High accuracy differentiating autoimmune pancreatitis from pancreatic ductal adenocarcinoma by immunoglobulin G glycosylation

  • Hsi Chang Shih (Contributor)
  • Ming Chu Chang (Creator)
  • Chein Hung Chen (Contributor)
  • I-Lin Tsai (Contributor)
  • San-Yuan Wang (Contributor)
  • Ya Po Kuo (Contributor)
  • Chung Hsuan Chen (Contributor)
  • Yu Ting Chang (Creator)



Abstract Background Misdiagnosis of autoimmune pancreatitis (AIP) as pancreatic cancer (PDAC) or vice versa can cause dismal patentsâ outcomes. Changes in IgG glycosylation are associated with cancers and autoimmune diseases. This study investigated the IgG glycosylation profiles as diagnostic and prognostic biomarkers in PDAC and AIP. Methods Serum IgG-glycosylation profiles from 86 AIP patients, 115 PDAC patients, and 57 controls were analyzed using liquid chromatographyâ electrospray ionization mass spectrometry. Classification and regression tree (CART) analysis was applied to build a decision tree for discriminating PDAC from AIP. The result was validated in an independent cohort. Results Compared with AIP patients and controls, PDAC patients had significantly higher agalactosylation, lower fucosylation, and sialylation of IgG1, a higher agalactosylation ratio of IgG1 and a higher agalactosylation ratio of IgG2. AIP patients had significantly higher fucosylation of IgG1 and a higher sialylation ratio of IgG subclasses 1, 2 and 4. Using the CART analysis of agalactosylation and sialylation ratios in the IgG to discriminate AIP from PDAC, the diagnostic accuracy of the glycan markers was 93.8% with 94.6% sensitivity and 92.9% specificity. There were no statistically significant difference of IgG-glycosylation profiles between diffuse type and focal type AIP. Conclusions AIP and PDAC patients have distinct IgG-glycosylation profilings. IgG-glycosylation could different PDAC from AIP with high accuracy.