DVH, Pneumonia, Lung Cancer Radiation Value Of ROC And Cytokine Forecast

OBJECTIVE (1) To investigate the clinical parameters correlated with radiation pneumonitis (RP) by independent sample t test, chi-square, and logistic regression.(2) To identify if clinical parameters and dose-volume histogram parameters improve the prediction of RP.(3)To identify if DVH combining receiver operating characteristic curve would improve the prediction efficiency for RP.(4) Further to investigate if DVH parameters using ROC combined with serum transforming growth factor β1(TGF-β1) could provide better prediction efficiency for RP.METHODS A total of85patients of non-small cell lung cancer who were treated with three dimensional conformal radiotherapy or intensity-modulated radiotherapy were prospectively enrolled. After treatment, the patients were followed-up for3months. The grade of RP was evaluated according to the NCI CTC3. The correlation between DVH parameters, TGF-β1or DVH parameters using ROC combined with TGF-β1and RP were analyzed.RESULTS (1) The ventilation injury classification and lung perfusion was not associated with RP. Even that almost all PFTs parameters are inclined to decrease in RP group, there were not significan difference.(2) Multi-parameters of DVH (V5-V40, MLD) were statistically relative with the RP development. MLD only is associated with RP by multivariate analysis (3) Receiver operator characteristic curve (ROC) displayed that the V5、V10、V13、V20were correlated with RP (P<0.05), and single DVH has no advantages to predicte the RP.(4) The concentration of cytokine (TGF-β1) was significant increased after RP than before RP. Furthermore, the ratio of TGF-β1after RT and before between group RP and no-RP>1. Cytokines if combined with multi-parameters, the predictive effency may be improved.CONCLUSION (1) Invesgating the fundamental lung function with SPECT lung perfusion examination and pulmonary function tests before treamnet is for selecting high risk populations of RP.(2)Multi-parameters of DVH (V5-V40, MLD) and FEV15were all statistically signifcantly relative with the RP development by univariate analysis (all values of P<0.05). MLD only was associated with RP by multivariate analysis (P=0.002).(3) ROC combing with DVH could offset the disadvantages of single DVH.(4) The concentration of TGF-β1was higher in RT patients. This result was more significant than on-RP patients. The multi-parameters combined with ROC could improve predictive efficiency.