Related Experiment Video
Updated: Jun 27, 2026
A Combined 3D Tissue Engineered In Vitro/In Silico Lung Tumor Model for Predicting Drug Effectiveness in Specific Mutational Backgrounds
Published on: April 6, 2016
Construction and evaluation of a predictive model for radiation-induced lung injury in lung cancer: a meta-analysis
Min Peng1, Zhiwei Sun2, Jing Zhang1
1Department of Integrated Traditional Chinese and Western Medicine, The Affiliated People's Hospital of Ningbo University Ningbo 315000, Zhejiang, China.
Objective:
To construct and evaluate a predictive model for radiation-induced lung injury (RILI) in lung cancer patients based on a meta-analysis of observational studies.
Methods:
A systematic search was conducted across various databases to identify observational studies on the prevalence and risk factors of RILI in lung cancer from the inception of each database up until June 2024. Meta-analysis was performed using Review Manager 5.3 software to calculate the latest prevalence data and pooled risk values for significant risk factors associated with RILI. A logistic regression model was developed using the natural logarithmic transformation of the combined risk values. External validation was performed with 180 lung cancer patients who underwent radiotherapy at Yinzhou Affiliated Hospital from July 2023 to June 2024. The predictive performance of the model was assessed using the area under the receiver operating characteristic (ROC) curve (AUC), and clinical practicability was evaluated by decision curve analysis.
Results:
A total of 27 studies were included in the analysis. The meta-analysis revealed that the prevalence of RILI in lung cancer patients was 33.0% (95% confidence interval [CI]: 23.0%-42.0%). Key risk factors identified for RILI included age, mean lung dose (MLD), volume of the lung receiving P20 Gray (V20), chronic obstructive pulmonary disease (COPD), radiotherapy dose, and volume of normal lung spared from irradiation at doses > 5 Gy (AVS5). The combined odds ratios (ORs) for these factors were 2.42 (95% CI: 1.08, 5.43) for age, 1.31 (95% CI: 1.16, 1.48) for MLD, and 1.64 (95% CI: 1.02, 2.64) for V20, among others. The resulting predictive model was: Logit(P) = -0.955 + 0.884X1 + 0.270X2 + 0.495X3 + 1.688X4 + 1.147X5 - 1.966X6, where X1, X2, X3, X4, X5, and X6 represent age, MLD, V20, COPD, radiotherapy dose, and AVS5, respectively. The model's AUC was 0.875 (95% CI: 0.799-0.951), with a sensitivity of 83.3% and specificity of 91.7%.
Conclusion:
Age, MLD, V20, COPD, radiotherapy dose, and AVS5 are significant risk factors for RILI in lung cancer patients. The constructed predictive model based on these factors demonstrates strong performance, with good evaluation results, making it useful for clinical risk assessment and management.
