Integrative machine learning reveals the biological function and prognostic significance of α-ketoglutarate in gastric cancer.

Gastric cancer (GC) has a poor response to treatment, an unfavorable prognosis and a lack of reliable biomarkers for predicting disease progression and therapeutic outcomes. α-Ketoglutarate (α-KG) is a critical metabolite involved in cellular energy metabolism and epigenetic regulation during tumor development, which has emerged as a potential prognostic biomarker for GC. The present study aimed to explore this potential using publicly available datasets from The Cancer Genome Atlas and Gene Expression Omnibus databases to analyze α-KG-related genes and establish the α-KG Index (AKGI). By assessing the predictive performance of the AKGI model, the results demonstrated its capability to predict survival outcomes in patients with GC. Notably, high AKGI scores were associated with worse prognoses. Building on these findings, the associations between AKGI and clinical variables, immune cell infiltration and tumor mutation characteristics were assessed, further identifying potential therapeutic drugs for patients with high AKGI scores. Additionally, by analyzing signaling pathways and biological functions correlated with AKGI, the regulatory mechanisms and biological roles of α-KG in GC were elucidated. The findings of these analyses were further evaluated using cellular experiments, where α-KG treatment was demonstrated to significantly inhibit GC cell proliferation, migration and invasion. In conclusion, the present study successfully constructed and validated the AKGI as a potential prognostic biomarker for GC. The findings indicate that AKGI can identify patients likely to benefit from immunotherapy, enhance diagnostic precision and improve clinical outcomes in GC management. Moreover, AKGI offers a valuable framework for advancing the understanding of the role and mechanisms of α-KG in GC.