Comprehensive radiation and imaging isocenter verification using NIPAM kV-CBCT dosimetry.

ABSTRACT

PURPOSE: To develop and demonstrate a comprehensive method to directly measure radiation isocenter uncertainty and coincidence with the cone-beam computed tomography (kV-CBCT) imaging coordinate system that can be carried out within a typical quality assurance (QA) time slot. METHODS: An N-isopropylacrylamide (NIPAM) three-dimensional (3D) dosimeter for which dose is observed as increased electron density in kV-CBCT is irradiated at eight couch/gantry combinations which enter the dosimeter at unique orientations. One to three CBCTs are immediately acquired, radiation profile is detected per beam, and displacement from imaging isocenter is quantified. We performed this test using a 5 mm diameter MLC field, and 7.5 and 4 mm diameter cones, delivering approximately 16 Gy per beam. CBCT settings were 1035-4050 mAs, 80-125 kVs, smooth filter, 1 mm slice thickness. The two-dimensional (2D) displacement of each beam from the imaging isocenter was measured within the planning system, and Matlab code developed in house was used to quantify relevant parameters based on the actual beam geometry. Detectability of the dose profile in the CBCT was quantified as the contrast-to-noise ratio (CNR) of the irradiated high-dose regions relative to the surrounding background signal. Our results were compared to results determined by the traditional Winston-Lutz test, film-based "star shots," and the vendor provided machine performance check (MPC). The ability to detect alignment errors was demonstrated by repeating the test after applying a 0.5 mm shift to the MLCs in the direction of leaf travel. In addition to radiation isocenter and coincidence with CBCT origin, the analysis also calculated the actual gantry and couch angles per beam. RESULTS: Setup, MV irradiation, and CBCT readout were carried out within 38 min. After subtracting the background signal from the pre-CBCT, the CNR of the dosimeter signal from the irradiation with the MLCs (125 kVp, 1035 mAs, n = 3), 7.5 mm cone (125 kVp, 1035 mAs, n = 3), and 4 mm cone (80 kVp, 4050 mAs, n = 1) was 5.4, 5.9, and 2.9, respectively. The minimum radius that encompassed all beams calculated using the automated analysis was 0.38, 0.48, and 0.44 mm for the MLCs, 7.5 mm cone, and 4 mm cone, respectively. When determined manually, these values were slightly decreased at 0.28, 0.41, and 0.40 mm. For comparison, traditional Winston-Lutz test with MLCs and MPC measured the 3D isocenter radius to be 0.24 mm. Lastly, when a 0.5 mm shift to the MLCs was applied, the smallest radius that intersected all beams increased from 0.38 to 0.90 mm. The mean difference from expected value for gantry angle was 0.19 ± 0.29°, 0.17 ± 0.23°, and 0.12 ± 0.14° for the MLCs, 7.5 mm cone, and 4 mm cone, respectively. The mean difference from expected for couch angle was -0.07 ± 0.28°, -0.08 ± 0.66°, and 0.04 ± 0.25°. CONCLUSIONS: This work demonstrated the feasibility of a comprehensive isocenter verification using a NIPAM dosimeter with sub-mm accuracy which incorporates evaluation of coincidence with imaging coordinate system, and may be applicable to all SRS cones as well as MLCs.