Purpose: We conducted metabolomics and spatial cell transcriptomics of intraductal papillary mucinous neoplasms (IPMN), recognized pancreatic cancer precursors, to identify oncometabolites that inform upon risk of malignancy of IPMNs.
Experimental design: Untargeted metabolomic analyses were performed on cystic fluid from 125 patients with low-grade (LG) dysplasia or high-grade (HG) dysplasia with/without concurrent pancreatic ductal adenocarcinoma (PDAC; IPMN/PDAC). Predictive performance of individual metabolites for identifying HG or PDAC/IPMN was determined and compared with CA19-9 performance. Data were intersected with metabolic profiles of resected IPMN tissues and murine Kras;Gnas IPMN cell lines as well as spatial and single-cell transcriptomics of IPMNs.
Results: A total of 388 metabolites were quantified in cystic fluid, of which 69 were differential (P-value < 0.05) between cases (HG IPMN + IPMN/PDAC) and patients with LG IPMN. Spermidine and spermine biosynthesis and catabolism was identified as the top perturbed metabolic pathway (FDR-adjusted P-value < 0.0001). Increases in cystic fluid spermidine, n-acetylputrescine, acetylspermidine, diacetylspermidine, diacetylspermine, and acetylcadaverine were associated elevated risk of harboring HG or IPMN/PDAC. An OR rule comprising CA19-9, n-acetylputrescine, acetylspermidine, and diacetylspermine achieved 54.8% sensitivity for detecting HG IPMN and IPMN\PDAC. CA19-9 alone yielded sensitivity of 11.9% (McNemar Test P-value < 0.001). Polyamines were elevated in IPMN\PDAC tissues compared with LG IPMN tissues; spatial and single-cell transcriptomic data revealed transcript levels of polyamine-metabolizing enzymes to be elevated in neoplastic epithelium and tumor-associated macrophages.
Conclusions: Cystic fluid polyamines offer utility for determining risk of malignancy of IPMNs that is complementary to CA19-9 and that has potential to aid in clinical management of patients with IPMNs.
©2025 American Association for Cancer Research.