RESUMEN
Electronic portal imaging (EPI) is commonly used to identify and correct for inter-fraction variability in tangential breast irradiation. Based on the institutional policy, EPI registration is performed by either radiation oncologist or therapist. Little data is available on the inter-observer agreement in EPI registration among different health practitioners. The aim of our study was to analyze inter-observer agreement among radiation oncologists and therapists in the evaluation of EPI for breast cancer radiotherapy verification. EPI data of 40 patients treated with tangential fields were independently reviewed by a radiation oncologist (on-line, just before treatment) and off-line by junior and senior therapists. Displacement of each EPI image with respect to the digital reconstructed radiographs (DRRs) was quantified using manual EPI registration based on bony marks with the corresponding DRRs. Agreement between observers was evaluated using weighted Cohen's Kappa statistics. In 95% out of 720 EPI-DRR comparisons, the EPI-DRR misalignment was < 5 mm. The difference between observers was < 2 mm in 666 (92.5%) out of all 720 delta values. High inter-observer agreement was found, with weighted Cohen's Kappa values attesting evaluation overlaps ranging from moderate (among therapists) to almost perfect (among radiation oncologist and therapists). The high agreement among the observers demonstrated the precision of breast localization using EPI. These findings suggest that routine EPI-based patient set-up verification in breast cancer radiotherapy can be safely entrusted to trained therapists (supervision should be assured based on the local tasks definition). Our study might be useful in quality assurance and in the optimization of workload in the radiotherapy departments. They might allow for wider implementation of complex and evolving radiotherapy technologies.