Trends and Challenges: A 15-Year Review of Imaging and Radiation Oncology Core Anthropomorphic Phantom Audits for Proton Therapy.

Purpose: This study aims to analyze the trends in institutional performance on Imaging and Radiation Oncology Core (IROC) proton phantoms and to investigate the impact of treatment parameters on proton therapy delivery accuracy. Materials and Methods: We analyzed 402 IROC anthropomorphic phantom audits performed by 57 institutions worldwide between 2009 and 2024, including brain, head and neck (H&N), spine, prostate, lung, and liver phantoms. Dosimetric assessments included thermoluminescent dosimeters (TLDs) for dose comparison and gamma analysis of radiochromic film. The lung and liver phantoms were irradiated with simulated motion. Univariate analysis was performed to evaluate the relationships between treatment parameters and phantom outcomes (pass/fail, TLD-to-treatment planning systems (TPS) ratio, and gamma passing rate). Linear regression was used to analyze the long-term trends in the institutional delivery accuracy across the 6 phantoms. Results: The inter-institutional dose variation in the target was approximately 3% across all phantoms. Of the participating institutions, 66.7% failed at least one IROC proton phantom. Pass rates were 97% for brain, 91% for H&N, 89% for spine, 76% for prostate, 63% for lung, and 57% for liver phantoms. Phantoms incorporating motion exhibited the poorest performance, particularly in gamma analysis, with 33 (N = 91) lung and 33 (N = 76) liver irradiations showing gamma values below 85%. No linear temporal trends were observed in the overall passing rates of the six phantoms, while TLD results improved for brain, liver, spine, and prostate phantoms (P < .05). Machine type was correlated with pass rates for H&N and liver phantoms (P < .05). Conclusion: Overall institutional phantom performance has not significantly improved over time. In particular, the proton lung and liver phantoms continuously demonstrate suboptimal results, likely due to complex geometries and motion management challenges. These findings underscore the need for careful selection of treatment parameters and optimization of motion management strategies. The IROC phantom program remains crucial for characterizing proton therapy systems and identifying clinically significant errors.