Significance of Serum N-glycan Profiling as a Diagnostic Biomarker in Urothelial Carcinoma

Abstract

BackgroundThe clinical diagnosis of urothelial carcinoma (UC) relies on invasive methods in patients with hematuria. Although more sensitive and noninvasive screening methods are required, a specific serum biomarker for UC is lacking. ObjectiveTo examine whether serum glycan-based biomarkers can be applied to UC detection. Design, setting, and participantsBetween April 1994 and June 2016, serum N-glycan concentrations were retrospectively measured in 212 patients with UC before treatment (UC group) and 212 pair-matched controls using glycoblotting and mass spectrometry. Outcome measurements and statistical analysisN-glycan levels were compared between the groups using receiver operating characteristic curves to select candidate N-glycans. We created an N-glycan score based on the combination of candidate N-glycans. The specificity and sensitivity of the candidate N-glycan score were evaluated using receiver operating characteristic curves. Results and limitationsThe N-glycan score was calculated using six N-glycans (m/z 1566, m/z 1687, m/z 1769, m/z 1871, m/z 2011, and m/z 2337) that were significantly associated with UC. The median N-glycan score was significantly higher in the UC group than in the pair-matched control group (5.0 vs 1.0, p<0.001). The N-glycan score correctly classified UC patients with a sensitivity, specificity, and area under the curve of 93%, 81%, and 0.95, respectively. The limitations of our study included its retrospective nature and nonclinical setting. ConclusionsSerum N-glycan content has the potential to be a specific and sensitive novel serum biomarker that may improve the accuracy of the detection for UC and reduce unnecessary invasive screening. Validation of this test in a large-scale prospective study is needed. Patient summaryCombination of serum N-glycan (N-glycan score) is a novel serum marker for urothelial carcinoma that is expressed by 93% of patients and thus is far more sensitive than classic urine cytology. Validation in a large patient cohort is needed.