Research on the correction method for radiotherapy verification plans based on displaced electronic portal imaging device

Background It has been observed that under the single isocenter conditions, the potential shifts of the electronic portal imaging devices (EPID) may be introduced when executing portal dosimetry (PD) plans for bilateral breast cancer, pleural mesothelioma, and lymphoma. These shifts are relative to the calibration positions of EPID and result in significant discrepancies in the plan verification results. Purpose To explore methods including correction model and specific correction matrices to revise the data obtained from displaced EPID. Methods Two methods, the correction model and the specific correction matrices, were applied to correct the data. Five experiments were designed and conducted to build correction model and to validate the effectiveness of these two methods. Gamma passing rates were calculated and data profiles along X‐axis and Y‐axis were captured. Results The gamma passing rates for the EPID‐displaced IMRT validation plans after applying correction model, along with the application of specific correction matrices to VMAT and IMRT validation plans, exhibit results that are comparable to the cases with non‐displaced EPID. Except for the VMAT plans applied correction model which showed larger discrepancies (0.041 ± 0.028, 0.049 ± 0.030), the other three exhibit minimal differences in discrepancy values. In all profiles, the corrected data from displaced EPID exhibit a high level of agreement with data obtained from non‐displaced EPID. Good consistency is observed in actual application of the correction model and the specific correction matrices between gamma passing rates of data corrected and those of non‐displaced data. Conclusions The proposed methods involving correction model and specific correction matrices can correct the data collected from the displaced EPID, and the gamma passing rates of the corrected data show results that are comparable to some extent with those of non‐displaced data. Particularly, the results corrected by specific correction matrices closely resemble the data from non‐displaced EPID.