Diagnosing autoimmune pancreatitis with the Unifying-Autoimmune-Pancreatitis-Criteria

Abstract

Background/Objectives

We had developed the Unifying-Autoimmune-Pancreatitis-Criteria (U-AIP) to diagnose autoimmune pancreatitis (AiP) within the M-ANNHEIM classification of chronic pancreatitis. In 2011, International-Consensus-Diagnostic-Criteria (ICDC) to diagnose AiP have been published. We had applied the U-AIP long before the ICDC were available. The aims of the study were, first, to describe patients with AiP diagnosed by the U-AIP; second, to compare diagnostic accuracies of the U-AIP and other diagnostic systems; third, to evaluate the clinical applicability of the U-AIP.

Methods

From 1998 until 2008, we identified patients with AiP using U-AIP, Japanese-, Korean-, Asian-, Mayo-HISORt-, Revised-Mayo-HISORt- and Italian-criteria. We retrospectively verified the diagnosis by ICDC and Revised-Japanese-2011-criteria, compared diagnostic accuracies of all systems and evaluated all criteria in consecutive patients with pancreatitis (2009 until 2010, Pancreas-Outpatient-Clinic-Cohort, n = 84). We retrospectively validated our diagnostic approach in consecutive patients with a pancreatic lesion requiring surgery (Surgical-Cohort, n = 98).

Results

Overall, we identified 21 patients with AiP. Unifying-Autoimmune-Pancreatitis-Criteria and ICDC presented the highest diagnostic accuracies (each 98.8%), highest Youden indices (each 0.95238), and highest proportions of diagnosed patients (each n = 20/21, U-AIP/ICDC vs. other diagnostic systems, p < 0.05, McNemar test). In the Pancreas-Outpatient-Clinic-Cohort, seven patients were diagnosed with AiP (n = 6 by U-AIP, n = 1 by Asian-criteria). International-Consensus-Diagnostic-Criteria confirmed the diagnosis in these individuals. Based on partial fulfillment of U-AIP, AiP was initially suspected in 13% (n = 10/77) of remaining patients from the Pancreas-Outpatient-Clinic-Cohort. In the Surgical-cohort, we identified one patient with AiP by U-AIP and ICDC.

Conclusions

Unifying-Autoimmune-Pancreatitis-Criteria revealed a satisfactory clinical applicability and offered an additional approach to diagnose AiP.

Introduction

Autoimmune pancreatitis (AiP) has been recognized as a rare form of pancreatic inflammation that represents the involvement of the pancreas in a systemic disorder [1], [2], [3]. The condition is characterized by typical features of clinical presentation, presence of serological markers, morphological appearance on cross-sectional imaging, characteristic imaging findings of the pancreatic-biliary system, and response to immunosuppressive medication [4], [5], [6], [7]. The disease may mimic pancreatic or biliary cancer, and its diagnosis often represents a clinical challenge [6], [8], [9]. Autoimmune pancreatitis has been recognized as a separate risk factor within the M-ANNHEIM multiple risk factor classification of chronic pancreatitis that has been developed in our institution [10].

In 2011, eastern and western experts agreed on International-Consensus-Diagnostic-Criteria (ICDC) for diagnosis of the disease [8]. Long before the international consensus was achieved, we had developed the Unifying-Autoimmune-Pancreatitis-Criteria (U-AIP) to diagnose AiP. The U-AIP were first published in 2009 [11]. At that period, the diagnosis of the disease rested on the Japan- [12], Korean- [13], Asian- [14], Mayo-HISORt- [4], Revised-Mayo-HISORt- [9], and Italian- [15] diagnostic criteria. These criteria represented the standard of care at that time. However, none of these systems comprised the entire spectrum of the disease, and some patients with AiP were left undiagnosed [16]. In an attempt to overcome these limitations, we combined key features of these available systems and developed the U-AIP-criteria (Fig. 1). Since 2009, the U-AIP have been slightly modified to their present form [17] and have been considered within the international consensus [8]. More recently, they were cited as the “German criteria” [16]. In clinical practice, we applied the U-AIP within the M-ANNHEIM classification system and we used the U-AIP together with all other available diagnostic systems until the ICDC were available.

Several studies have retrospectively tested the clinical use of the ICDC [16], [18], [19], [20]. These studies revealed that the ICDC currently represent the standard for the diagnosis of AiP and were more effective in diagnosing the disease than the previously published systems from throughout the world [4], [9], [12], [13], [14], [15]. Limitations of these previous criteria systems include restriction to IgG4-associated disease, refusal of possible diagnostic features such as presence of other autoimmune diseases or response to immunosuppressive medication, and requirement of a diagnostic endoscopic pancreatogram. The ICDC allow for the categorization of AiP into the major categories of “type 1” or “type 2”, with definite or probable diagnostic certainty, respectively. Patients with a more atypical presentation of the disease may meet the criteria of “AiP-not other specified” (AiP-NOS). Finally, the system offers the possibility to categorize forms of the disease that do not fulfill these criteria, including histologically unconfirmed but clinically suspected type 2, as ‘‘probable-AIP’’. However, several diagnostic features that were well-accepted in previous diagnostic systems, such as the presence of autoantibodies in AiP [6], [12], [13], [21] or the association of AiP with other autoimmune diseases [6], [12], [15], [21], [22], were not incorporated into the ICDC. The classification of patients according to the ICDC requires the detailed combination of various clinical features [8]. The resulting complexity of the ICDC may hinder their use in daily clinical practice [16], and has encouraged the publication of the Revised-Japanese-2011 diagnostic criteria [23], [24]. Recently, eastern and western experts agreed that further modifications of the ICDC might be required [19], [25]. Thus, the search for helpful diagnostic criteria continues.

A comparison of the diagnostic accuracy of the U-AIP with the ICDC and all other diagnostic systems has not yet been performed, and the clinical applicability of the U-AIP has not been evaluated. Thus, the aims of the present investigation were, first, to describe a cohort of patients with AiP that were diagnosed by the U-AIP; second, to compare the diagnostic accuracies of the U-AIP with all other available diagnostic criteria systems; and third, to evaluate the clinical applicability of the U-AIP.