Purpose: The Halcyon linear accelerator (Varian Medical Systems, Palo Alto, CA) has an integrated x-ray imaging system capable of performing cone-beam CT (CBCT) acquisitions four times faster than conventional systems, including a free-breathing fast acquisition protocol for 3DCBCT in which 491 projections are acquired in 16.6 seconds. A conventional 4DCBCT acquires 1320 projections over 240 seconds to provide enough information for conventional reconstructions to produce images of sufficient quality across the respiratory cycle. This study presents a method by which the fast <20 second acquisition can produce 4DCBCT images and quantify image quality relative to conventional 4DCBCT.
Methods: We used two 4DCT scans of the same patient from the Cancer Imaging Archive 4D-Lung dataset, defining the first as a “planning” 4DCT and the second as a “ground-truth” for 4DCBCT simulation and image quality quantification. The open-source Reconstruction Tool Kit (RTK) was used to simulate conventional and rapid CBCT acquisitions over a 200-degree arc by generating 1320 and 491 forward projections respectively, assuming 13-15 breaths per minute. Both acquisitions were reconstructed using the 4D Feldkamp-Davis-Kress (4DFDK) algorithm. The rapid acquisition was also reconstructed using the Motion Compensated FDK (MCFDK) algorithm, in which data is backprojected along curved paths to account for motion estimated from the planning 4DCT. Image quality was quantified in terms of contrast-to-noise ratio (CNR), root-mean-square error to ground-truth (RMSE-GT) and structural similarity index to ground-truth (SSIM-GT).
Results: The phase averaged CNR, RMSE-GT and SSIM-GT for the conventional 4DFDK were 2.695, 78.93 and 0.415 respectively. The phase averaged CNR, RMSE-GT and SSIM-GT for the rapid 4DFDK and MCFDK reconstructions were 2.006, 292.8, 0.072 and 2.864, 134.7, 0.205.
Conclusion: The data acquired in the <20 second protocols offered on new generation linacs can deliver 4DCBCT images of quality comparable to 4-minute protocols but only when motion compensated reconstruction is implemented.
Funding Support, Disclosures, and Conflict of Interest: This work was funded by an NHMRC project grant 1138899 and partially by the Cancer Australia PdCCRS project grant number 1123068. RO was funded by a Cancer Institute NSW career development fellowship.