ESTRO 2024 - Abstract Book

S4949

Physics - Quality assurance and auditing

ESTRO 2024

Data from 2688 irradiations of the IROC IMRT phantom, carried out by hundreds of institutions between 2004 and 2019 as part of clinical trial credentialing, were reviewed. Each irradiation was assessed by measurements performed with 6 TLDs centrally located in the two targets in the phantom. Delivered dose versus calculated dose was evaluated at each location as part of normal IROC credentialing activities. To estimate biological outcome, these 6 values were used to generate CTV differential DVHs (dDVH) and then to calculate Tumour Control Probability (TCP) using the Marsden model [2] (BioSuite Vn 12.01). H&N TCP model parameters were derived from the literature to be α=0.3 (1/Gy) σ α =0.09 and α/β=10 Gy, adjusting σ α to fit clinical data [3]. The “ideal” TCP was obtained by assigning 100% of the prescribed dose to the CTV. Correlation between TCP variations obtained as DeltaTCP (ideal_TCP-calculated TCP), and physical dose variations given as DeltaDose (Dose prescribed to the TLD– maximum deviation measured) were calculated. Irradiation results were reported as failed if the TLD had either <93% or >107% disagreement between prescribed and measured dose. TCP-based criteria were evaluated considering 3%, 5% and 10% tolerances, as well as defining the tolerance that yielded the same absolute number of failing phantom results. Considering the currently used criteria for audits of 7% at the point of measurements, 90.1% of the phantom measurements (2421 out of 2688) passed the dosimetric criteria. The same 90.1% pass rate was achieved with a TCP tolerance of 7.4%, which is relatively loose compared to desired tumour control outcomes. When the dosimetric tolerance results were compared to the biological results, there was generally good agreement (Table 1). Most plans that passed/failed the dosimetric criteria also passed/failed the biologic criteria, although this was not universally true and in 3% (passDeltaTCP-fail DeltaDose) and 3.1% (failDeltaTCP-passDeltaDose) of cases there was disagreement about whether the plan should pass/fail. The degree of disagreement between dosimetric and biologic evaluation increased as the TCP tolerances used were tightened to 5%. In this case agreement (passDeltaTCP-passDeltaDose) was found in only 74.5% of the phantoms irradiations, and in 15.5% (fail-pass) and 1.3% (pass-fail) of the cases there was disagreement between the dosimetric and biological criteria, largely identifying cases where the dosimetric metrics were deemed acceptable but the biological impact (delta TCP) was high. The relation between dosimetric errors and biological impact for each phantom is shown in Figure 1, including the key metrics considered for the phantom. Results: