Integrative Multiomics Analysis Identifies a Novel Gene Signature That Predicts Chemotherapy Resistance and Poor...

Chemotherapy resistance is the primary barrier to improving survival in osteosarcoma (OS), yet reliable predictive biomarkers remain limited. Here, we developed and validated a 13-gene signature via weighted gene coexpression network analysis (WGCNA) of chemotherapy-resistant and sensitive OS transcriptomes. The risk model demonstrated robust prognostic value across four independent cohorts (TARGET-OS, GSE21257, GSE16091, and GSE39055). Multiscale analysis revealed that high-risk tumors exhibit proliferative hyperactivation, genomic instability with elevated TMB, and a paradoxical immune-cold phenotype despite the high mutational load. Single-cell and spatial transcriptomics demonstrated that high-risk malignant cells display dedifferentiated, stem-like properties, preferentially localize to hypoxic necrotic peripheries, and engage in extensive tumor-stromal crosstalk. Drug sensitivity predictions confirmed resistance to first-line chemotherapy agents. Collectively, these findings establish a clinically actionable 13-gene biomarker and provide a mechanistic framework linking transcriptional profiles to chemoresistance biology, exposing novel therapeutic vulnerabilities in OS.