Prospective multicenter study on the usefulness of EUS-guided FNA biopsy for the diagnosis of autoimmune pancreatitis

Background and AimsIn the International Consensus Diagnostic Criteria (ICDC), autoimmune pancreatitis (AIP) is classified into types 1 and 2, and its definite histopathology diagnosis can be made based only on surgical or core biopsy specimens. Although EUS-guided FNA (EUS-FNA) biopsy is a safe technique for the collection of pancreatic tissue, no consensus viewpoint has been reached with regard to the role of EUS-FNA biopsy in the diagnosis of AIP. This study investigated the utility of pancreatic tissue collected by EUS-FNA biopsy by using a standard 22-gauge aspiration needle in the diagnosis of AIP.MethodsPatients with suspected AIP were prospectively enrolled in Nagoya University Hospital and Nagoya University–affiliated institutions. Pancreatic tissue was collected from each by EUS-FNA biopsy with a standard 22-gauge aspiration needle.ResultsFifty patients were registered, including 45 with a final diagnosis of AIP. Lymphoplasmacytic infiltration and abundant immunoglobulin G4–positive plasmacyte infiltration (>10/high-power field) were detected in 36 (72%) and 27 (54%) patients, respectively. Obliterative phlebitis and storiform fibrosis were not detected in our study. Granulocytic epithelial lesions (GEL) were observed in 3 patients. The sensitivity, specificity, positive predictive value, and negative predictive value of EUS-FNA biopsy to definitively diagnose AIP were 7.9% (3/38), 100% (12/12), 100% (3/3), and 25.5% (12/47), respectively. Pathology evaluation of pancreatic tissue collected by EUS-FNA biopsy improved the diagnostic accuracy in 8 (16%) of the 50 patients.ConclusionsIn this study, EUS-FNA biopsy by using a standard 22-gauge aspiration needle is not an effective diagnostic method for most patients with AIP. The combination of level 2 histology diagnosis of AIP with other findings specified in the ICDC slightly improved the diagnostic accuracy, although it still remains insufficiently accurate for routine clinical use.(Clinical trial registration number: 000006297.)