Evidence-based design of a patient-specific quality assurance program through retrospective analysis of pretreatment and in vivo results

Background and purpose As radiotherapy techniques become increasingly complex, patient-specific quality assurance (PSQA) must evolve to detect a broad spectrum of potential failure modes across the treatment chain. This study presents a extensive evaluation of a multi-modal PSQA program by assessing its ability to detect clinically relevant failure modes and quantifying its efficacy. Materials and Methods A five-year retrospective analysis was conducted of an integrated PSQA program implemented at Maastro, combining secondary dose calculations, log-file based verification, and Electronic Portal Dosimetry (EPD) in both pretreatment and in vivo workflows. The efficacy of individual QA methods and the overall program was quantified using positive predictive values (PPVs). PSQA outcomes were linked to potential improvements in treatment planning strategies and delivery techniques. Results A total of 18,558 treatment plans were analyzed in the pretreatment workflow and 1817 treatment plans in the in vivo workflow. Nearly all high-priority failure modes could be detected except machine output miscalibration. Combining PSQA methods substantially improved PPVs across failure categories. The in vivo workflow flagged the most failure modes and additionally gives insight into possible improvements in planning strategies or treatment techniques. Conclusions This retrospective analysis demonstrated that a comprehensive framework incorporating independent, purpose-specific PSQA methods is essential to detect a wide range of treatment deviations. Combining complementary methods improves the reliability and interpretability of results. The clinical impact of a PSQA program can be further enhanced by adopting more clinically meaningful evaluation metrics and using the results to guide the optimization of new treatment strategies.