LAUSR

To develop a superior diagnostic approach for pancreatic adenocarcinoma (PAAC), the present study prospectively included 338 PAAC patients, 294 normal healthy volunteers (NHV), 122 chronic pancreatitis (CP) patients and 100 patients with non‐PAAC malignancies. In the identification phase, HuProt Human Proteome Microarray, comprising 21 065 proteins, was used to identify serum tumor‐associated autoantibodies (TAAbs) candidates differentiating PAAC (n = 30) from NHV (n = 30). A PAAC‐focused array containing 165 differentially expressed TAAbs identified was subsequently adopted in the validation phase (n = 712) for specificity and sensitivities. The multivariate TAAbs signature for differentiation PAAC from controls (NHV + CP) identified five candidates, namely the IgG‐type TAAbs against CLDN17, KCNN3, SLAMF7, SLC22A11 and OR51F2. Multivariate logistic performance model of y = (22.893 × CA19‐9 + 0.68 × CLDN17 − 4.012) showed a significant better diagnostic accuracy than that of CA19‐9 and CLDN17 in differentiating PAAC from controls (NHV + CP) (AUC = 0.97, 0.92 and 0.82, respectively, P‐value < .0001). We further tested the autoantigen level of CLDN17 by ELISA in 82 sera samples from PAAC (n = 42), CP (n = 24) and NHV (n = 16). Similarly, the model showed superior diagnostic performance than that of CA19‐9 and CLDN17 (AUC = 0.93, 0.83 and 0.81, respectively, P‐value < .0001) in differentiating PAAC from controls. In conclusion, our study is the first to characterize the circulating TAAbs signatures in PAAC. The results showed that CLDN17 combined with CA19‐9 provided potentially clinical value and may serve as noninvasive novel biomarkers for PAAC diagnosis.