ESTRO 2024 - Abstract Book

S4823

Physics - Quality assurance and auditing

ESTRO 2024

establish and examine such a cohort of reference plans and then to introduce purposeful errors into the plans such that the ability to identify these by the individual groups could be tested.

Material/Methods:

Five international GHG members participated in developing this IMRT dose assessment framework. Sixteen audit plans were taken as reference plans using RayStation treatment planning system software. Eight plans were designed to be standard modulated plans with the remaining 8 being purposely over modulated versions (~0.5 – 1.0 increased MU/Gy). To validate the 16 reference plans, they were then delivered to a multi-ion (A1SL chamber) chamber phantom consisting of a cube made of PMMA (Figure 1) and using an Elekta Synergy linear accelerator. The point dose differences were evaluated at 8 different locations for each plan with restrictions on dose gradient (<10%) and dose level (>50%). Plans were classified as acceptable if differences between measured and calculated doses were less than 4% at all points in the multi-ion chamber phantom.

Figure. 1 Multi-ion chamber phantom.

Next, the sensitivity of these reference plans to dose errors was examined in the planning system based on the individual audit phantom geometry. Dosimetric errors were introduced into the reference plans, and the impact on dose to CTV and OAR volumes was calculated on each of the 5 audit phantoms. This was done by modifying the TPS parameters (MLC offset, transmission, leaf-tip width, PDD), beam calibration, and MLC position based on documented community distributions of error at the 2.5, 25, 75 and 97.5 percentile level. 3-6 A total of 8 different perturbations were created for each of the 16 cases. Dose volume histograms were assessed to evaluate the differences in the mean dose to CTV and parallel OAR, and the maximum dose to serial OAR structures in the audit phantoms on which the reference plans were originally designed. These differences were compared to the original unperturbed cases.

Results:

While 16 acceptable reference plans were ultimately identified, 4 initial candidates needed to be re-planned for failing to meet the reference plan criteria (agreement within 4% at all appropriate locations). When errors were introduced into these reference plans, based on community errors, different results were identified based on the type of plan