ESTRO 2024 - Abstract Book

S4060

Physics - Inter-fraction motion management and offline adaptive radiotherapy

ESTRO 2024

Results:

The maximum deviation between HyperSight CBCT and CT was 35 HU with the phantom was positioned centrally. Therefore, the currently clinically implemented CT-to-ED calibration curve was considered sufficiently accurate for HyperSight CBCT, and was used for all dose calculations. For prostate cancer patients the median time between planning CT and first fraction was 10 days (range 8-24 days). All DVH parameters were accurate within 0.9% between the dose calculation on the HyperSight CBCT and the planning CT. PTV Dmean was accurate within 0.5 Gy, anus Dmean within 0.6 Gy and femoral heads Dmax within 1.0 Gy. On average, 0.78% of voxels failed in the gamma analysis using 2%/2mm reference criteria (range 0.2-1.6%) and 3.9% for 1%/1mm (range 2.6%-6.5%). One patient showed a much higher failed voxels percentage (4.8% for 2%/2mm and 12.4% for 1%/1mm) due to a mismatch in the patient contour as a result of differences in positioning between the planning CT and the first treatment fraction. For lung cancer patients the median time between planning CT and first fraction was 14 days (range 11-33 days). Notably, the median PTV V 95% was 1.7% lower, while the median PTV V 107% was 1.5% higher for the HyperSight CBCT (Figure 1). In one patient, the PTV V 95% dropped by 20.3%. Upon closer inspection, this was due to a motion artifact on the CBCT, leaving patches of HU≈-1000 in the lungs (Figure 2). Similar artifacts were seen in the CIRS dynamic thorax phantom, especially in case of large tumor motion and/or low breathing frequency. On average, 3.6% of voxels failed in the gamma analysis for 2%/2mm (range 0.7-5.7%) and 14.8% for 1%/1mm (range 3.6-25.2%).