm vs 2.8 mm) and RV lead compensation (2.0 mm) (p<0.001).
Does 6DoF respiratory motion modelling using three cardiac leads improve target localization accuracy compared to no compensation or single lead compensation in patients undergoing cardiac radioablation?
Six-degree-of-freedom respiratory motion modelling using three implanted cardiac leads significantly improves target localization accuracy for cardiac radioablation.
Absolute Event Rate: 0.8% vs 2.8%
p-value: p=<0.001
Purpose:To investigate the time-resolved translations and rotations of the heart through respiration and their impact on target localization accuracy in cardiac radioablation (CR). Methods and Materials:Twelve patient datasets, from six patients, were acquired with 5 Hz, bi-planar kV, fluoroscopy for 15-20 seconds in preparation for CR.Each patient was imaged twice, with and without abdominal compression.Included CR patients had implanted cardiac leads in the right ventricle (RV), right atrium (RA), and left ventricle (LV).Time-resolved respiratory motion for each cardiac lead was determined by monitoring the lead tip in bi-planar images, triangulating its 3D position, and low-pass filtering its motion.Three motion compensation strategies to model the target's position were simulated:(i) no respiratory motion compensation, (ii) RV lead respiratory compensation, and (iii) six-degree-offreedom (6DoF) respiratory motion modelling using all three cardiac leads.The 6DoF model also enabled quantification of the time-resolved translation and rotations of the cardiac lead cluster through respiration.Each scenario was evaluated on its ability to predict the position of an independent pseudo-target, represented by the most proximal LV lead electrode, on the lateral wall of the LV. Results:The average rotational amplitude of the cardiac lead cluster through respiration was 2.40.6 right-left, 1.30.4sup-inf, 1.70.7 ant-post degrees.For each patient, 6DoF respiratory motion compensation signficantly (p0.001)reduced respiratory motion localization errors compared to no motion compensation and RV lead only compensation.The average magnitude of 3D localization errors in respiratory motion compensation was (2.81.1)mm without motion compensation, (2.01.0)mm with RV lead compensation, and (0.80.4) mm with 6DoF motion modelling. Conclusions:The rotation of the heart through respiration in CR, and its importance for real-time motion monitoring, is presented for the first time.For each patient dataset, 6DoF modelling significantly improved the accuracy of respiratory motion localization for pseudo-targets on the lateral wall of the LV.
Marshall et al. (Sun,) conducted a other in Cardiac radioablation (n=6). Six-degree-of-freedom (6DoF) respiratory motion modelling vs. No motion compensation or RV lead only compensation was evaluated on 3D localization errors in respiratory motion compensation (p=<0.001). Six-degree-of-freedom respiratory motion modelling significantly reduced 3D localization errors compared to no compensation (0.8 mm vs 2.8 mm) and RV lead compensation (2.0 mm) (p<0.001).