Room: AAPM ePoster Library
Purpose: compare differences in institution irradiation accuracy between different Imaging and Radiation Oncology Core (IROC) phantoms, and across multiple irradiations of the same phantom
Methods: compared the dose deviations (treatment planning system (TPS) vs. measured) from institutions who each irradiated the IROC spine, head & neck (H&N) and lung phantoms at different points in time. One-way ANOVA with repeated measures and pairwise regression analyses were performed on the phantom groups to assess their differences in dose deviation. Additionally, phantom irradiations repeated (not to remedy a failed result) by an institution 3 or more times on any given phantom (n=11), were separately examined. This was to assess whether increased familiarity with the phantom and process improved results.
Results: from 60 institutions who each irradiated all 3 phantoms: spine, H&N and lung were assessed, only 5 of which were failing results. The average dose deviation of the three groups were all within 1% of each other (spine: 2.42%, H&N: 2.23%, lung: 3.02%) and the H&N and lung results were found to be significantly different (p<0.05). No direct correlation was found between any pair of phantom results (R²<0.15), suggesting that performance in one was not a predictor of performance in another. Assessment of repeated phantom irradiations showed no consistent improvement of results over time.
Conclusion: performed similarly on average among the three phantoms irradiated, indicating that irradiation performance is not largely influenced by phantom type, but by the institution’s radiation therapy (RT) process. The absence of direct correlations between phantoms validates their capability to test distinct aspects of the RT process, as performance in one does not appear to predict performance in another. Assessment of multiple irradiations did not show improvements in accuracy over time, which demonstrates that the RT process, not phantom familiarity, is driving performance, as is desired for an audit.