Expression characteristics and clinical significance of PSCA in pancreatic cancer

To investigate PSCA expression in pancreatic cancer, evaluate its correlation with clinicopathological features, and explore its potential diagnostic and prognostic value. Gene expression and clinical data were obtained from TCGA‑PAAD and GEO databases (GSE16515, GSE28735, GSE71729). Bioinformatic analyses included differential expression, survival analysis, and GO/KEGG/GSEA enrichment. Immunohistochemistry (IHC) was performed in 20 pancreatic cancer tissues and 1 normal pancreatic tissue for preliminary validation. (1) PSCA mRNA is significantly upregulated in pancreatic adenocarcinoma (PAAD) tissues, and this finding is consistently validated in the TCGA-PAAD cohort and three independent GEO cohorts (GSE16515, GSE28735, GSE71729). ROC analysis demonstrates that PSCA has favorable diagnostic performance for PAAD, with an area under the ROC curve (AUC) of 0.895 in the TCGA cohort, and AUC values of 0.927, 0.771, and 0.862 in the three GEO validation cohorts, respectively. (2) Analysis of the TCGA cohort reveals that high PSCA expression is significantly associated with poorer overall survival (OS) in PAAD patients, and serves as an independent adverse prognostic factor for PAAD (HR = 1.18, 95% CI: 1.09–1.28, P < 0.001). Time-dependent ROC curves show that the AUC values for predicting 1-year, 3-year, and 5-year survival of patients are 0.642, 0.800, and 0.741, respectively. External validation using the TISIDB database further confirms that PAAD patients with high PSCA expression have significantly worse overall survival ( P = 0.00029), which is consistent with the results from the TCGA cohort. (3) In the TCGA cohort, clinicopathological correlation analysis stratified by PSCA expression shows a significant difference in AJCC stage distribution between the high and low PSCA expression groups ( P = 0.042), while no statistically significant differences are observed in age, gender, T stage, N stage, M stage, or chemotherapy status between the two groups (all P > 0.05). Validation from the TISIDB database indicates that PSCA expression level is significantly positively correlated with both tumor grade ( P = 0.00449) and AJCC clinical stage ( P = 0.00405) of PAAD — that is, higher PSCA expression corresponds to higher tumor grade and more advanced clinical stage. (4) Gene Set Enrichment Analysis (GSEA) shows that samples with high PSCA expression are enriched in pathways related to epidermal development and keratinization. Kyoto Encyclopedia of Genes and Genomes (KEGG) and Gene Ontology (GO) enrichment analyses suggest that PSCA may be involved in biological processes and signaling pathways including neuroactive ligand-receptor interaction, pancreatic secretion, and membrane potential regulation. (5) The expression distribution and intensity of PSCA protein are heterogeneous across different pathological subtypes of pancreatic cancer, and no PSCA expression is detected in pancreatic sarcomatoid carcinoma. In pancreatic ductal adenocarcinoma (PDAC), PSCA protein expression level is significantly correlated with tumor differentiation degree (MOD: P = 0.0048; PAR: P = 0.0059), and no PSCA expression is observed in normal pancreatic tissues. In the 20-case matched IHC cohort, multivariate analysis confirms that chemotherapy is an independent protective factor for patients’ overall survival (HR = 0.07, P = 0.032). In the 50-case clinical cohort, multivariate analysis verifies that N stage and tumor differentiation degree are independent prognostic factors affecting patients’ overall survival (all P < 0.05). Public transcriptomic datasets confirm that PSCA mRNA is significantly upregulated in tumor tissues, with favorable differential diagnostic efficacy and independent adverse prognostic value; its expression level is positively correlated with AJCC clinical stage and tumor grade. Validation in our institutional immunohistochemistry (IHC) cohort reveals that PSCA protein expression exhibits heterogeneity across pathological subtypes, and its expression intensity is negatively correlated with tumor differentiation degree in pancreatic ductal adenocarcinoma. Limited by the sample size, the independent prognostic value of PSCA at the protein level has not been verified. In summary, PSCA is a candidate biomarker with potential diagnostic and prognostic value for pancreatic adenocarcinoma, and further validation in larger-scale multicenter clinical cohorts and functional experiments is warranted.