Abstract
Although serum tumor biomarkers alpha-fetoprotein (AFP), carbohydrate antigen 19-9 (CA19-9) and carcinoembryonic antigen (CEA) have been used in digestive cancer risk prediction, the prediction efficiency remains unsatisfactory. The aim of this study was to evaluate whether genetic correction could improve the efficiency of these biomarkers for prediction of digestive cancer risk. We conducted a prospective analysis in 9,808 healthy individuals based on a cohort study in the elderly Chinese population. The genotypes of reported single nucleotide polymorphisms (SNPs) associated with serum AFP, CA19-9 and CEA were used to estimate the genetic corrected levels of these markers. Unconditional logistic regression analysis was performed to evaluate the risk of digestive cancer. The Harrell's C-statistic was used to evaluate the discriminative ability of the raw levels and genetic corrected levels of biomarkers on digestive cancer risk. Up to October 2013, a total of 172 individuals were newly diagnosed with digestive cancer. With the genetic correction, higher odds ratios (ORs) for digestive cancer risk were found for the genetic corrected levels of tumor biomarkers compared with their raw serum levels (1.57 vs. 1.65 for AFP; 1.19 vs. 1.21 for CA19-9; 1.09 vs. 1.10 for CEA, respectively). The same results were observed in the Harrell's C-statistic analyses. Genetic correction improved the prediction efficiency of tumor biomarkers on the digestive cancer risk in an elderly Chinese population. Our findings provide evidence for further studies of genetic effects on tumor biomarker to improve the predictive efficiency on cancer risk.