Abstract
This study introduces a novel prognostic model for glioma outcomes based on disulfidptosis, a unique programmed cell death pathway, highlighting its potential in cancer progression. Analyzing 15 disulfidptosis genes across various cancers, significant prognostic disparities were observed in GBMLGG, ACC, LIHC, KIRC, and others. A detailed investigation in GBMLGG utilized one training and two testing groups to identify seven target genes (ACTN4, IQGAP1, DSTN, MYH9, PDLIM1, FLNB, ACTB) using 101 machine learning approaches across 10 models. Their predictive accuracy for patient prognosis was confirmed through ROC and KM analyses across three datasets. A comprehensive nomogram prediction model incorporating clinical data was developed and independently validated. The study also explored correlations between target genes, immune cells, tumor mutational burden (TMB), and 19 immune checkpoints, uncovering significant associations with 13 cell lines and 19 immune checkpoint-related genes. Validation techniques included single-cell analysis, PCR, immunohistochemistry, and summary data-based Mendelian randomization. The research underscores the importance of disulfidptosis in cancer development and its strong correlation with GBMLGG prognosis, facilitated by immune microenvironment interactions and specific immune checkpoint inhibitors. This multifaceted analysis not only affirms disulfidptosis's role in prognostic predictions for GBMLGG but also its broader impact on cancer research, offering a deep dive into the molecular mechanisms of tumor heterogeneity and the immune landscape.