One essential requirement is reproducible spatial alignment between the kV imaging system and the MV treatment-beam isocentre. It's also important that the CT images are correctly scaled and as sharp as possible. Some QA techniques, however, can be time-consuming or limited to use with just one manufacturer's system. With this in mind, a UK research team has come up with a new QA scheme based on the QUASAR Penta-Guide: an acrylic phantom designed for cone-beam registration and kV-MV alignment (Phys. Med. Biol. 53 5275).

The researchers - from the St James's Institute of Oncology, Leeds Teaching Hospitals Trust and the University of Leeds - are investigating a possible replacement for their existing QA routine, in which they check different aspects of their IGRT machine (an Elekta Synergy system) using different phantoms. Currently, isocentre alignment is examined monthly using a ball bearing phantom and a utility in the Synergy software. Image sharpness, meanwhile is assessed less often using a Catphan phantom.

The new scheme replaces these separate checks with a single cone-beam scan of the Penta-Guide. The team developed a way to assess image sharpness by measuring the modulation transfer function (MTF) of the same set of images used to check the isocentre alignment. This procedure should reduce the time taken to perform quality-control checks on cone-beam CT-based IGRT systems, while maintaining levels of accuracy and precision.

"We already had a good system up and running, using a ball bearing phantom and software that was integral to the Elekta system," explained Jonathan Sykes, principal radiotherapy imaging physicist at St James's Institute for Oncology. "But, we at Leeds have been involved in a project to evaluate IGRT systems, for the UK national health Centre for Evidence-based Purchasing, and I wanted to be able to evaluate both Varian and Elekta cone-beam-CT IGRT systems using the same method."

The Penta-Guide (from Modus Medical Devices of Canada) is a 16 cm cube containing five spherical air cavities. It benefits from being manufacturer-independent but is less accurate than the Synergy's inherent QA scheme. To combat this limitation, the Leeds team developed software based on the Elekta method, but utilizing the Penta-Guide's central air cavity instead of the ball bearing. This method enables the Penta-Guide to test cone-beam CT to MV isocentre coincidence with sufficient accuracy and precision for QA of IGRT.

Sykes and his team also wanted to use the Penta-Guide to assess image sharpness. "The initial motivation was both academic - can it be done? - and dissatisfaction with the Catphan for measuring MTF in cone-beam CT systems," he told medicalphysicsweb. "It turned out to have great potential, with the added benefit that it can be integrated into the QA of the kV-MV alignment."

The researchers measured the MTF on the cone-beam CT images by fitting an edge response function to scan data taken across the phantom-central air cavity interface. MTF could be estimated robustly, even when the images were reconstructed with the 1 mm3 voxel size used in clinical practise.

Next, the team plans to validate the Penta-Guide QA scheme further on the Elekta system, and possibly on both Varian's and Siemens' IGRT kit. The sensitivity of the QA scheme and the stability of the IGRT systems will be tested over a longer time period by seeing whether the Penta-Guide can pick up step changes or drift in geometric alignment and image sharpness.

"I would also like to integrate a basic QA of contrast-to-noise and CT number stability using the air and acrylic densities," said Sykes. "This might show up any drift in tube output, image calibration, increased noise and so on."

Ultimately, the aim is to use one scan of the Penta-Guide to perform all the QA necessary for a monthly or even daily check. A more thorough, yearly QA check using the Catphan would still be employed.

"I am looking to find the resources to turn the software from a research tool into something that any physicist could use," said Sykes. "I am currently in discussion with Modus Medical about how we can take this forward as a product to be supplied with the Penta-Guide phantom."