ESTRO 2024 - Abstract Book

S4824

Physics - Quality assurance and auditing

ESTRO 2024

and reference geometry. While calibration errors were uniform (by nature) across all plans and phantom geometries, MLC offset and MLC transmission (combined with calibration and leaf bank modification) errors showed a clearly greater impact in more complex plans (particularly the overmodulated cases). In general, the overmodulated plans were on average 20% more sensitive to the introduced errors (Figure 2), and in some instances demonstrated twice the dosimetric error. The largest dosimetric errors observed were up to 12.1% for CTV and 30.1% for OAR. In general, errors in the OAR were on average 50% larger in magnitude than errors in the CTV (Figure 2), and in some instance those differences were greater than twice as much.

Conclusion:

This study developed a cohort of reference plans that meet reasonable and clinically relevant criteria. Not all plans were found to be suitable reference plans. Introduction of known errors in beam modelling created a variety of impact across the different plans. These reference plans provide a robust methodology for reference plan validation and rigorously defining and comparing the sensitivity and specificity of different audit methodologies. The potential for aligning international dosimetry audits through the GHG is the overall goal of our ongoing work.