Decoding the Circulating Proteome: Matrix and Immune Context Markers Shape Early Multi-Cancer Detection
Discuss this preprint
Start a discussion What are Sciety discussions?Listed in
This article is not in any list yet, why not save it to one of your lists.Abstract
Background
Blood-based proteomics offers a complementary path to multi-cancer early detection (MCED) by capturing the tumor secretome and host response. We analyze recent (2020–2025) evidence and add pathway/hallmark context, cross-platform validation, and proteome-scale protein–protein interaction (PPI) inference to guide translational panel design.
Methods
We reviewed extensive prospective and multi-cancer studies using Olink, SomaScan, and mass spectrometry, contrasting case-control versus prospective performance. Candidates were organized into Known and Novel sets and mapped with GeneCodis (GO/KEGG/Reactome) and Cancer Hallmarks. Clinical relevance was assessed using GEPIA 3.0 Cox Forest plots and TCGA-survival Kaplan–Meier curves (median split; log-rank). Protein-level corroboration used TPCPA/RPPA Z-score distributions across tumor types. To contextualize molecular crosstalk, we incorporated in silico PPI prediction to evaluate whether candidates cluster into interaction sub-modules relevant to secretory/TGF-β, matrix remodeling, and immune-follicular biology.
Results
Beyond classic antigens/inflammatory markers (e.g., CEACAM5, WFDC2/HE4, GDF15), the Novel set converged on a matrix-immune-secretory axis comprising ECM proteases (MMP12, ADAM8), antigen presentation/B-cell programs (CD74, CXCL13), secretory/TGF-β signaling (TGFB1), and epithelial invasion (CDCP1). Forest plots show adverse hazards for secretory/matrix genes across multiple epithelial cancers, while immune-follicular genes exhibited context-dependent effects; TCGA-survival curves reproduced these directions. TPCPA demonstrated tumor-type–specific protein elevation, supporting detectability. PPI inference organized candidates into coherent interaction modules (e.g., TGFB1, CDCP1 protease, and CXCL13; CD74 hubs), reinforcing multiplex, not single-marker, readouts.
Conclusions
Cross-platform agreement, augmented by PPI-defined interaction modules, supports a two-bucket MCED strategy that pairs high-risk secretory/matrix markers with immune-context sensors to enhance sensitivity, tissue-of-origin interpretability, and clinical triage. Prospective validation and down-selection to a cost-scalable targeted assay are warranted for population screening.
