Author: Denisa R. Goia, M. Saiful Huq, Ronald John Lalonde, Fang Li, Noor Mail, Joseph Shields, Christopher Tyerech 👨🔬
Affiliation: UPMC Hillman Cancer Center, UPMC Hillman Cancer Center and University of Pittsburgh School of Medicine, Department of Radiation Oncology, University of Pennsylvania, UPMC 🌍
Purpose: HyperSight is a new platform for image-guided radiation therapy, offering advanced reconstruction algorithms, a large field-of-view, and rapid acquisition times. To validate the performance of its reconstruction algorithms under different scatter conditions, a quantitative comparison of image-quality is needed. In this study, four different reconstruction algorithms are used to describe and compare the image-quality offered by HyperSight under standardized metrics for two scatter scenarios.
Methods: A CatPhan phantom, wrapped in bolus to simulate patient anatomy (26 cm thickness in the anterior-posterior and 31.5 cm in the lateral direction), was used for CBCT acquisition. Scans were performed with a half-arc (20°–211°) at 124-kVp and 470.35 mAs for 15 and 25 cm fields of view (FOV) in the superior-inferior-direction. The CBCT data were reconstructed using Standard-FDK (Feldkamp-Davis-Kress-back-projection), iterative-CBCT (ICBCT), ICBCT-Acorus, and metal-artifact reduction (ICBCT-MAR). Image-quality was assessed using standard-metrics, including contrast-to-noise ratio (CNR), uniformity, shading-artifacts, modulation-transfer-function (MTF), CT-number accuracy, and linearity. Image analysis was performed using MATLAB 2021.
Results: HyperSight Acuros images demonstrated superior image-quality compared to the other three algorithms. CNR values for Acuros were 5.95±0.54 and 5.90±0.40 at 15 and 25 cm FOV, respectively, closely matching CT-sim values. In contrast, CNR for Standard-FDK images was 4.50±0.35 and 3.13±0.38 at 15 and 25 cm FOV, respectively. Shading-artifacts were 1.50±0.60 and -3.85±3.51 for Acuros at 15 and 25 cm FOV compared to CT -0.88±1.41. Shading-artifacts was 14.2±1.0 for FDK standard images at 24 cm FOV. Uniformity for HyperSight CBCT images were comparable to CT-simulation. The MTF with ICBBCT-Acuros was marginally superior.
Conclusion: This study quantitatively evaluated HyperSight CBCT image-quality across four reconstruction algorithms. The Acorus algorithm delivered superior performance, comparable to CT simulation, under both low and high scatter conditions. The findings support the use of HyperSight Acorus for daily patient alignment and adaptive therapy applications.