Author: Bing-Hao Chiang, Eric C. Ford, Juergen Meyer, Timothy D. Solberg 👨🔬
Affiliation: Department of Radiation Oncology, Fred Hutchinson Cancer Center, University of Washington 🌍
Purpose: This study evaluated the potential risk reduction of various pretreatment patient-specific quality assurance (PSQA) programs for Intensity Modulated Radiation Therapy (IMRT) in mitigating failure modes.
Methods: A Failure Modes and Effects Analysis (FMEA) was conducted by 12 medical physicists with IMRT experience across multiple sites within a single radiation oncology department, including a medical center, cancer center, community practice, and proton center. 18 IMRT QA-related failure modes were identified and assigned FMEA scores—Occurrence, Severity, and Detectability—based on TG-100 guidelines. The baseline assumed routine linear accelerator QA, secondary dose calculation, and plan checks only. Detectability scores were assigned based on the PSQA programs used: array-measurement-based, logfile-based, and EPID-based QA. Risk Priority Numbers (RPN) were calculated for each failure mode across the different PSQA programs.
Results: Three high-risk failure modes (RPN > 150) were identified: small field modeling, plan deliverability, and Multi-Leaf Collimator (MLC) modeling. The reduction in RPN was primarily due to improved detectability scores. Array-measurement-based QA reduced the RPN for small field modeling from 321.8 to 210.4 and for MLC modeling from 155.3 to 97.5. All three QA approaches effectively reduced the RPN for plan deliverability from 194.7 to 43.9. However, detectability improvements for the remaining 15 failure modes were minimal, reinforcing the need for routine plan checks and linear accelerator QA.
Conclusion: A practical method for evaluating PSQA programs in IMRT was developed and demonstrated consensus across diverse clinical settings. The results suggest that array-measurement-based QA may reduce risks for SBRT or highly modulated plans, particularly for MLC and small field modeling, while logfile-based or EPID-based QA are equally effective for plan deliverability.