Comparison of Proteomic Technologies for Blood-Based Detection of Colorectal Cancer

Abstract

Blood-based protein biomarkers are increasingly being explored as supplementary or efficient alternatives for population-based screening of colorectal cancer (CRC). The objective of the current study was to compare the diagnostic potential of proteins measured with different proteomic technologies. The concentrations of protein biomarkers were measured using proximity extension assays (PEAs), liquid chromatography/multiple reaction monitoring-mass spectrometry (LC/MRM-MS) and quantibody microarrays (QMAs) in plasma samples of 56 CRC patients and 99 participants free of neoplasms. In another approach, proteins were measured in serum samples of 30 CRC cases and 30 participants free of neoplasm using immunome full-length functional protein arrays (IpAs). From all the measurements, 9, 6, 35 and 14 protein biomarkers overlapped for comparative evaluation of (a) PEA and LC/MRM-MS, (b) PEA and QMA, (c) PEA and IpA, and (d) LC/MRM-MS and IpA measurements, respectively. Correlation analysis was performed, along with calculation of the area under the curve (AUC) for assessing the diagnostic potential of each biomarker. DeLong's test was performed to assess the differences in AUC. Evaluation of the nine biomarkers measured with PEA and LC/MRM-MS displayed correlation coefficients >+0.6, similar AUCs and DeLong's p-values indicating no differences in AUCs for biomarkers like insulin-like growth factor binding protein 2 (IGFBP2), matrix metalloproteinase 9 (MMP9) and serum paraoxonase lactonase 3 (PON3). Comparing six proteins measured with PEA and QMA showed good correlation and similar diagnostic performance for only one protein, growth differentiation factor 15 (GDF15). The comparison of 35 proteins measured with IpA and PEA and 14 proteins analyzed with IpA and LC/MRM-MS revealed poor concordance and comparatively better AUCs when measured with PEA and LC/MRM-MS. The comparison of different proteomic technologies suggests the superior performance of novel technologies like PEA and LC/MRM-MS over the assessed array-based technologies in blood-protein-based early detection of CRC.

Keywords: LC/MRM-MS; biomarkers; colorectal cancer; diagnosis; microarray; plasma proteins; proximity extension assays.

Figure 1

STARD (Standards for the Reporting of Diagnostic Accuracy) flow diagram (BLITZ study). Abbreviations: CRC, colorectal cancer; IpA, immunome protein array; LC/MRM-MS, liquid chromatography/multiple reaction monitoring–mass spectrometry; n/N, number; SD, standard deviation; PEA, proximity extension assay; QMA, quantibody microarray. * The exclusion criteria for the selection of CRC cases were not applicable after this point.

Figure 2

Distribution of AUCs of nine protein biomarkers overlapping between PEA and LC/MRM-MS measurements for detecting CRC cases from participants in screening colonoscopy. Abbreviations: AUC, area under the receiver operating curve; CDH5, cadherin 5; CRC, colorectal cancer; IGFBP2, insulin-like growth factor binding protein 2; LC/MRM-MS, liquid chromatography/multiple reaction monitoring–mass spectrometry; MASP1, mannan binding lectin serine protease 1; MMP9, matrix metalloproteinase 9; MPO, myeloperoxidase; PEA, proximity extension assay; PON3, serum paraoxonase lactonase 3; PRTN3, myeloblastin; SPARC, secreted protein acidic and rich in cysteine; TR, transferrin receptor protein 1.

Figure 3

Distribution of AUCs of six protein biomarkers overlapping between PEA and QMA measurements for detecting CRC cases from participants in screening colonoscopy. Abbreviations: AREG, amphiregulin; AUC, area under the receiver operating curve; CEA, carcinoembryonic antigen; CRC, colorectal cancer; FASLG, Fas ligand; GDF15, growth differentiation factor 15; KRT19, keratin Type I cytoskeletal 19; OPN, osteopontin; PEA, proximity extension assay; QMA, quantibody microarray.

Figure 4

Distribution of AUCs of 11 protein biomarkers overlapping between PEA and IpA measurements for detecting CRC cases from participants in screening colonoscopy. Abbreviations: AREG, amphiregulin; AUC, area under the receiver operating curve; CHI3L1, chitinase-3-like protein 1; CRC, colorectal cancer; CTSD, cathepsin D; GAL1, galectin-1; GDF15, growth differentiation factor 15; IpA, immunome protein array; KRT19, keratin Type I cytoskeletal 19; MMP2, matrix metalloproteinase-2; OPG, osteoprotegerin; PEA, proximity extension assay; PSIP1, PC4 and SFRS1-interacting protein; TR, transferrin receptor protein 1; VIM, vimentin.

Figure 5

Distribution of AUCs of 14 protein biomarkers overlapping between LC/MRM-MS and IpA measurements for detecting CRC cases from participants in screening colonoscopy. Abbreviations: A1BG, alpha-1B-glycoprotein; A2HS, alpha-2-HS-glycoprotein; AUC, area under the receiver operating curve; C4BPA, C4b, binding protein alpha chain; C6, complement component C6; CA1, carbonic anhydrase 1; CD5, CD5 antigen; CRC, colorectal cancer; HSPA5, 78 kDa glucose-regulated protein; IGHG1, immunoglobulin gamma-1 chain C region; IpA, immunome protein array; LC/MRM-MS, liquid chromatography/multiple reaction monitoring–mass spectrometry; PLS2, plastin-2; S100A9, protein S100-A9; SAP, serum amyloid P-component; SERPINF2, alpha-2-antiplasmin; SPARC, secreted protein acidic and rich in cysteine; TR, transferrin receptor protein 1.