Exploration of Multiomic Profiles and Biomarkers as Predictors of Neoadjuvant Chemoradiotherapy Responsiveness in Esophageal Squamous Cell Carcinoma

Abstract

The current gold standard of care for resectable locally advanced esophageal cancer is neoadjuvant chemoradiotherapy (NCRT) followed by surgery. Given that only 30-40% of patients with esophageal squamous cell carcinoma (ESCC) achieved a pathologic complete response (pCR) following neoadjuvant chemoradiotherapy, it is critical to understand the biological basis of NCRT resistance in esophageal cancer and identify biomarkers for these patients in order to further personalize treatment plans. We aim to depict the biological landscape of ESCC responsiveness and resistance to neoadjuvant chemoradiotherapy. Endoscopic biopsied specimens of the primary tumors and paired peripheral blood samples were obtained from 24 patients before neoadjuvant chemoradiotherapy and tested for whole exosome sequencing, RNA sequencing, and DIA mass spectrometry. Genomic data were analyzed for significantly mutated genes, copy number alterations, MSI, TMB, and mutational signatures. Transcriptomics and proteomics data were used to examine differentially activated pathways. GSEA and ActivePathways were used for the single omics level and joint multi-omics analysis, respectively. Tumor microenvironment (TME) characteristics were deconvoluted by xCell upon RNA-seq data. Treatment resistance biomarkers were identified and validated in a separate cohort using mIHC. In the study cohort, 54% (13/24) of the patients achieved pCR. WES data suggested that FBXW7 was more frequently mutated in the pCR group (fisher test p-value = 0.029), and the most significant cytoband loss in the pCR group was 9p21.3 (qval = 0.001). Differences in TMB, MSI, and mutational signatures were not significant between groups. Combined transcriptomics and proteomics analysis revealed that type I interferon signaling pathways and RIG-I-like receptor signaling pathways(p<0.05) were enriched in non-pCR tumors. Esophageal cancer cohort RNA-seq data from TCGA verified the correlation between the genetic variances (FBXW7 mutation and 9p21.3 loss) and the decreased expression of type I interferon signaling pathway genes. In TME analysis, tolerogenic dendritic cells and exhausted T cell signatures were significantly enriched in non-pCR tumors, indicating an immunosuppressive status in treatment resistant patients. Based on proteomics PPI network and differential expression genes from RNA-seq data, a biomarker panel consisted of 12 proteins predictive of non-pCR tumors was identified: STAT1, EIF2AK2, MX1, BST2, TRIM21, SAMHD1, IFI44L, GBP1, PARP14, ISG15, IFIT3, and HLA-B. The expression of selected genes was validated by mIHC in an independent cohort. Through a multiomics approach, we described the biological characteristics of ESCC with distinct responses to neoadjuvant chemoradiotherapy and proposed a panel of 12 proteins as predictive biomarkers for non-pCR patients.