Predictive risk model for radiation-induced pneumonitis in patients diagnosed with advanced esophageal cancer receiving radiation therapy alone or in combination with immunotherapy: A retrospective study

Abstract

Background Radiation therapy (RT) is an essential treatment for esophageal cancer with radiation induced pneumonitis (RP) as a possible adverse event. With the emergence of immunotherapy for patients with esophageal cancer (EC), whether in combination with RT or alone it increases RP incidence, which leads to the need of reevaluating the RP risk predication paraments. Methods Clinical and physical parameters were collected from patients with clinicopathologically confirmed unresectable locally advanced EC from 1/2020 and 7/2023. The endpoint was Grade ≥ 2 RP occurrence within 6 months after radiotherapy. The χ² test and Logistic regression analysis were used to analyze the relationship of categorical and continuous variables with RP occurrence respectively. Multivariate Cox analysis was used to construct a RP risk model by R software, the accuracy of which was further evaluated by ROC and risk curves. Results After strict screening, 92 patients receiving RT alone and 84 patients receiving RT + Immunotherapy were eligible for inclusion in this study. The incidence of grade ≥ 2 RP in patients with EC who received RT and immunotherapy was 30.95%, which is higher compared to those who received RT alone (17.39%). Several factors were included for the construction of RP Risk Model by multivariate Cox regression in group RT and RT-I seperatly. Four factors were used for RP prediction risk model in patients who received the RT alone, and five factors were used for RP prediction risk model in patients who received the RT + Immunotherapy. The ROC curve indicated the satisfactory accuracy with AUC value was 0.734 and 0.805 in group RT and RT-I respectively. Risk curves confirmed favorable accuracy that the higher risk score, the higher RP risk, the lower interval times when RP happens after RT. Conclusions (1) The immunotherapy may increase the risk of RP. (2) We screened out five indicators (age, total irradiation dose, irradiation segmentation frequency, V15 and V20) for predicting RP incidence in EC patients receiving RT and immunotherapy combination, which provided an important theoretical basis for the RT treatment plan.