Doctor of Philosophy (PhD)
Purpose: Volumetric modulated arc therapy-computed tomography (VMAT-CT) was proposed in 2010 as the CBCT reconstruction technique using portal images collected during VMAT. Yet, its clinical applications were impeded by the poor image quality due to motion blurs from fast movements of multi-leaf collimator (MLC) and projection data deficiency due to truncations from MLC. The goal of this study is to develop systematic preprocessing methods, and a novel iterative reconstruction algorithm to generate quality VMAT-CT and lower the number of failure cases of VMAT-CT reconstruction.
Methods: Systematic image preprocessing methods, including Online region-based active contouring (ORACM), MLC motion modeling and outlier filtering were developed to remove MLC motion blurs and residual artifacts on portal images. An iterative algorithm based on Compressed sensing, including two constraints: total variation (TV) and block matching 3D filtering (BM3D), was also proposed to reconstruct VMAT-CT. For phantom studies, 50 phantom cases were collected by delivering clinical VMAT plans for multiple treatment sites (lung, esophagus, and head & neck), to a Rando phantom. For real patient studies, 17 real-patient cases of VMAT in thoracic regions were also acquired. VMAT-CT reconstruction was conducted for all collected cases using the original analytical algorithm and our proposed methods, and image quality was evaluated and compared.
Results: 48 out of 50 phantom cases and 15 out of 17 patient cases were successfully reconstructed using our proposed preprocessing methods and TV-BM3D iterative algorithm, while only 39 phantom cases and 8 patient cases could be reconstructed by the original algorithm. The ANOVA test showed that contrast-to-noise ratios (CNR) of VMAT-CT using our methods were significantly improved (p
Conclusions: The systematic preprocessing methods and TV-BM3D iterative algorithm could improve VMAT-CT quality significantly and boost the success rate of VMAT-CT reconstruction. With these progresses, we believe that VMAT-CT could be a promising and convenient imaging approach for on-site treatment monitoring without introducing additional imaging dose, hardware, and cost.
Chien, Chia-Lung, "Reconstruction of Volumetric Modulated Arc Therapy-Computed Tomography" (2023). LSU Doctoral Dissertations. 6170.
Available for download on Sunday, May 24, 2026