Abstract
Lymph node metastasis (LNM) is closely related to the postoperative recurrence of colorectal cancer (CRC), and greatly affects patient survival. Conducting Gene set variation analysis (GSVA) and gene set enrichment analysis (GSEA), we found that the epithelial-mesenchymal transition (EMT) signaling pathway is the signaling pathway most relevant to the process of LNM. An EMT-related gene signature was identified from a discovery dataset obtained 489 patients using LIMMA and LASSO Cox methods. Six external independent dataset analyses including a total of 1,045 CRC patients and stratification analysis showed that EMT-related gene signature could sort out those high- and low-risk CRC patients accurately. Functional analysis and loss-of-function exploration in vitro and in vivo indicated that the EMT-related-signature-associated coding genes might play functional roles in the sophisticated regulation of CRC proliferation and metastasis. Prognostic nomograms integrating the EMT-related gene signature and clinicopathological risk factors were constructed for use as numerical prediction tools to assess clinical prognosis and clinical decision-makings. The comprehensive transcriptomic analysis in this article highlights the prognostic value of an EMT-related gene signature for postoperative disease recurrence in CRC patients and reveals a potential prognostic and therapeutic biomarker for CRC.