Tianjin Medical Journal ›› 2026, Vol. 54 ›› Issue (10): 1081-1086.doi: 10.11958/20260425

• Clinical Research • Previous Articles     Next Articles

Construction and validation of a nomogram prediction model for symptomatic radiation pneumonitis after radiotherapy in patients with NSCLC

MAO Yixiang(), WANG Haidong△(), HANG Ye, SONG Jin   

  1. Department of Radiation Oncology, Yancheng Third People’s Hospital, Yancheng 224000, China
  • Received:2026-02-06 Revised:2026-03-27 Published:2026-10-15 Online:2026-10-10
  • Contact: △E-mail:1499299393@qq.com

Abstract:

Objective To construct and validate a nomogram prediction model for symptomatic radiation pneumonitis (SRP) in patients with non-small cell lung cancer (NSCLC) after radiotherapy. Methods A total of 360 patients with NSCLC were selected and divided into the training set (270 cases) and the validation set (90 cases) at a 3∶1 ratio using a simple random sampling method. All patients were followed up for 6 months after the end of radiotherapy. Patients in the training set were divided into the SRP group (83 cases) and the non-SRP group (187 cases) according to whether SRP occurred. General clinical data, laboratory indicators and dosimetry parameters were compared between the two groups. Multivariate Logistic regression was used to analyze factors influencing the occurrence of SRP in patients with NSCLC after radiotherapy. A nomogram prediction model for SRP was constructed and validated. Results Compared with the non-SRP group, the SRP group showed higher proportions of patients with history of underlying pulmonary disease, patients with Eastern Cooperative Oncology Group (ECOG) scores of 2 and patients receiving sequential radiotherapy and chemotherapy, and a lower proportion of patients with central-type lesions, higher procalcitonin levels and neutrophil count, lower levels of platelet count and eosinophil count, larger planning target volume (PTV), higher percentage of total lung volume receiving≥5 and ≥20 Gy irradiation (V5, V20, P<0.05). Multivariate Logistic regression analysis showed that having history of underlying pulmonary disease, ECOG score of 2, increased PTV, sequential radiotherapy and chemotherapy and high V20 were independent risk factors for SRP in patients with NSCLC, while central-type lesion was an independent protective factor (P<0.05). The areas under the receiver operating characteristic (ROC) curve of the constructed model for predicting SRP in the training set and the validation set were 0.826 (95%CI: 0.775-0.869) and 0.897 (95%CI: 0.815-0.951), respectively. Hosmer-Lemeshow goodness of fit test showed χ2=3.913, P=0.865 in the training set, and χ2=6.482, P=0.593 in the validation set, indicating good model fit. The calibration curve was relatively close to the ideal curve, and decision curve analysis indicated that the model could obtain net benefits. Conclusion The nomogram prediction model constructed based on risk factors in patients with NSCLC after radiotherapy demonstrates good predictive efficacy.

Key words: carcinoma, non-small-cell lung, radiotherapy, radiation pneumonitis, nomogram, symptomatic radiation pneumonitis

CLC Number: