The study, published in the journal Clinical Proteomics, took an unusually personal approach to tumor analysis. Rather than pooling samples and averaging away individual differences, the researchers collected twenty-five matched pairs of tumor tissue and adjacent normal tissue from colorectal cancer patients spanning different stages of the Tumor-Node-Metastasis classification system, the standard framework surgeons and oncologists use to describe how far a cancer has advanced. By analyzing each patient’s tumor against their own healthy tissue, the team could build what they call personalized proteome profiles, capturing the unique molecular signature of every individual’s disease.

The technical engine behind the work was quantitative mass spectrometry, specifically a tandem mass tag labeling strategy coupled with liquid chromatography and high-resolution tandem mass spectrometry. This approach allowed the researchers to measure the relative abundance of thousands of proteins simultaneously across dozens of samples with remarkable precision. The depth of coverage was extraordinary: the team identified more than 12,000 distinct proteins, a figure that approaches the most comprehensive inventories of the human proteome ever assembled and far exceeds what typical cancer proteomics studies achieve.

From this vast dataset, the researchers sifted out differentially abundant proteins, molecules whose levels shifted significantly between tumor and normal tissue and, crucially, whose abundance changed as the disease progressed through increasingly advanced stages. Among the proteins that caught their attention were candidates with well-defined roles in cell biology: annexin A11, a membrane-associated protein implicated in vesicle trafficking; high mobility group box 3, a DNA-binding protein linked to chromatin dynamics; crystallin alpha B, a molecular chaperone with roles in stress response; and flap endonuclease 1, an enzyme central to DNA repair. Others included translocase of outer mitochondrial membrane 34, eukaryotic translation initiation factor 4 gamma 1, and the cell adhesion molecule L1 like protein, each pointing toward distinct vulnerabilities in the cancer cell’s machinery.

One of the study’s most conceptually interesting moves was its treatment of time. Instead of simply cataloging which proteins went up or down, the team clustered the dynamically changing proteins into six distinct profiling patterns, essentially grouping molecules according to the trajectory their abundance followed as tumors advanced. Proteins sharing a pattern, the researchers argue, are likely to be co-opted into the same phase of tumor progression, whether that involves early loss of normal epithelial function, the metabolic rewiring of intermediate disease, or the invasive and metastatic programs of late-stage cancer. This temporal framing transforms a static protein list into something closer to a molecular narrative of how colorectal cancer evolves within a patient.

The functional implications of these patterns were reinforced by pathway analysis. Enrichment of proteins involved in epithelial-mesenchymal transition, the developmental program cancer cells hijack to become mobile and invasive, emerged as a recurring theme, alongside alterations in DNA replication, metabolic pathways, and immune-related signaling. The findings align with and extend earlier large-scale efforts, including work by the Clinical Proteomic Tumor Analysis Consortium and genomic data from The Cancer Genome Atlas, but the personalized, paired-sample design gives the new data a resolution that population-level studies cannot match.

Perhaps the most clinically consequential result came from survival analysis. When the researchers examined progression-free survival, the length of time patients remained free of disease recurrence after surgery, they found a significant correlation between the abundance of their newly identified biomarker candidates and the later appearance of metastasis. In other words, the protein signatures measured in the original tumor tissue carried predictive information about which patients would go on to develop disseminated disease, information that current clinical markers such as carcinoembryonic antigen and carbohydrate antigen 19-9 do not reliably provide.

The candidate biomarkers were not left as abstract mass spectrometry measurements. The team verified their findings using complementary techniques, including immunohistochemistry on tissue sections, confirming that the protein abundance changes observed in the mass spectrometry data could be reproduced with methods already standard in hospital pathology laboratories. That matters enormously for translation: a biomarker that can only be measured in a research-grade mass spectrometry facility faces a long road to the clinic, whereas one that can be detected with an antibody-based stain on a routine biopsy specimen could, in principle, be adopted much more quickly.

Several of the highlighted proteins also suggest mechanistic stories worth pursuing. Flap endonuclease 1, for example, sits at the heart of DNA repair and its overexpression in tumors may help cancer cells tolerate the genomic chaos of rapid division, making it a potential therapeutic target as well as a marker. High mobility group box 3 has been linked to tumor immune evasion and metastatic behavior in other cancers. The cell adhesion molecule L1 like protein connects directly to the epithelial-mesenchymal transition program that underpins invasion and spread. Each of these molecules, the study suggests, may be doing more than passively marking disease stage; they may be active participants in driving it.

The researchers, led by Qi Zhang, Wenyuan Zhu, Jianguo Ji, Minzhe Li, and Qingsong Wang, frame their work as a demonstration of what proteomic technology can reveal about clinically relevant cancer signatures that genomics alone cannot capture. Proteins are the functional workhorses of the cell, and their abundance often diverges from the predictions of gene expression data. By mapping those divergences patient by patient, the study offers both a resource for the broader cancer research community and a concrete shortlist of molecules that could form the basis of new diagnostic tests for metastatic risk. Larger validation cohorts will be needed before any of these candidates reaches the clinic, but the study adds weight to a growing conviction in oncology: the next generation of cancer biomarkers will be written in protein.

Subject of Research: Quantitative proteomic profiling of colorectal cancer tumor progression and biomarker discovery

Article Title: Personalized proteomic analysis of human colorectal cancer identifies novel biomarkers associated with tumor progression

Article References: Zhang, Q., Zhu, W., Ji, J., Li, M., & Wang, Q. (2026). Personalized proteomic analysis of human colorectal cancer identifies novel biomarkers associated with tumor progression. Clinical Proteomics. https://doi.org/10.1186/s12014-026-09633-0

Image Credits: AI Generated

DOI: 10.1186/s12014-026-09633-0

Keywords: colorectal cancer, proteomics, biomarkers, tumor progression, mass spectrometry, metastasis, TNM staging, differentially abundant proteins, epithelial-mesenchymal transition, progression-free survival, tumor biology, Clinical Proteomics