A Potential Biomarker — Vitronectin Predicting for Grade ≥2 Radiation Pneumonitis in Lung Cancer Patients Receiving Thoracic Radiotherapy

Our previous findings have identified vitronectin (VN) as a potential biomarker for radiation pneumonitis (RP) through proteomics and molecular mechanism studies. In the present study, we further explore associations between plasma level and single nucleotide polymorphism (SNP) analysis of VN and the risk of developing grade ≥2 RP in lung cancer patients receiving radiotherapy. We identified a cohort of 173 patients with lung cancer received radiotherapy. Blood samples were collected from all study subjects within one week before radiation. VN reference SNP rs704 and rs2227721 were genotyped in these patients. Plasma VN concentrations were detected by enzyme-linked immunosorbent assay (ELISA)before irradiation. Clinical variables and genotypes associated with risk of grade ≥2 RP were analyzed by univariate and multivariate Cox regression. Kaplan-Meier curves were applied to estimate the cumulative RP probability, and receiver operating characteristic (ROC) curve was used to identify cutoff values. T-test and one-way ANOVA were conducted to evaluate the expression of plasma levels. A total of 63 (36.4%) patients developed grade ≥2 RP and 19(11.0%) patients suffered grade 3 RP. Tumor histology, use of chemotherapy, tumor status, nodal status, V20 and mean lung dose (MLD) were associated with grade ≥2 RP risk in univariate analysis (P< 0.05). On multivariate analysis, the VN rs704 GA/GG and rs2227721 AA/AC genotypes had a statistically significantly lower risk of grade ≥2 RP than rs704 AA and rs2227721 CC genotypes (adjusted hazard ratio [HR],0.479; 95% confidence interval [CI],0.269-0.850; P = 0.012; adjusted HR,0.467; 95% CI,0.247-0.883; P = 0.019, respectively). The baseline secretion level of VN in patients with grade ≥2 RP (143.8±71.3μg/mL)was significantly higher than that in grade≤1 RP ones(66.1±28.9μg/mL) (P<.0001), and it was proportional to the severity of RP (grade 3 vs grade 2, P<.0001 & grade 2 vs grade 1, P<.0001).Moreover, we found significantly elevated VN plasma level in AA genotype of rs704 compared with GA/GG genotypes(P = 0.048). In addition, combining the cutoff points of MLD and plasma level, RP risk groupings were as followed: high risk, MLD≥11Gy and plasma level ≥90μg/mL(31of 44 patients,70.5%); intermediate risk, MLD≥11Gy and level<90μg/mL or MLD<11Gy and level ≥90μg/mL(26 of 73 patients,35.6%); and low risk, MLD<11Gy and plasma level<90μg/mL(6 of 56 patients,10.7%) (P<.0001). Patients with relatively high plasma levels of VN before radiotherapy were associated with the higher risk of RP, and polymorphisms at rs704 and rs2227721 each had a significant effect on radiation sensitivity. In addition, we identified three risk groups in combining MLD with plasma level to evaluate the incidence of grade ≥2 RP. Based on our previous findings, this study further indicated that VN may serve as a blood biomarker for susceptibility to RP in patients with lung cancer before irradiation.