A novel algorithm to improve specificity in ovarian cancer detection

Abstract

BackgroundMeasurement of autoantibodies (AAbs) to tumor associated antigens has been proposed to aid in the early detection of ovarian cancer with high specificity. Here we describe a multiplex approach to evaluate selected peptide epitopes of p53 protein, and propose a novel approach to increase specificity and potentially sensitivity for discrimination between healthy women and women with cancerous masses. Materials and methods20-mer overlapping peptide epitopes of p53, generated by mapping the complete p53 sequence, were evaluated in a multiplex immunoassay for their detection of serum AAbs in patients with ovarian cancer, using Luminex technology. AAbs to the selected peptides and to p53 full length protein were then detected in a multiplex immunoassay evaluating 359 sera from healthy women and 285 sera from patients with early and late stage ovarian cancer. CA-125 levels were measured in all p53 AAb-positive sera. ResultsWe considered the AAb results together to identify sera where both the full length protein and at least one selected peptide epitope were positive and chose cutoffs that reduced false positives from these AAbs to 1/359 samples, improving specificity. Using this combined approach, we could identify 7 AAb-positive patients that were negative for CA-125 (concentrations below 35 IU/mL); this represents 26% of the p53 positive patients in the total population. ConclusionBy detecting p53 AAbs in CA-125-negative sera, we demonstrated that combining measurement of AAbs to the full length p53 protein and one or more selected epitopes can potentially improve sensitivity and specificity for ovarian cancer detection.