ACR benchmark testing of a novel high-speed ring-gantry linac kV-CBCT system

Abstract

A new generation cone-beam computed tomography (CBCT) system with new hardware design and advanced image reconstruction algorithms is available for radiation treatment simulation or adaptive radiotherapy (HyperSight CBCT imaging solution, Varian Medical Systems-a Siemens Healthineers company). This study assesses the CBCT image quality metrics using the criteria routinely used for diagnostic CT scanner accreditation as a first step towards the future use of HyperSight CBCT images for treatment planning and target/organ delineations. Image performance was evaluated using American College of Radiology (ACR) Program accreditation phantom tests for diagnostic computed tomography systems (CTs) and compared HyperSight images with a standard treatment planning diagnostic CT scanner (Siemens SOMATOM Edge) and with existing CBCT systems (Varian TrueBeam version 2.7 and Varian Halcyon version 2.0). Image quality performance for all Varian HyperSight CBCT vendor-provided imaging protocols were assessed using ACR head and body ring CT phantoms, then compared to existing imaging modalities. Image quality analysis metrics included contrast-to-noise (CNR), spatial resolution, Hounsfield number (HU) accuracy, image scaling, and uniformity. All image quality assessments were made following the recommendations and passing criteria provided by the ACR. The Varian HyperSight CBCT imaging system demonstrated excellent image quality, with the majority of vendor-provided imaging protocols capable of passing all ACR CT accreditation standards. Nearly all (8/11) vendor-provided protocols passed ACR criteria using the ACR head phantom, with the Abdomen Large, Pelvis Large, and H&N vendor-provided protocols produced HU uniformity values slightly exceeding passing criteria but remained within the allowable minor deviation levels (5-7 HU maximum differences). Compared to other existing CT and CBCT imaging modalities, both HyperSight Head and Pelvis imaging protocols matched the performance of the SOMATOM CT scanner, and both the HyperSight and SOMATOM CT substantially surpassed the performance of the Halcyon 2.0 and TrueBeam version 2.7 systems. Varian HyperSight CBCT imaging system could pass almost all tests for all vendor-provided protocols using ACR accreditation criteria, with image quality similar to those produced by diagnostic CT scanners and significantly better than existing linac-based CBCT imaging systems.

Keywords: ACR; CBCT; treatment planning.

FIGURE 1

ACR CT head and body ring design and setup. (a) ACR head phantom module design and size, (b) head phantom setup, (c) body ring and head phantom setup.

FIGURE 2

ACR head phantom module analysis. (a) CNR module analysis ROIs, (b) spatial resolution bar patterns, (c) HU accuracy ROI placements, (d) uniformity ROIs and image scaling measurement location.

FIGURE 3

ACR phantom passing criteria for CNR, spatial resolution, HU accuracy, uniformity, and image scaling based on imaging protocol selected.

FIGURE 4

ACR head phantom HU accuracy module slice taken with window leveling set to pelvis (−160 to 240 HU) on the (a) Varian HyperSight, (b) Siemens SOMATOM Edge, (c) Varian TrueBeam 2.7, and (d) Varian Halcyon 2.0. All modalities using default Head imaging protocol.

FIGURE 5

CT number accuracy for four imaging protocols for four HyperSight imaging protocols with a repeated set of scans several months later (green = ACR passing criteria HU range for each material, blue stars = original scan values, red pluses = repeated scan values). Each repeated imaging protocol used the same default imaging parameters with a slice thickness of 3 mm, with the Head protocol using the ACR head phantom and the Pelvis, Pelvis Large, and Breast protocols using the ACR body phantom.