Identification of Anoikis‐Related Subgroups and Prognosis Model in Colon Cancer

Background Anoikis represents a distinct type of programmed cell death that serves a critical function in tumor development and metastatic. Despite its significance, there remains a notable gap in comprehensive studies that explore the role of Anoikis specifically within the context of colon cancer (CC). Thus, it becomes imperative to pinpoint Anoikis‐related genes (ARGs) that could provide valuable prognostic insights into the progression and outcomes of colon cancer. Identifying these ARGs may ultimately enhance our understanding of the disease and contribute to improving patient management strategies. Methods RNA sequencing data, along with clinical details from cancerous tissue samples, were sourced from the TCGA and GEO databases, in addition to the Genecards and Harmonizome platforms, to identify differentially expressed ARGs (DE‐ARGs). The ensuing analyses encompassed Gene ontology (GO) enrichment, Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment, and protein–protein interaction (PPI) examinations. To discover DE‐ARGs associated with prognosis, single‐factor Cox regression analysis was employed, alongside evaluations of single nucleotide variations (SNVs), copy number variations (CNVs), and methylation levels. Consensus clustering was applied to categorize patient subtypes and to investigate differences in survival outcomes, clinical characteristics, and immune cell infiltration (ICI) across these groups. A risk assessment model was constructed utilizing LASSO‐ Cox regression analysis. The model's viability in a clinical context was validated through survival curve analysis, receiver operating characteristic (ROC) curve assessment, and the generation of array plots. Additionally, the analysis explored variations in the ICI, tumor mutation burden (TMB), microsatellite instability (MSI), immune checkpoint (ICP) characteristics, and drug sensitivity of the model. Finally, immunohistochemistry (IHC) analysis confirmed the discrepancies in expression levels and prognostic implications between NATI and OGT. Results A total of 161 DE‐ARGs and 32 prognosis‐related DE‐ARGs were identified. Consensus clustering analysis divided CC patients into two unique subtypes. These subtypes showed differences in clinicopathological features, ICI, and prognostic outcomes. A prognostic risk model consisting of five ARGs (TIMP1, NAT1, EDAR, HOTAIR, and OGT) was constructed. Significant variations were observed in ICI, TMB, drug sensitivity, MSI, and expression of ICP‐related genes. Drawing from the findings of both single‐factor and multi‐factor prognostic assessments, a nomogram was developed that incorporates risk score, age, and stage classification. This nomogram's clinical effectiveness was validated by using calibration curve, cumulative hazard curve, and decision curve analysis (DCA). IHC findings supported the distinct expression levels of NAT1 and OGT in CC tissues as well as in adjacent tissues. A low level of NAT1 paired with a high level of OGT suggested a negative prognosis for the patients. Conclusion According to the analysis of ARGs, CC can be divided into two categories. A prognostic risk model related to Anoikis for patients with CC was developed, highlighting its potential effectiveness in forecasting patient outcomes, understanding immune microenvironment changes, and evaluating chemotherapy responses. This work thus establishes a crucial theoretical basis for upcoming personalized precision therapies.