Using kV-kV and CBCT imaging to evaluate rectal cancer patient position when treated prone on a newly available belly board.

The goal of this work was to use daily kV-kV imaging and weekly cone-beam CT (CBCT) to evaluate rectal cancer patient position when treated on a new couch top belly board (BB). Quality assurance (QA) of the imaging system was conducted weekly to ensure proper performance. The positional uncertainty of the combined kV-kV image match and subsequent couch move was found to be no more than ± 1.0 mm. The average (1 SD) CBCT QA phantom match was anterior-posterior (AP) = -0.8 ± 0.2 mm, superior-inferior (SI) = 0.9 ± 0.2 mm, and left-right (LR) = -0.1 ± 0.1 mm. For treatment, a set of orthogonal kV-kV images were taken and a bony anatomy match performed online. Moves were made along each axis (AP, SI, and LR) and recorded for analysis. CBCT data were acquired once every 5 fractions for a total of 5 images per patient. The images were all taken after the couch move but before treatment. A 3-dimensional (3D-3D) bony anatomy auto-match was performed offline and the residual difference in position recorded for analysis. The average (± 1 SD) move required from skin marks, calculated over all 375 fractions (15 patients × 25 fractions/patient), were AP = -2.6 ± 3.7 mm, SI = -0.3 ± 4.9 mm, and LR = 1.8 ± 4.5 mm. The average residual difference in patient position calculated from the weekly CBCT data (75 total) were AP = -1.7 ± 0.4 mm, SI = 1.1 ± 0.6 mm, and LR = -0.5 ± 0.2 mm. These results show that the BB does provide simple patient positioning that is accurate to within ± 2.0 mm when using online orthogonal kV-kV image matching of the pelvic bony anatomy.