Research has confirmed that transforming growth factor-beta1 (TGF-beta1) is one of the cytokines related to radiation pneumonitis. But the level of TGF-beta1 in serum needed to predict radiation pneumonitis is still not clear. This study assessed the value of TGF-beta1 in both serum and induced sputum in predicting radiation pneumonitis, providing a reference for the radiotherapy of patients with non-small cell lung cancer (NSCLC). A total of 23 patients with NSCLC treated with three-dimensional conformal radiotherapy (3D-CRT) or intensity-modulated radiation therapy (IMRT) in our department between November 2007 and January 2009 were analyzed and evaluated. TGF-beta1 levels in both serum and sputum were detected before and near the end of radiotherapy for all the patients. The TGF-beta1 level in serum was measured with enzyme-linked immunosorbent assay (ELISA). Immunocytochemistry assays were used to detect TGF-beta1 expression in sputum sediment. Radiation pneumonitis was graded according to Radiation Therapy Oncology Group (RTOG) radiation scoring criteria every 3 weeks from the start to 3 months after the end of treatment. Radiation pneumonitis was noted in 9 patients in this cohort. The total incidence of radiation pneumonitis was 39.1% (9/23) and those with Grade II or worse was 30.4% (7/23). The absolute TGF-beta1 level in serum after radiotherapy was higher than before radiotherapy, but there was no statistical difference (P = 0.139). Patients with increased levels of TGF-beta1 had a higher incidence of radiation pneumonitis (45.5%) than those with decreased TGF-beta1 levels post-radiotherapy (40.0%). Though there was a tendency of higher incidence of radiation pneumonitis with increases in TGF-beta1 level, no statistical difference was found (P = 1.000). Patients with tumor response had higher incidence of radiation pneumonitis (50.0%) than patients without when TGF-beta1 levels in serum increased, but there was no statistical difference (P = 0.792). TGF-beta1 was positively expressed (brown yellow) in sputum on immunocytochemistry assays and located in the cytoplasm of either macrophages or epithelial cells. Macrophages were the main cells expressing TGF-beta1. A significantly higher positive expression rate (71.4%) was found in sputum post-radiotherapy than pre-radiotherapy (28.6%) (P = 0.015). The higher incidence of radiation pneumonitis (46.7%) was found in patients with positive TGF-beta1 expression in sputum post-radiotherapy than those with negative expression post-radiotherapy (14.3%) (P = 0.193). It may be more reasonable to predict radiation pneumonitis by combining the change of TGF-beta1 levels in serum with tumor response than just the change of TGF-beta1 levels in serum alone. TGF-beta1 can positively express in the sputum of patients with NSCLC, located in macrophages and epithelial cells, with macrophages as the main areas of expression. Patients with positively expressed TGF-beta1 in sputum after radiotherapy have a higher incidence of radiation pneumonitis than those with negative expressions. The positive expression of TGF-beta1 in sputum is expected to become a factor for predicting radiation pneumonitis.