Evaluating Image Quality of Cone Beam Computed Tomography Using HyperSight® on the Varian TrueBeam

The study evaluated the image quality of the cone beam computed tomography (CBCT) system on Varian TrueBeam linear accelerators with HyperSight (TB-HS) and compared it with the standard TrueBeam system (TB-S). Particular emphasis was placed on changes relevant to offline adaptive radiotherapy (ART).

Image quality was assessed using a Catphan 504 phantom on two TB-HS and two TB-S machines. Quantitative parameters included HU accuracy, uniformity, contrast-to-noise ratio (CNR), modulation transfer function (MTF), low-contrast visibility (LCV), and noise power spectrum (NPS). Measurements were performed at 100 kV and 125 kV using default clinical CBCT protocols and reconstruction algorithms. Planning CT images acquired with matching clinical protocols served as reference values. Statistical comparisons of image quality parameters between techniques and agreement with internationally recommended tolerance criteria were performed using the pairwise linear mixed models and McNemar's tests, respectively. Additionally, eight radiographers conducted a blinded subjective image quality evaluation of CBCT images from 15 patients before and after the HyperSight installation RESULTS: Catphan analysis showed that TB-HS improved HU accuracy for high-density inserts, while uniformity and CNR showed minor, mostly non-significant differences. Spatial resolution, noise characteristics, and LCV remained unchanged. Both techniques generally met the tolerance criteria, though TB-HS did so more consistently. Subjective image quality evaluation showed no overall preference.

The HyperSight upgrade provided minor image quality enhancements but demonstrated no clear clinical advantage. Nevertheless, findings support the potential of HyperSight for offline ART. However, further investigations in patient-based applications are needed to confirm its clinical value.