Evaluation of MVCT images with skin collimation for electron beam treatment planning

Abstract

This study assessed the potential of using megavoltage CT (MVCT) images taken with high density skin collimation in place for electron beam treatment planning. MVCT images were taken using the TomoTherapy Hi·Art^® system, and the CT numbers were converted to density by calibrating the Hi-Art system using an electron density phantom. Doses were computed using MVCT images and kVCT images and compared by calculating dose differences in the uniform dose region (>90%, excluding buildup region) and calculating distance-to-agreement (DTA) in high dose-gradient regions (penumbra and distal falloff, 90%-10%). For 9 and 16 MeV electron beams of 10x10 cm² calculated on a homogeneous CIRS plastic water^® phantom, the maximum dose differences were +2.3% for both beams and the maximum DTAs were 1.9 mm and 1.3 mm, respectively. The same phantom was then MV CT scanned nine times with square skin collimators of Cerrobend^® on its surface - field sizes of 3x3, 6x6, and 10x10 cm² and thicknesses of 6, 8, and 10 mm. Using the Philips Pinnacle³ treatment planning system, a treatment plan was created for each combination of electron energy and field size. The same treatment plans were calculated using the kVCT images of the phantom with regions-of-interest (ROI) manually drawn to duplicate the sizes, shapes, and density of the skin collimators. With few exceptions, the maximum dose differences exceeded ±5% and the DTAs exceeded 2 mm. We determined that the dose differences were due to residual distortions in the MVCT images manifested as errors in the phantom CT numbers and in the shape of the skin collimator edges. These results suggest that MVCT images without skin collimation have potential for use in patient electron beam treatment planning. However, the small residual distortion in images with skin collimation makes them unsuitable for clinical use.