Author: Laura I. Cervino, Karen Episcopia, Hsiang-Chi Kuo, Sangkyu Lee, Seng Boh Gary Lim, Shih-Chi Lin, Grace Tang 👨🔬
Affiliation: Department of Medical Physics, Memorial Sloan Kettering Cancer Center 🌍
Purpose: This study evaluated the performance of the HyperSight Cone-Beam Computed Tomography (CBCT) system on a TrueBeam C-arm LINAC (TB) and two Ethos ring-gantry LINACs (ES) for adaptive radiation therapy.
Methods: HyperSight CBCTs from TB (CBCTTB) and ES (CBCTES) were compared to a standard fan-beam CT simulator (FBCT) using Pelvis and Thorax protocols. A matched field-of-view (350x350 cm2) and similar CT dose indices were used to scan a Catphan-604 phantom. Iterative reconstruction with Boltzmann-based scatter correction, Acuros (BSC). were applied. Metrics included spatial resolution (MTF50%), contrast-to-noise ratio (CNR), noise (coefficient of variation), and non-uniformity. The HU to relative electron density (RED) calibration curves for CBCTTB (HUTB) and CBCTES (HUES) were compared with vendor reference (HUvendor) using an 062M Electron Density Phantom reconstructed with and without BSC. Extended CT curves (HUTB_m and HUES_m) were generated with a titanium plug (RED 3.79) and metal artifact reduction. Phantoms offset by 5 cm superiorly evaluated cone-beam geometry effects on HU.
Results: MTF50% differences were ≤5.7% for Pelvis and up to 21.2% better for Thorax. CNR for CBCTES differed by ≤9.7%, while CBCTTB showed up to 18.7% higher than FBCT. Non-uniformity was marginally inferior (0.3–0.8%), and noise differences were ≤0.2%. Without BSC, HU deviations from HUvendor, in the density range of [Lung, Soft tissue, Bone] were within [52, 41, 428] (HUTB) and [33, 33, 382] (HUES), improving with BSC to be within [20, 21, 147] (HUTB) and [37, 25, 89] (HUES). For the extended curves, HUES_m saturated at 7000 HU with titanium, while HUTB_m was 399 HU over HUvendor (6000 HU). The 5 cm offset caused changes within [9, 7, 79] HU.
Conclusion: HyperSight CBCT demonstrated high image quality comparable to FBCT, with acceptable HU differences from the reference curve. These findings support its feasibility for organ delineation and dose calculation in adaptive radiation therapy.