A Clinical Validation of a Diagnostic Test for Esophageal Adenocarcinoma Based on a Novel Serum Glycoprotein Biomarker Panel: PromarkerEso

Abstract

Background: Esophageal adenocarcinoma (EAC) diagnosis involves invasive and expensive endoscopy with biopsy, but rising EAC incidence has not been reduced by increased surveillance. This study aimed to develop and clinically validate a novel glycoprotein biomarker blood test for EAC, named PromarkerEso.

Methods: Serum glycoprotein relative concentrations were measured using a lectin-based magnetic bead array pulldown method, with multiple reaction monitoring mass spectrometry in 259 samples across three independent cohorts. A panel of glycoproteins: alpha-1-antitrypsin, alpha-1-antichymotrypsin, complement C9 and plasma kallikrein, were combined with clinical factors (age, sex and BMI) in an algorithm to categorize the samples by the risk of EAC.

Results: PromarkerEso demonstrated a strong discrimination of EAC from the controls (area under the curve (AUC) of 0.91 in the development cohort and 0.82 and 0.98 in the validation cohorts). The test exhibited a high sensitivity for EAC (98% in the development cohort, and 99.9% and 91% in the validation cohorts) and a high specificity (88% in the development cohort, and 86% and 99% in the validation cohorts). PromarkerEso identified individuals with and without EAC (96% and 95% positive and negative predictive values).

Conclusions: This less invasive approach for EAC detection with the novel combination of these glycoprotein biomarkers and clinical factors coalesces in a potential step toward improved diagnosis.

Keywords: Gastrointestinal diseases; LeMBA; MRM-MS; PromarkerEso; biomarkers; cancer diagnosis; esophageal adenocarcinoma (EAC); glycoprotein; targeted mass spectrometry.

Figure 1

Participant recruitment details and allocations of samples for modelling. EAC = esophageal adenocarcinoma, GenPop = general population controls, NC = control by negative endoscopy.

Figure 2

Differential glycoprotein fold change abundance in esophageal adenocarcinoma compared to negative controls in PROBE-NET cohort for development of assay. Relative peptide concentration is represented here with first three letters of peptide sequence by log2 fold change of peak area ratio of each glycoprotein over internal standard. Full sequences are available in Supplementary Table S2. p values are from bivariate analysis using Mann–Whitney U Testing between EAC and negative controls. EAC = esophageal adenocarcinoma, SERPINA1 = alpha-1-antitrypsin, C9 = complement C9, HP = haptoglobin, HPX = hemopexin, PON1 = serum paraoxonase/arylesterase, KLKB1 = plasma kallikrein.

Figure 3

Overall Receiver Operator Curve for PromarkerEso assay in development cohort and validation cohorts. The grey line represents the line of no discrimination (area under the curve = 0.5).

Figure 4

Distribution of esophageal adenocarcinoma (EAC) and control sample probability scores as classified by PromarkerEso into Low-, Moderate- and High-Risk categories in development cohort (A) PROBE-NET and validation cohorts (B) Ochsner and (C) VCB. Actual outcomes are represented as blue dots (EAC) and black dots (Negative Controls or General Population). Lower and upper cutoffs (20% and 75% respectively) are represented by red dotted lines.