非小細胞肺癌における免疫チェックポイント阻害剤反応予測のための統合マルチオミクスアプローチ：ゲノミクス、プロテオミクス、メタボロミクスからの洞察

Immune checkpoint inhibitors (ICIs) have improved outcomes in non-small cell lung cancer (NSCLC), yet durable benefit is limited to a subset of patients. Reliable predictive biomarkers are therefore essential. We reviewed genomic, proteomic, and metabolomic studies to evaluate how multi-omics integration advances prediction of ICI efficacy in NSCLC.

A systematic search of PubMed, ClinicalTrials.gov, and Google Scholar was conducted on April 11, 2024, covering studies published from 2016 through January 2025, to identify omics-based biomarkers of ICI response in NSCLC. In total, 33 genomic, 9 proteomic, and 9 metabolomic studies met inclusion criteria. Each was evaluated using a standardized evidence rubric (0-14) assessing effect robustness, validation, cohort size, and clinical endpoint relevance.

Genomic predictors of poor response included EGFR and ALK/RET/ROS1 fusions, as well as KRAS co-mutations with STK11, KEAP1, or SMARCA4, all linked to immune-cold phenotypes with low tumor mutational burden (TMB) and poor T-cell infiltration. In contrast, KRAS/TP53 co-mutations, NOTCH family alterations, and BRAF V600E aligned with immune-hot signatures characterized by interferon signaling, PD-L1 upregulation, and cytotoxic T-cell infiltration. Proteomic studies consistently identified chemokines CXCL9 and CXCL10, apoptotic regulators (CASP8, FASLG), and checkpoint proteins (soluble PD-1, PD-L1, LAG-3) as predictive, while acute-phase proteins (SAA1/2, S100A8/9) correlated with resistance. Multi-analyte platforms such as PROphet demonstrated promising risk-stratification potential. Metabolomic profiling linked ICI benefit to higher baseline tryptophan, histidine, and short-chain fatty acids, while resistance was associated with increased 3-hydroxyanthranilic acid, pyruvate, and lipid metabolites indicating immunosuppressive IDO pathway activity.

Multi-omics approaches converge on pathways governing antigenicity, interferon signaling, and immune-metabolic crosstalk. Although promising, most biomarkers require prospective validation in large, uniformly treated cohorts. Integrative strategies-particularly when combined with AI-driven analytics-hold potential to refine patient stratification and guide clinical use of ICIs in NSCLC.