Sensitivity and Specifity of IgG4 Staining in Ampullary Mucosa for Autoimmune Pancreatitis
WM Yu, A Gupta, HE Remotti. Columbia Presbyterian Medical Center, New York, NY
Background: Autoimmune pancreatitis (AIP) is a type of chronic pancreatitis that may present as a mass lesion and mimic pancreatic carcinoma. Distinguishing AIP from neoplastic disease in the pancreas is critical to clinical management, since it is steroid responsive and usually does not require surgical intervention. AIP may be associated with increased numbers of IgG4-positive plasma cells in multiple organs. Studies have suggested immunostaining duodenal ampullary biopsies for IgG4 antibody as a useful method for diagnosing AIP. The presence of >10 IgG4-positive plasma cells per high power field (HPF) in a duodenal ampullary biopsy has been proposed by a group as a finding that supports the diagnosis of AIP. We examined duodenal ampullary mucosa in AIP and non-AIP specimens and investigated the feasibility of using the >10 IgG4-positive plasma cell/HPF cutoff in diagnosing AIP.
Design: Formalin-fixed, paraffin-embedded tissue sampling the ampullary region of Whipple resection specimens with autoimmune pancreatitis (3 cases) and control cases with adenocarcinoma (9 cases) were retrieved and were immunohistochemically stained with IgG4 antibody (Calbiochem 411492). For each case, IgG4-positive plasma cells/HPF were recorded in 10 random HPFs sampling superficial ampulla and duodenal papilla. The average number was computed for each case; sensitivity and specificity were calculated based on the averages.
Results: The average number of IgG4-positive cells in a high power field for the AIP cases range from 4.4 to 36.6, with a mean of 18.53 and median of 14.6. The average number of IgG4-positive cells in a high power field for the 9 non-AIP cases ranged from 1.3 to 16.4, with a mean of 5.38 and median of 2.6. Using the criterion that the presence of > 10 IgG-positive plasma cells /HPF supports AIP, our analysis reveals sensitivity = 67% and specificity =78%; positive predictive value =25%; negative predictive value= 88%.
|Avg # of IgG4 + cells/hpf||4.4||36.6||14.6|
|Range of IgG4 + cells/hpf||1-13||12-55||3-36|
Conclusions: The proposed threshold for diagnosing AIP (presence of >10 IgG4 positive plasma cells/HPF in duodenal ampullary mucosa) is not reliable for distinguishing between AIP and pancreatic adenocarcinoma. There is overlap in cases of AIP and pancreatic adenocarcinoma with false negative and false positive cases identified.
Category: Liver & Pancreas
Wednesday, March 11, 2009 9:30 AM
Poster Session V # 222, Wednesday Morning