Prognostic and microenvironmental analysis of pancreatic cancer by tumor location

Pancreatic cancer exhibits distinct biological and clinical features according to tumor location. Differences in survival and molecular programs among head (PH), body (PB), and tail (PT) tumors remain unclear.

Clinical data from the SEER database (n = 7302) and transcriptomic profiles from the TCGA cohort (n = 160) were analyzed. Patients were grouped by tumor site as PH, PB or PT. Overall survival was compared using Kaplan-Meier and Cox regression analyses, and location-specific nomograms were constructed. Differentially expressed genes (DEGs) were identified using edgeR, followed by GO, KEGG, and GSEA enrichment to explore molecular mechanisms.

Overall survival differed significantly by tumor site (P < 0.0001), with tail tumors showing the best prognosis, followed by PB, and PH tumors the poorest. Tumor location was an independent prognostic factor, and the nomograms demonstrated good predictive performance. Transcriptomic analysis revealed enrichment of lipid and steroid metabolism in tail versus FH tumors, and neuroactive ligand-receptor and endocannabinoid signaling in body-tail comparisons. GSEA identified apical junction, epithelial-mesenchymal transition (EMT), and hypoxia as key pathways associated with location-specific outcomes.

Integrating large-scale epidemiologic and transcriptomic data, this study demonstrates that pancreatic tumor location independently influences survival and reflects distinct metabolic and microenvironmental programs. Recognition of these differences may aid individualized prognostic assessment and guide tailored therapeutic strategies.