Comprehensive machine learning identifies anoikis signatures predicting therapeutic resistance and survival in gastric cancer

Gastric cancer (GC), the fifth most common malignancy worldwide, presents a significant challenge to global public health, with drug resistance emerging as a leading cause of treatment failure. Anoikis is a specialized form of cell death that occurs when cells detach from the extracellular matrix, and it has been shown to play a pivotal role in tumor progression. In this study, we explored potential biomarkers for prognosis and resistance decision-making in GC, focusing on anoikis. By utilizing data from 10 publicly available GC databases, we enriched immune-related and anoikis-related genes and identified a set of drug-naturalized anoikis genes (DNAGs). These DNAGs exhibit a high anoikis index, are highly sensitive to immunotherapy, show significant alterations in the immune microenvironment, and are closely associated with chemotherapy resistance. Through unsupervised clustering analysis of 125 DNAGs, we categorized the gastric cancer samples into two clusters. Cluster 1 was characterized by higher tumor mutation burden, greater immune cell infiltration, and increased oncogene mutation frequency. In contrast, Cluster 2 included a higher proportion of younger patients, was predominantly enriched in the GS molecular subtype, and exhibited a more pronounced immunosuppressive microenvironment with fibroblast enrichment. Subsequently, we constructed a treatment-oriented prognosis signature (TOPS) by applying combinations of 101 machine learning algorithms to 25 genes highly expressed in Cluster 2. This approach resulted in the identification of 11 key genes that constitute the final signature. Using combinations of 101 machine learning algorithms, we identified 11 key genes that constitute this signature. Significant differences were observed between the high and low TOPS groups in terms of survival rate, immune microenvironment, tumor stage, and invasive subtypes, indicating that patients in the high TOPS group have a poorer survival prognosis and more aggressive tumors. Moreover, the high TOPS group demonstrated marked resistance to chemotherapy and immunotherapy, particularly to immune checkpoint inhibitors and various chemotherapeutic agents. In conclusion, through a comprehensive analysis of the molecular and immune characteristics of GC, this study revealed the clinical significance of DNAGs and TOPS in GC, providing insights into new potential targets for personalized treatment strategies.