ESTRO 2024 - Abstract Book

S5472

RTT - Patient care, preparation, immobilisation and IGRT verification protocols

ESTRO 2024

setup data were computed, and statistical significant differences between standard deviations were determined using Levene’s test to compare mask stability. Linear correlation analyses were conducted to assess possible dependence of residual setup errors and the executed pitch and roll following kV-kV guided setup. In addition, the percentages of acquired verification CBCTs per mask system were examined to illustrate the differences in exceeding the protocol tolerances.

Results:

Couch correction shifts based on the set-up kV-kV indicate the inter-fraction variations. Our results (see Table 1) show that the range of AP (-6.52 - 0.53), CC (-1.03 - 0.72), and pitch (-2.7 - 2.4) setup inaccuracies is larger in DSPS masks than for Brainlab masks. For AP (0.82) and pitch (0.87), the SD is significantly higher in DSPS masks. However, for roll (0.66) and yaw (0.76), the SD is significantly higher for Brainlab masks. The residual setup corrections, indicating intra-fraction variability, reveal that for AP (-0.12 - 0.07), roll (-1.1 - 0.6) and yaw(-1.9 - 0.9) DSPS masks exhibit a greater range, while there is a significantly larger SD in CC direction (0.04). For Brainlab masks the SD is significantly higher for roll setup corrections (0.23). Verification CBCT and mid treatment CBCT showed similar results for intra-fraction setup accuracy. The values of the SD for intra-fraction errors are shown to be smaller than 0.04 cm /0.25deg for both mask systems.

Patients in a DSPS mask underwent approximately 20% more verification CBCTs (absolute difference) surpassing the protocol tolerance limits more frequently, primarily due to an overall larger range.

The correlation between residual deviations in CC direction after pitch- and left-right (LR) direction after roll maneuvers appeared to be statistically significant (p =<0.001). The correlation coefficients were nearly twice as high in R2 (0.3 vs 0.13) and slope (0.03 vs 0.015) for DSPS masks as for Brainlab masks (see Figure 1), although for both mask systems the correlation was weak.