Tumor‐associated autoantibody signature for the early detection of gastric cancer

Abstract

Autoantibodies against tumor-associated antigens are very attractive biomarkers for the development of noninvasive serological tests for the early detection of cancer because of their specificity and stability in the sera. In our study, we applied T7 phage display-based serological analysis of recombinant cDNA expression libraries technique to identify a representative set of antigens eliciting humoral responses in patients with gastric cancer (GC), produced phage-antigen microarrays and exploited them for the survey of autoantibody repertoire in patients with GC and inflammatory diseases. We developed procedures for data normalization and cutoff determination to define sero-positive signals and ranked them by the signal intensity and frequency of reactivity. To identify autoantibodies with the highest diagnostic value, a 1,150-feature microarray was tested with sera from 100 patients with GC and 100 cancer-free controls, and then the top-ranked 86 antigens were used for the production of focused array that was tested with an independent validation set comprising serum samples from 235 patients with GC, 154 patients with peptic ulcer and gastritis and 213 healthy controls. The receiver operating characteristic curve analysis showed that 45-autoantibody signature could discriminate GC and healthy controls with area under the curve (AUC) of 0.79 (59% sensitivity and 90% specificity), GC and peptic ulcer with AUC of 0.76 and GC and gastritis with AUC of 0.64. Moreover, it could detect early GC with equal sensitivity than advanced GC. Interestingly, the autoantibody production did not correlate with histological type, H. pylori status, grade, localization and size of the primary tumor, whereas it appeared to be associated with the metastatic disease.