Developing a Prediction Model for Radiation Pneumonitis in Lung Cancer Patients Treated With Stereotactic Body Radiation Therapy Combined With Clinical, Dosimetric Factors, and Laboratory Biomarkers

BackgroundThe study aims to identify the risk factors and develop a model for predicting grade ≥ 2 radiation pneumonitis (RP) for lung cancer patients treated with stereotactic body radiation therapy (SBRT).Materials and MethodsClinical data, dosimetric data, and laboratory biomarkers from 186 patients treated with lung SBRT were collected. Univariate and multivariate logistic regression were performed to determine the predictive factors for grade ≥ 2 RP. Three models were developed by using the clinical, dosimetric, and combined factors, respectively.ResultsWith a median follow-up of 36 months, grade ≥ 2 RP was recorded in 13.4% of patients. On univariate logistic regression analysis, clinical factors of age and lung volume, dosimetric factors of treatment durations, fractional dose and V10, and laboratory biomarkers of neutrophil, PLT, PLR, and Hb levels were significantly associated with grade ≥2 RP. However, on multivariate analysis, only age, lung volume, fractional dose, V10, and Hb levels were independent factors. AUC values for the clinical, dosimetric, and combined models were 0.730 (95% CI, 0.660-0.793), 0.711 (95% CI, 0.641-0.775) and 0.830 (95% CI, 0.768-0.881), respectively. The combined model provided superior discriminative ability than the clinical and dosimetric models (P<0.05).ConclusionAge, lung volume, fractional dose, V10, and Hb levels were demonstrated to be significant factors associated with grade ≥ 2 RP for lung cancer patients after SBRT. A novel model combining clinical, dosimetric factors, and laboratory biomarkers improved predictive performance compared with the clinical and dosimetric model alone.Micro-AbstractThe age, lung volume, fractional dose, V10, and Hb levels were found to be independent factors for grade ≥ 2 RP by analyzing 186 patients with primary or metastatic lung tumor treated with SBRT. The age and Hb levels are novel predictors reported for the first time. A prediction model combining clinical, dosimetric factors and laboratory biomarkers was proposed and achieved an AUC value up to 0.830, higher than models reported in the literature.