Multi-omics integration identifies serum FTL as a novel diagnostic biomarker for esophageal squamous cell carcinoma

Background Esophageal squamous cell carcinoma (ESCC) is a common malignant tumor in China, with a poor five-year survival rate due to the lack of effective early diagnostic methods. This study aimed to identify potential non-invasive biomarkers and clarify their underlying mechanisms. Methods We employed an integrative approach combining serum proteomic profiling, bulk RNA sequencing, and single-cell RNA-seq (scRNA-seq) analyses. Quantitative proteomics was conducted on 15 serum samples of treatment-naive early-stage ESCC patients and 5 from healthy controls. Differentially expressed proteins (DEPs) were cross-referenced with differentially expressed genes (DEGs) from the GSE53624 dataset, and a diagnostic model was developed using LASSO regression. Cross-validation was performed using scRNA-seq data. The enzyme-linked immunosorbent assay (ELISA) was utilized to validate FTL in 155 serum samples from two centers. In silico FTL knockouts were conducted using scTenifoldKnk, followed by GO, KEGG, and PPI network analyses. Results Proteomic profiling identified a total of 129 DEPs. Intersection with GSE53624 DEGs yielded 8 overlapping genes, and a 6-biomarker (TF, CHAD, CSRP1, FCGR3A, RCN1, FTL) diagnostic model was built, with AUC 0.882–1.0 across validation cohorts. scRNA-seq intersection analysis identified FTL as the only overlapping DEG, exhibiting high expression levels in ESCC epithelial cells and macrophages. FTL demonstrated strong standalone diagnostic performance, with AUC values ranging from 0.872 to 0.970 across three GEO datasets. ELISA validation yielded an AUC of 0.75 (95% CI: 0.674–0.816) and 0.838 (95% CI: 0.751–0.905) for early-stage ESCC, surpassing the performance of CEA and SCCA. The combination of FTL with SCCA further enhanced discriminatory efficacy. Mechanistically, FTL was associated with altered immune cell infiltration, characterized by increased resting dendritic cells and M2 macrophages and decreased follicular helper T cells, as well as activation of the ferroptosis pathway. In silico FTL knockouts revealed DEGs enriched in stress and metabolic biological processes, extracellular vesicles (cellular component), transcriptional regulation (molecular function, CHOP-ATF3 complex), IL-17/TNF pathways, and lipid/ROS-related pathways. Protein-protein interaction hubs included FOS, ATF3, NFKBIA, and CXCL8, with FTL linked to NFE2L2, TFRC, and TXNRD1. Conclusions FTL represents a credible non-invasive serum diagnostic biomarker for ESCC. Computational analyses suggest it is a candidate molecule requiring further mechanistic and therapeutic investigation.