Implementation of High-Quality Motion-Compensated Simultaneous Algebraic Reconstruction Technique (mc-SART) Cone-Beam CT (CBCT) Imaging Using the 5D Model in a Prospective Patient Study
K Singhrao1*, D O'Connell1, J Fu1, M Lauria1, B Stiehl1, A Raldow1, J Hegde1, A Lee1, A Santhanam1, D Low1, J Lewis2, (1) Dept. of Radiation Oncology, University of California, Los Angeles, Los Angeles, CA, (2) Dept. of Radiation Oncology, Cedars-Sinai Medical Center, Los Angeles, CA
Presentations
SU-F-TRACK 6-3 (Sunday, 7/25/2021) 4:30 PM - 5:30 PM [Eastern Time (GMT-4)]
Purpose: Clinically used pre-treatment imaging methods for external-beam radiotherapy of lung cancer such as 3D/4D-based CBCT imaging can result in low quality images. Here, we implement a novel 5D-model-based motion-compensated simultaneous algebraic reconstruction technique (mc-SART) method to a create a patient‐ and treatment-day-specific volumetric image and motion model from free‐breathing cone‐beam projections and respiratory surrogate measurements.
Methods: Four lung cancer patients receiving external beam radiotherapy were prospectively enrolled in this study. Patient-specific breathing traces were recorded using an in-house respiratory monitoring system. Raw CBCT projections were saved after a single fraction for each patient. CBCT projections were sorted into 8 bins based on breathing amplitude and flow. The mc-SART framework involves creating an initial reference image and motion model, and then iteratively improving them using a modified version of SART. The initial reference image was defined using a reconstructed SART image from end exhale/zero flow projections. The motion model was created using deformable image registration between the reconstructed bins and least squares fitting to the 5D model parameters. The initial reference image was deformed to a particular amplitude and flow and a modified version of SART was applied to include motion compensation. The model output is a volumetric reference image and a motion model that can be used to generate volumetric images at any other time point from breathing traces.
Results: Signal-to-noise-ratio (SNR) measurements were performed in soft tissue and lung regions for FDK-reconstructed, 4D-SART and mc-SART reference images. The soft tissue and lung SNR measurements for mc-SART images was 160% and 28% greater than FDK-reconstructed image and, 26% and 43% greater than 4D-SART reconstructed images.
Conclusion: We have developed a method to create a high-quality motion-compensated reference image and motion model with real patient data. This framework opens up the possibility of creating model-based volumetric images during treatment delivery.
Funding Support, Disclosures, and Conflict of Interest: Funding Provided Through a Master Research Agreement with Varian Medical Systems
Handouts
Keywords
Taxonomy
Contact Email