High-Throughput Olink Proteomics Elucidates the Immuno-Neurological Landscape of Gastric Cancer

Clinical Snapshot

Key Findings

• P Value: Not reported in abstract; statistical thresholds for differential expression not specified

• Effect Size: Not reported in abstract; fold-change magnitudes and NPX differences for individual DEPs not specified

• Primary Outcome: 124 differentially expressed proteins (DEPs) identified between gastric tumour tissue and paired adjacent/normal tissues, predominantly associated with inflammation, immunity, and neurology; protein families including chemokines, interleukins, and C-type lectin domain-containing proteins (CLECs) were major contributors

• Nnt Or Sensitivity: No diagnostic accuracy metrics (AUC, sensitivity, specificity) reported for candidate biomarkers CCN4, FNDC1, or PGF; machine learning model performance parameters not disclosed in abstract

• Confidence Interval: Not reported

Clinical Application

Olink PEA technology is commercially available but requires specialised laboratory infrastructure and is not currently part of routine clinical pathology workflows. The identified candidate biomarkers (CCN4, FNDC1, PGF) are not yet validated for clinical use. ELISA-based measurement of these proteins is technically feasible but requires clinical-grade assay development and regulatory approval before clinical deployment. Gastric cancer represents approximately 2,200 new diagnoses annually in Australia (AIHW data), with higher incidence in populations of East Asian, Southern European, and Indigenous background. The RACGP and Cancer Australia do not currently recommend proteomic biomarker panels for GC diagnosis or surveillance. None of the candidate biomarkers (CCN4/WISP1, FNDC1, PGF) are listed on the PBS or approved by the TGA for GC diagnostic use. Australian GC management follows NHMRC-endorsed guidelines emphasising endoscopic diagnosis, HER2 testing (for trastuzumab eligibility on PBS), and PD-L1 assessment (for pembrolizumab eligibility). The immuno-neurological proteomic signatures described may have future relevance to understanding treatment resistance in Australian patients receiving immune checkpoint inhibitors, but this requires prospective validation in ethnically diverse cohorts including Australian patients. Adult patients with confirmed gastric adenocarcinoma undergoing surgical resection; findings are most directly applicable to East Asian populations given the study demographics. Potential future applicability to biomarker-guided diagnosis or tumour microenvironment characterisation in GC patients broadly, pending independent validation.

Abstract

Gastric cancer (GC) ranks as the third leading cause of cancer-related mortality globally. A comprehensive and precise proteomic study of GC tissues can yield detailed insights into its pathogenesis. This study employed the highly sensitive Olink technology to quantify the levels of 1034 selected proteins in tumor tissues, paired adjacent tissues, and normal tissues from Chinese GC patients. Through an examination of expression changing trends and differential expression across these tissue types, it was determined that the protein profile of tumor tissue exhibits distinct characteristics, whereas the protein profiles of adjacent and normal tissues are more similar. The 124 differentially expressed proteins (DEPs) are predominantly associated with inflammation, immunity, and neurology, with major contributions from protein families such as chemokines, interleukins, and CLECs. We also employed machine learning algorithms to identify potential biomarkers, including CCN4, FNDC1, and PGF. Furthermore, validation results obtained through transcriptome data and ELISA demonstrated a concordance with findings from Olink analysis. To the best of our knowledge, this study represents the inaugural proteomic investigation of GC tissue utilizing high-throughput Olink technology. It reveals distinctive proteomic characteristics at the tissue level and offers significant insights for the identification and validation of novel biomarkers.