Sensitivity Analysis of Acquisition Precision in Fast Low-Dose 4DCBCT
B Lau1*, O Dillon2, S Vinod3, R O'Brien2, T Reynolds2, (1) University Of Sydney, Sydney, NSW, AU, (2) University Of New South Wales, Sydney, NSW, AU
Presentations
TH-E-201-6 (Thursday, 7/14/2022) 1:00 PM - 2:00 PM [Eastern Time (GMT-4)]
Room 201
Purpose: Prior to thoracic radiotherapy treatments, patients are commonly positioned using conventional 4DCBCT scans (1320 projections over 4 minutes). Adaptive 4DCBCT is designed to address the limitations of conventional 4DCBCT (long scan time and high imaging dose) by acquiring evenly spaced fast low-dose scans. Adaptive 4DCBCT modulates projection acquisition and gantry rotational velocity in accordance with real-time patient breathing rate. This enables evenly spaced fast low-dose 200-projections to be acquired with 20 projections binned per respiratory phase (ideal scenario). We assess the impact of projection spacing in adaptive 4DCBCT scans, utilizing data from all 30 patients of the ADAPT clinical trial (NCT04070586).
Methods: To isolate the impact of projection spacing, patient angles from the ADAPT trial were used to simulate projections from the XCAT phantom. In addition, angular separation gaps of 20°, 30° and 40° were added into the ideal scenario projections. The Feldkamp-Davis-Kress (FDK) and Motion Compensated McKinnon Bates (MCMKB) reconstruction methods were used to reconstruct images. Reconstructed image quality was quantified using Structural-Similarity-Index (SSIM), Signal-to-Noise-Ratio (SNR), Contrast-to-Noise-Ratio (CNR), Tissue-Interface-Width-Diaphragm (TIW-D) and Tissue-Interface-Width-Tumor (TIW-T). Evenly spaced 200-projection reconstructions (Ideal) and XCAT ground truth were used as the reference.
Results: Adaptive 4DCBCT patient projection angles had an average angular separation of 9.95°±0.07° (Ideal 10°). For MCMKB-reconstructions, average patient angular spacing produced SSIM 0.98±0, SNR 28.8±6.7, CNR 12.4±2.0, TIW-D 1.2±0.1 and TIW-T 2.4±0.3, angular gap 40° produced SSIM 0.98±0, SNR 34.7, CNR 11.9, TIW-D 1.8 and TIW-T 2.0 and ideal scenario produced SSIM 0.97±0, SNR 34.8, CNR 13.0, TIW-D 1.5 and TIW-T 2.0. Motion compensated reconstruction (MCMKB) produced sharp images with low noise and streaks as opposed to standard FDK-reconstruction.
Conclusion: World’s first implementation of adaptive 4DCBCT achieved even angular spacing within sufficient tolerance. Large angular gaps did not negatively impact image quality provided the full scan range was acquired.
Funding Support, Disclosures, and Conflict of Interest: We would like to acknowledge the patients who participated in the ADAPT trial, offering their time and imaging dose. This study was funded by NHMRC grant 1138899 and partly by a Cancer Australia (Priority-driven Collaborative Cancer Research Scheme) project grant number 1161748.
Keywords
Taxonomy
Contact Email