Serum CRP Is Predictive if Survival in Non-Small-Cell Lung Cancer Patients Treated With Definitive Chemoradiation Therapy

had both DNA samples and preand post-CRT PFTs from Nov 1, 1998 through July 31, 2009 were included in this study. There were 73 males and 58 females with a median age of 64 years, and 71% were ever smokers with a median pack year of 40. There were 40% adenocarcinoma, 37% squamous cell carcinoma (SCC), and 23% NSC, NOS. Of these patients, 48.1%, 35.1%, and 16.8% received 3D conformal radiation therapy (3DCRT), intensity modulated radiation therapy (IMRT), and proton therapy (PBT), respectively. The percentage of predicted DLCO and forced expiration volume in 1 second (FEV1) before and within 6 months or 1 year post-RTwere collected. The end points were the mean and the lowest ratio of DLCO or FEV1 from postto pre-RT. Linear regression was used to test the association between SNPs and the change in DLCO and FEV1. Clinical covariates with P values <0.20 were enrolled into multivariate model. Results: Gender was the only risk factor for ratio of FEV1 (P<0.001), and the mean lung dose (MLD) was a significant risk factors only for the ratio of DLCO (P<0.001). For ratios of FEV1, the CC/CT genotype of rs228590 in ATM gene (P<0.05) associated with both lower mean and lowest ratios within 1 year after CRT. For ratios of DLCO within 6 months after CRT, the CC/CT genotype of the rs1892073 in the TGF-b1 gene (P Z 0.046) and the GG genotype of rs1800629 in the TNFa gene (P Z 0.01) were associated with lower mean DLCO ratios; the GG genotype of rs1800629 in the TNFa gene (P Z 0.024) was associated with lower lowest DLCO ratios. For ratios of DLCO within 1 year after CRT, the CC genotype of rs2010963 in VEGF gene (PZ 0.01) was associated with both lower mean and lowest DLCO ratios. In addition, the AG/GG genotype of rs25487 in the XRCC1 gene was also associated with lower lowest DLCO ratios within 1 year after CRT (P Z 0.048). Conclusions: We found associations between PFT changes and a number of high risk SNPs for RP that were also identified in our previous studies for RP. The next step is to incorporate the SNPs into the conventional NTCP MLD model or effective dose model to determine if SNPs can improve the predictive accuracies of these models. Author Disclosure: Y. Song: None. T. Xu: None. D. Gomez: None. L. Levy: None. Z. Liao: None.