(Sunday, 7/14/2019) 2:00 PM - 3:00 PM
Room: 304ABC
Purpose: Conventional pre-treatment thoracic 4DCBCT scans take 1320 projections over 4 minutes. This study seeks to evaluate a “quick scan� (200 exposure, sub 2-minute) protocol. To account for the reduced data, advanced reconstruction algorithms are implemented and evaluated. The “quick scan� is also Respiratory-Motion-Guided (RMG) i.e. patient respiration guides gantry rotation to ensure equal angular separation between projections of a given respiratory phase.
Methods: This study used CT scans from 14 patients in the Virginia Commonwealth University 4D-Lung database. For each patient, a planning CT was used for motion estimation and parameter tuning, and a later CT used as a “ground-truth� for simulating conventional and quick scan data. Reconstructions were then computed with the FDK, McKinnon-Bates (MKB), RecOnstructiOn using Spatial and TEmporal Regularization (ROOSTER) and Motion Compensated FDK (MCFDK) algorithms. A novel reconstruction algorithm where MCFDK motion is estimated from MKB reconstructions, dubbed MCMKB, is also implemented. Reconstructions were then evaluated with Root Mean Square Error (RMSE), Signal-to-Noise Ratio (SNR), Structural SIMilarity index (SSIM) and Tissue Interface Sharpness (TIS).
Results: The quick scan delivers 80% less dose and is on average 75% faster. Relative to conventional 4DCBCT FDK, Quick scan FDK reconstructions have on average 14% higher RMSE, 20% lower SNR, 37% lower SSIM and 37% lower TIS. The MCFDK algorithm has comparable computational cost to FDK, and quick scan MCFDK reconstructions had on average 46% lower RMSE, 390% higher SNR, 190% higher SSIM and 23% higher TIS than conventional scan 4D FDK reconstructions.
Conclusion: Scan time and dose can be significantly reduced without compromising on reconstruction quality by implementing RMG acquisition and motion compensated reconstruction.
Funding Support, Disclosures, and Conflict of Interest: This research was supported by NHMRC project grants #1138899 and #1123068, awarded through the Priority-driven Collaborative Cancer Research Scheme and funded by Cancer Australia. Ricky O'Brien acknowledges the support of a Cancer Institute NSW Career Development fellowship. Paul Keall acknowledges the support of an NHMRC Senior Principal Research Fellowship.
Keywords
Cone-beam CT, Respiration, Reconstruction
Taxonomy
Contact Email
