Room: AAPM ePoster Library
To quantify the Hounsfield Unit (HU) variation of CBCT caused by different image acquisition setups for application of direct dose calculation in adaptive radiotherapy.
The measurements were performed on Halcyon™ 3.0 linear accelerator (Varian Medical Systems, Palo Alto, USA) using the default pelvis protocol (125 kV) and CIRS Model 062MA CBCT Electron Density phantom. To assess the variation of the HU values, different tests have been set up considering repeatability, consistency, scan length and object size. Consistency is defined as the HU difference between two plugs of the same tissue equivalent material positioned in two different locations of the phantom. The scan length was changed from 28 to 8 cm by changing kV collimator position longitudinally. Scans were reconstructed using 2 mm slice thickness once with the commercially available iterative reconstruction algorithm (iCBCT Halcyon 3.0) and once with an improved prototype version. The improved version uses Acuros® CTS scatter correction and a spectrum-based HU assignment. The measured HU values were calculated in a cylindrical ROI for every plug and were compared for both algorithms.
The variation of HU values between 10 consecutive scans was maximum 4 HU. The highest inconsistency of 41 HU for soft tissue materials decreased to 10 HU. Nevertheless, HU inconsistency of bone-200 and lung-inhale increased by 10 HU. The HU variation, caused by a 20 cm reduction in scan length, decreased by 73% in average. The overall HU variation caused by a 12 cm increase in phantom total lateral extent decreased in average by 34%.
Using Acuros® CTS and a spectrum-based HU assignment in iCBCT, improves overall HU consistency and reduces overall HU variation caused by different scan lengths and object sizes. It is expected that this helps to improve the accuracy of the direct dose calculation using CBCT in adaptive radiotherapy.
Funding Support, Disclosures, and Conflict of Interest: I am employed by Varian Medical Systems.