A novel prognostic model based on migrasome-related LncRNAs for gastric cancer

Gastric cancer (GC) represents a substantial public health challenge, characterized by elevated morbidity and mortality rates. Migrasomes, a newly discovered type of extracellular vesicle, have been highlighted as important contributors to cancer progression, though their specific role in GC remains unclear. To address this issue, we developed the first prognostic model utilizing migrasome-related long non-coding RNAs (MRLs). This model aims to deepen the understanding of GC pathogenesis and improve patient outcomes. Clinical and transcriptional data for 407 GC patients from TCGA were classified as training and testing sets. Through Pearson correlation analysis, 537 MRLs were recognized, and LASSO and Cox regression analyses further refined the list to four key lncRNAs (AC012055.1, LINC01150, AC053503.4, AC107021.2) for constructing the prognostic model. Kaplan-Meier survival analysis indicated a significantly poorer prognosis for the high-risk group. PCA confirmed the model's robustness, and univariate and multivariate analyses validated it as an independent predictor of clinical outcomes. The ROC curve and C-index evaluations further affirmed the model's predictive power. We developed a nomogram combining the MRLs signature with clinical parameters to enhance prognostic accuracy. GO, KEGG and GSEA were performed on migrasome-related genes associated with GC. Furthermore, high-risk patients exhibited increased immune cell infiltration and reduced tumor mutation burden, both associated with poorer outcomes. Additionally, twenty-nine potential therapeutic agents were identified. This novel MRLs-based model provides crucial insights into GC biology and represents a valuable tool for improving patient management and therapeutic strategies.