Background and purpose: Recent researches demonstrated that single nucleotide polymorphisms (SNPs) of genes involving inflammation, DNA repair, etc. were associated with risk of radiation pneumonitis (RP). However, these studies were single-centered, from single ethnic origin, without validation from independent cohort studies from other populations. In order to identify clinical valuable SNPs for RP, in this study we selected 19 RP-related SNPs candidates previously published before 2016 for validation in our cohort. Material and methods: 359 lung cancer patients with radiotherapy were included in our prospective study (NCT02490319). Peripheral blood samples from these patients were genotyped by MassArray and Sanger Sequence method. Multivariate Cox hazard and other analyses were applied to estimate the hazard ratio (HR) and 95% confidence intervals (CIs) of all factors possibly related to the risk of RP. Results: Patients with elder age, MLD >= 15 Gy, V-20 >= 24% had higher risk of RP >= grade 3 compared with their counterparts (HR = 2.020, 95% CI: 1.045-3.906, P = 0.037; HR = 2.502, 95% CI: 1.346-4.652, P = 0.004; HR = 2.256, 95% CI: 1.191-4.272, P = 0.013, respectively). Moreover, patients receiving IMRT were associated with decreased incidence of RP (HR = 0.520, 95% CI: 0.280-0.963, P = 0.037). Importantly, CT + TT genotype of IL4: rs2243250 was strongly related to decreased risk of RP >= grade 3 (HR = 0.195, 95% CI: 0.090-0.424, P = 0.000037, P-c = 0.0006). Conclusion: IL4: rs2243250 was validated to be significantly related to RP of grade >= 3 in our cohort. Our results further emphasized the prevalence and clinical value of IL4: rs2243250 on RP, and may thus be one of the important predictors of severe RP before radiotherapy. (C) 2019 Elsevier B.V. All rights reserved.