PURPOSE: This study aims to test the hypothesis that an enzyme-linked antibody-lectin sandwich assay for a glycovariant of CA15-3 can deliver better diagnostic performance, defined by classification accuracy, sensitivity and specificity, for breast cancer compared to an existing FDA-approved CA15-3 test. METHODS: A genetically engineered lectin (SubB2M) that specifically binds N-glycolylneuraminic acid (Neu5Gc) was used as a detection reagent in a CA15-3 capture antibody-lectin sandwich (neuCA15-3) assay. In a case: control cohort equivalence study the classification accuracy for the neuCA15-3 assay was determined and compared to an FDA-approved CA15-3 IVD test (Elecsys CA15-3 II, Roche Diagnostics). RESULTS: Classification accuracy and AUC for neuCA15-3 were 81% and 0.886 ± 0.015 (standard error, n = 567) and for Elecsys CA15-3 II, 55% and 0.642 ± 0.023 (n = 558), respectively. At a threshold cut-off serum concentration of 23.6 units/ml, overall breast cancer classification accuracy of the neuCA15-3 was 81% (compared to 55% for the comparator assay, p < 0.001). At 95% specificity, the sensitivity of the neuCA15-3 assay was 69.5%, significantly greater than the comparator assay (11.9%, p < 0.001). neuCA15-3 concentrations did not vary significantly with breast cancer receptor subtype or comorbidities tested. CONCLUSIONS: The diagnostic performance of neuCA15-3 was substantially improved by specifically targeting both a CA15-3 protein epitope and a pan-cancer glycan (Neu5Gc) epitope (the specific binding target of SubB2M). The reporter signal generated depends on the colocalization of the cancer antigen protein epitope and the aberrant sialylation of the protein, thus increasing the assay specificity. The presence of multiple Neu5Gc lectin-binding sites per glycoprotein molecule increases signal generation and assay sensitivity. The inclusion of additional cancer biomarkers in a multivariate index assay format may further increase diagnostic performance for breast cancer.