Machine Learning-Based WGCNA Approach for Developing an Immunogenic Cell Death-Related Hub Gene Signature and Identification of AJM1 as a Prognostic Biomarker in Pancreatic Adenocarcinoma

<b>Background & Aims:</b> Pancreatic adenocarcinoma (PAAD) remains a highly lethal malignancy with limited therapeutic options, primarily due to the absence of reliable prognostic biomarkers. Immunogenic cell death (ICD) plays a pivotal role in anti-tumor immunity and has potential as both a prognostic marker and a predictor of immunotherapy response. This study aimed to identify ICD-related hub genes and establish a robust prognostic gene signature for PAAD using weighted gene co-expression network analysis (WGCNA). <b>Methods & Results:</b> Transcriptomic and clinical data of PAAD patients were obtained from the TCGA and GEO databases. ICD enrichment scores were calculated using single-sample gene set enrichment analysis (ssGSEA), and ICD-associated gene modules were identified through WGCNA. A prognostic ICD-related gene signature was then constructed, and patients were stratified into high- and low-score groups based on the median risk score. Functional enrichment analysis was performed using the Molecular Signatures Database (MsigDB). Correlations between the signature score, immune cell infiltration, and drug sensitivity (IC₅₀ values from the GDSC2 database) were further assessed. Among the identified genes, AJM1 emerged as a key prognostic marker, validated in an independent PAAD cohort and through <i>in vitro</i> functional assays. <b>Conclusion:</b> This study developed and validated an ICD-related gene signature capable of predicting prognosis and immunotherapy responsiveness in PAAD. The identification and validation of AJM1 highlight its potential role as a prognostic biomarker and a novel contributor to the pathogenesis of PAAD.