Introduction

Since Yoshida et al. [1], first proposed the concept of autoimmune pancreatitis (AIP) in 1995, AIP has been accepted worldwide as a distinctive type of pancreatitis [16]. Due to the increasing numbers of cases, several issues in the management of AIP were raised in Japan. These issues are (1) diagnosis and management of atypical or indeterminate AIP, (2) differentiation from pancreas cancer, (3) evaluation of OOIs, (4) diagnosis and treatment of recurrent cases, and (5) different diagnostic criteria in Japan and other countries [4, 5]. To resolve these issues, the Japan Pancreas Society (JPS) and the Research Committee for Intractable Pancreatic Disease supported by the Ministry of Health, Labour and Welfare of Japan (RCIPD-MHLWJ), proposed the Japanese consensus guidelines for the management of AIP in 2009 [6]. In 2011, the International Consensus Diagnostic Criteria for AIP (ICDC) [9] were proposed. The ICDC proposed two subtypes, type 1 AIP, which is associated with IgG4, and type 2 AIP, which is associated with granulocytic epithelial lesion (GEL). Lymphoplasmacytic sclerosing pancreatitis (LPSP) is a pancreatic manifestation of IgG4-related disease (IgG4-RD) characterized by increased serum IgG4 and abundant infiltration of IgG4-positive plasmacytes, obliterative phlebitis and storiform fibrosis. In Japan, LPSP is more often observed, whereas idiopathic duct-centric chronic pancreatitis (IDCP) characterized by GEL is rare [48]. One of the major differences between the 2002 and 2006 Japanese criteria and the ICDC is in the therapeutic use of steroids. The previous Japanese criteria [35] did not recommend facile therapeutic use of steroids. The revised version of the JPS criteria (JPS-2011) for type 1 AIP [10, 11] was proposed in response to the ICDC’s inclusion of response to steroid treatment. The number of publications on AIP increased from 871 to 1,843 between 2008 and 2012 (in the PubMed database). In light of this additional research, the Japanese consensus guidelines need to be revised. Most of the evidence levels of the specific clinical statements and a secondary database were still lower than grade III as proposed by the Agency for Health Care Policy and Research in 1993. Therefore, we have developed the revised version of the consensus guidelines using the modified Delphi approach [58, 12]. Briefly, to establish consensus, three committees (the professional committee for making clinical questions and statements by Japanese specialists for AIP, the expert panel committee for rating statements using the modified Delphi method, and the evaluating committee comprised of moderators) were organized. During the first phase, 15 specialists (11 pancreatologists, 2 radiologists, 1 respiratory system expert and 1 pathologist) were selected from the members of the RCIPD-MHLWJ. These specialists revised the 36 clinical questions (CQs) and statements for (1) concept and diagnosis (13 CQs), (2) extra-pancreatic lesions (6 CQs), (3) differential diagnosis (6 CQs) and (4) treatment (11 CQs) based on the selected papers [68], which focus on the concept and diagnosis CQs.

The expert panelists (ten pancreatologists) individually rated the clinical statements for appropriateness, and discussed areas of disagreement and uncertainty [58, 12]. Ratings of appropriate methods for the management of AIP were developed using a modified Delphi approach. Rating was on a nine-point scale, with one being highly inappropriate and nine being highly appropriate. A clinical statement receiving a median score greater than seven was regarded as valid. The specialists revised some of the clinical statements after discussion with expert panelists, and then the revised clinical statements were rated again. Based on the two-round modified Delphi approach, guideline statements for diagnosis and management of AIP were developed. In addition to the specialist and expert panels, the moderators included one pancreatologist, one surgeon, one pathologist and one internist, each of whom were also familiar with epidemiology and the modified Delphi approach [58]. The moderators reviewed the literature, collected clinical statements from the literature as well as from a survey of the professionals, facilitated the panelist meetings, and analyzed the data. Because available clinical evidence regarding the diagnosis and management of AIP is limited, we could not set a suitable recommendation level for some clinical statements. In the revised consensus-based guidelines, the statements for clinical practice were evaluated as “strongly recommendable” (level A) or “strongly unrecommendable (level D)” for receiving a score of nine, and “ordinarily recommendable” (level B), “unrecommendable” (level C), or “conflicting benefits and harms” (level I) for that less than nine according to the grading proposed by United States Preventive Services Task Force [13].

Clinical questions and statements

I. Concept and Diagnosis

CQ-I-1. What is “autoimmune pancreatitis (AIP)”?

  • AIP is a distinct form of pancreatitis clinically characterized by frequent presentation with obstructive jaundice with or without a pancreatic mass, histologically by a lymphoplasmacytic infiltrate and fibrosis and therapeutically by a dramatic response to steroids.

  • AIP is classified as two subtypes, type 1 and type 2. Type 1 AIP is more prevalent in Japan; references to AIP in Japanese literature usually mean type 1 AIP.

  • Type 1 AIP is lymphoplasmacytic sclerosing pancreatitis (LPSP) characterized by massive infiltration of lymphocytes and plasmacytes, especially IgG4-positive plasmacyte; storiform fibrosis; and obliterative phlebitis. It is a pancreatic manifestation of a systemic disorder, IgG4-related disease (IgG4-RD).

  • Type 2 AIP (also called IDCP) or AIP with GEL, is more commonly observed in Europe and the United States. Type 2 AIP exhibits neutrophilic lesions and, therefore, is a different condition from type 1 AIP.

Description

AIP is a distinct form of pancreatitis characterized clinically by frequent presentation with obstructive jaundice with or without a pancreatic mass, histologically by a lymphoplasmacytic infiltrate and fibrosis and therapeutically by a dramatic response to steroids.

The original concept of AIP was proposed in Japan [1] and was defined as a pancreatitis whose pathogenesis could possibly involve autoimmune mechanisms [18]. Autoimmune mechanisms were suspected due to characteristic findings, such as hypergammaglobulinemia, increased serum levels of IgG or IgG4, presence of autoantibodies, and effective response to steroid therapy. Patients with AIP occasionally exhibit other organ involvement (OOI) such as sclerosing cholangitis, sclerosing sialadenitis, retroperitoneal fibrosis, enlarged coeliac and hilar lymph nodes, chronic thyroiditis, or interstitial nephritis. All of them show similar pathological findings with abundant infiltration of IgG4-positive cells as well as high serum IgG4, which support the possibility that AIP is a systemic disorder associated with pancreatic lesions. After several proposals of nomenclatures such as the IgG4-related systemic sclerosing disease [14], systemic IgG4-related plasmacytic syndrome (SIPS) [15] and IgG4-positive multi-organ lymphoproliferative syndrome (IgG4 MOLPS), [16] the consensus nomenclature of IgG4-related disease (IgG4-RD) [1720] was proposed. Therefore, AIP related to IgG4 is now regarded as a pancreatic manifestation of IgG4-RD. Histopathological findings show lymphoplasmacytic sclerosing pancreatitis (LPSP) [21] characterized by (1) massive infiltration of lymphocytes and plasmacytes, (2) especially IgG4-positive plasmacytes more than 10 cells/high power field (400×), (3) storiform fibrosis, and (4) obliterative phlebitis. It is commonly seen in elderly males, and is comparable to lymphoplasmacytic sclerosing pancreatitis (LPSP), which is characterized by histopathological findings of pronounced infiltration of lymphocytes and plasmacytes, infiltration of IgG4-positive plasmacytes, storiform fibrosis, and obstructive phlebitis.

Cases associated with ulcerative colitis in young patients, mainly reported in Europe and the US, show typical pathological neutrophilic lesions called IDCP [22] or AIP with GEL [23, 24]. In addition to histopathological findings, no hematological markers suggest that their pathological conditions are different from LPSP [11]. Although typical pancreatic images in both LPSP and IDCP show diffuse swelling of more than one-third of the pancreas, some atypical and indeterminate cases of segmental/focal swelling or mass-forming type [911] are necessary to be differentiated from pancreatic cancer. Based on these findings, the recent international consensus diagnostic criteria (ICDC) for AIP [9] proposed the current concept of AIP and classification as two subtypes, type 1 (LPSP) and type 2 AIP (IDCP). As most of AIP are type 1 AIP in Japan [10, 11], use of the simple term of AIP usually means type 1 AIP in the present guidelines (Table 1).

Table 1 Recommendation levels based on consensus

CQ-I-2. Are there characteristic clinical symptoms in AIP?

  • There are no specific symptoms seen in patients with AIP. However, in many cases, patients with type 1 AIP show; obstructive jaundice; symptoms of diabetes mellitus; accompanying extra-pancreatic lesions, and minor to no abdominal pain. Those with type 2 AIP commonly have abdominal pain and acute pancreatitis.

Description

Patients with type 1 AIP do not show the type of severe abdominal pain seen in those with acute pancreatitis or with acute exacerbation of chronic pancreatitis. Abdominal pain is mild to nonexistent, [5, 19, 2528] although there have been a few cases reported where the disease started as acute pancreatitis or severe pancreatitis [29, 30]. One-third to one-half of patients show obstructive jaundice or mild abdominal pain, and 15 percent have shown back pain or weight loss [31, 32]. More than half of cases are associated with sclerosing cholangitis, diabetes mellitus, sclerosing sialoadenitis/dacryoadenitis, or retroperitoneal fibrosis, showing obstructive jaundice, polydipsia,/polyuria or malaise, xerostomia/xerophthalmia, or hydronephrosis, respectively [31]. Those with type 2 AIP commonly have abdominal pain and acute pancreatitis [9] (Table 2).

Table 2 Clinical symptoms in AIP

CQ-I-3. How is AIP found?

  • In many cases, patients go to see doctors with complaints such as minor abdominal pain, general malaise, jaundice, or dry mouth (Level of recommendation: B).

  • In many cases, AIP is found when patients who have increased levels of biliary enzymes, obstructive jaundice, or diabetes mellitus are tested for pancreatic or biliary duct cancers in a differential diagnosis (Level of recommendation: B).

  • In many cases, an enlarged pancreas demonstrated by abdominal ultrasonography leads to the detection of AIP (Level of recommendation: B).

Description

In more than half of AIP cases, patients visit the hospital for symptoms such as minor abdominal pain, general malaise, jaundice, or dry mouth [1, 2532]. A urine test or general blood biochemical test shows abnormal levels of pancreatic or biliary enzymes. In other cases, an increased level of CA19-9 is observed; pancreatic imaging tests such as abdominal ultrasound, CT or MRI show a diffusely or locally enlarged pancreas, or a pancreatic mass may also be found. In many cases the disease is found in the course of a differential diagnosis against pancreatic or biliary cancers [1, 611, 2532]. AIP is also found during the close examination of extra-pancreatic lesions; e.g., during the differential diagnosis against primary sclerosing cholangitis (PSC); in examination of suspected Sjögren’s syndrome by a head/neck-otolaryngologist, ophthalmologist, or collagen disease-rheumatologist; or in examination for retroperitoneal fibrosis by an urologist. The rate of association with other autoimmune diseases is not clear. There have been reports, mainly in Europe and the United States, of cases associated with juvenile ulcerative colitis showing evidence of IDCP [9, 22] or GEL [23, 24]. Conversely, cases associated with ulcerative colitis or primary biliary cirrhosis are rarely seen in Japan [68].

CQ-I-4. What are the characteristic blood-biochemical and immunological findings in AIP?

  • Although there are no disease-specific serum biochemical findings, increased serum levels of pancreatic enzymes, biliary enzymes and total bilirubin are commonly observed in AIP (Level of recommendation: A).

  • Serum levels of IgG4 have the highest diagnostic value as a single serological diagnostic method among all the available ones; however, it is not disease specific (Level of recommendation: A).

  • High serum IgG or the presence of non-specific antibodies such as antinuclear antibodies or rheumatoid factor suggest the possibility of AIP (Level of recommendation: B).

Description

Most cases of AIP are discovered when patients show increased levels of biliary enzymes, obstructive jaundice, or diabetes mellitus, which are usually reflected in biochemical tests. Abnormal biliary findings are seen in many cases; 60–82 % of cases exhibit an increase of biliary enzymes: and 39–62 % of cases exhibit an increase of total bilirubin [3235]. Compared to cases of acute pancreatitis or acute exacerbation of chronic pancreatitis, the occurrence rate of abnormal levels of serum pancreatic enzymes is lower, 36–64 % [32, 33], and the levels rarely become abnormally high. There have been reports of increased levels of peripheral eosinophil granulocytes [32] and activated T-lymphocytes (CD4-positive, CD8-positive) [33].

Immunological examinations show high incidences of hypergammaglobulinemia (43 %), increased levels of serum IgG (62–80 %), increased levels of serum IgG4 (68–92 %) [3235] ), antinuclear antibodies (40–64 %), rheumatoid factor (25 %) [32, 33] and Th2 predominance over Th1 in the local lesions [36, 37]. However, these results are not disease-specific. Some reports have shown the presence of autoantibodies, such as anti-carbonic anhydrase II antibodies (55 %) or anti-lactoferrin antibodies (75 %) in patients with AIP, although they generally cannot be tested [32, 33]. Anti-SSA/B antibodies or anti-mitochondrial antibodies, on the other hand, are rarely seen [32, 33]. Among all serological diagnostic methods, an increased level of serum IgG4 has the highest diagnostic value as a single method because of its sensitivity (80 %) and its specificity (98 %) in differentiating AIP from pancreatic cancer; however, it is not disease specific. The sensitivity and specificity of serum IgG are 70 and 75 %, respectively, and the positive ratios of antinuclear antibodies and rheumatoid factor are 60 and 20–30 %, respectively. Even when IgG is combined with antinuclear antibodies or rheumatoid factor, the sensitivity is 91 % but the specificity is 61 %; the specificity is lower than that for IgG4, although the sensitivity is equivalent to that for IgG4 [34]. (Refer to CQ-III-2).

CQ-I-5. Are there pancreatic exocrine and endocrine dysfunctions?

  • AIP is often associated with pancreatic exocrine and endocrine dysfunctions (e.g., diabetes mellitus). Occurrence ratios are about 80 and 70 % for exocrine and endocrine dysfunctions, respectively, (Level of recommendation: A).

Description

AIP is in many cases associated with pancreatic exocrine and endocrine dysfunction (e.g., diabetes mellitus.) According to the fact-finding survey conducted in 2000 by the Ministry of Health and Welfare Investigation Research Team for Special Intractable Pancreas Disease, 80.6 % of the cases studied showed abnormal pancreatic exocrine function [in which the abnormality is defined as 70 % or lower secretion in the BT-PABA (PFD test)], and 70.0 % of the cases showed exocrine dysfunction (as determined by the secretin test), comparable to that of confirmed cases of chronic pancreatitis. Additionally, 77.0 % of the cases were reported to be associated with diabetes mellitus [31]. Studies by individual medical facilities have reported that 83–88 % of the cases were associated with secretion dysfunction, and 42–78 % with diabetes mellitus [3840]. The diabetes mellitus accompanying AIP was analyzed in detail in a national fact-finding survey conducted in 2006 [35]. Among those AIP patients who sought medical treatment in 2002, 66.5 % of cases were found to have associated diabetes mellitus; Of those, 33.3 % had diabetes mellitus prior to the onset of AIP, and 51.6 % started developing diabetes mellitus around the same time as the onset of pancreatitis. Among those patients having diabetes mellitus, 14 % developed diabetes after steroid treatment [35], suggesting that such diabetes may be caused by long-term steroid treatment. There are some cases where pancreatic endocrine dysfunction was improved by steroid treatment; however, since not all cases improved, it can be stated that medical conditions that have progressed far enough to cause some degree of organic change can not be reversed. (Refer to CQ-IV-9.)

In AIP, the mechanism of pathogenesis of pancreatic exocrine dysfunction is assumed to involve the following: decreased secretion of pancreatic enzymes associated with collapsed acinar cells caused by pronounced cellular infiltration mainly of plasmacytes and fibrosis; and obstructed flow of pancreatic juice due to inflammatory cell infiltration around the pancreatic ducts and subsequent narrowing of pancreatic ducts [35, 3941]. A recent study suggested that mislocalization of the cystic fibrosis transmembrane conductance regulator (CFTR), which plays a central role in pancreatic duct HCO3 secretion, and up-regulation of aquaporin-1 (AQP1) on the plasma membrane and in the cytoplasm of pancreatic duct cells may be involved in the development of AIP [42]. Corticosteroids reduce inflammation and restore both digestive enzyme and HCO3 secretion in patients with AIP by regenerating acinar cells and correcting CFTR localization in pancreatic duct cells [42]. In contrast, the mechanism of pathogenesis of diabetes mellitus is assumed to be affected by both obstructed blood flow of endocrine glands (islets of Langerhans) associated with the fibrosis of exocrine glands, and damaged islets of Langerhans due to the spreading of inflammation [40, 41].

CQ-I-6. What are the characteristic findings of abdominal ultrasonography in AIP ?

  • Abdominal ultrasonography is effective for the diagnosis of AIP. Ultrasonic findings in patients with AIP are characterized by a diffusely or locally enlarged pancreas with low echo; A diffusely enlarged pancreas is called a “sausage-like” pancreas. (Level of recommendation: A).

Description

The Japanese Clinical Diagnostic Criteria of Autoimmune Pancreatitis [35, 10, 11] states that a “diffusely or locally enlarged pancreas is detected by an abdominal ultrasound test, an abdominal X-ray test, or an abdominal MRI test.” Ultrasonography is the initial clinical test performed, and serves as a tool to diagnose AIP. However, in some cases, patients are found to have AIP during physical examinations [43].

A diffusely enlarged pancreas appears as a low-echo area in general (Fig. 1a) and has a so-called “sausage-like” appearance [44]. No dilatation of the main pancreatic duct is seen in most cases. The enlarged area shows a low echo image, in some cases with scattered high echo spots [44, 45]. A segmentally/focally enlarged pancreas must be distinguished from pancreatic cancer or mass-forming pancreatitis through a differential diagnosis. Although dilatation of the main pancreatic duct is not seen in most cases, some patients may show minor dilation, which makes the differential diagnosis difficult. Conversely, if the main duct is found to penetrate through the mass (Fig. 1b), duct penetration may be a useful sign to rule out pancreatic cancer [46]. In some cases, there may be many low echo mass images in the pancreatic parenchyma (Fig. 1c), which makes it difficult to differentiate AIP from malignant lymphoma or metastatic pancreatic tumors.

Fig. 1
figure 1

Abdominal ultrasonography in AIP. a Diffuse type: a diffusely enlarged pancreas appears as a low-echo area with high echoic spots and has a so-called “sausage-like” appearance. b Tumor forming type: The main duct (arrows) is found to penetrate through the mass (duct-penetrating sign) in the case of a locally enlarged pancreas with a clear margin. c Multiple mass forming type: low echoic masses are observed in the pancreas head and body (arrows) and normal parenchyma is seen in the interstitial segment of these masses. d IgG4-related sclerosing cholangitis: moderately homogenous low-echoic wall thickness is observed in the upper and middle bile duct

It has been reported that around 60 % of patients with AIP show a thickened bile duct wall [45] (Fig. 1d). A thickened bile duct wall is characterized by layered or parenchymal low echo wall thickening [47, 48]. There have been some cases where the thick wall centered around the extrahepatic bile duct extends over to the intrahepatic bile duct or gallbladder [47]. The wall thickening has been studied in detail with intraductal ultrasonography (IDUS) [45, 48]. Although wall thickening of narrowed areas is not clearly visible in US, since areas other than the narrowed area show thickening of the internal low-echo layer while maintaining the high-echo image for the outer area, it is assumed that the thickening is happening on the bile duct wall itself [49].

Some recent reports have discussed the usefulness of contrast-enhanced ultrasonography in the diagnosis to differentiate AIP from pancreatic cancer [5052]. Reports have shown that while in the case of pancreatic cancer, only the rim of the mass was stained with the presence of tumor vessels; in the case of AIP, the entire mass was stained without presence of tumor vessels. However, reports have also shown that for AIP, findings varied depending on the stage of the disease. The areas of stronger inflammation and immature fibrosis were stained strongly, whereas the areas of weaker inflammation and older fibrosis were stained weakly [52].

CQ-I-7. What are the characteristic findings of abdominal computed tomography (CT)?

  • Abdominal CT images of patients with AIP show a diffusely or locally enlarged pancreas. The dynamic CT shows a distinctive delayed enhancement pattern with various images depending on the activity or stage of the disease. (Level of recommendation: A).

  • If a capsule-like rim is observed, the patient is highly suspected of having AIP. (Level of recommendation: A).

Description

Typical AIP exhibits a diffusely enlarged pancreas [3]. The pancreatic parenchyma is replaced by fibrosis, which causes a reduced enhancement effect during the “pancreatic parenchymal phase” and shows less absorption compared to a normal pancreas. Due to the delayed enhancement in the area of fibrosis, a certain level of enhancement is seen in the “portal phase” and the enhancement continues into the “delayed phase” where the enhancement becomes stronger than in a normal pancreas. Consequently, the dynamic CT enhancement pattern of AIP shows a slow and delayed enhancement pattern. However, because weak fibrosis shows an enhancement pattern similar to normal pancreas, even in the absence of delayed enhancement, the possibility of AIP cannot be ruled out [44].

A “capsule-like rim” is a relatively distinctive CT feature of AIP (Fig. 2a, b) [44, 53]. It is a band-like structure that appears to surround all or part of the lesions; it shows lower absorption than pancreatic parenchyma of the lesion during the pancreatic parenchymal phase, and a delayed enhancement pattern with dynamic CT [44, 53]. While these findings may indicate fibrosis of the rim of the lesion, the frequency of such findings varies depending on the report [53, 54]. This finding, however, is specific to AIP and is not seen in any other diseases. If a capsule-like rim is observed, the chance of the patient having AIP is high. A locally enlarged pancreas is an especially useful sign to distinguish AIP from pancreatic cancer [44, 53]. (Refer to CQ-III-3) Although an actual positive rate of capsule-like rim in AIP on CT images remains unclear because of few reports, a recent single study suggested 48 % [55] .

Fig. 2
figure 2

Abdominal CT in AIP. a Parenchymal phase: The pancreatic parenchyma shows irregularly reduced enhancement. The marginal zone of the pancreas shows capsule-like rim with more reduced enhancement (arrows). b Delayed phase: the enhancement in the parenchyma and capsule-like rim become stronger in the delayed phase.(arrows)

AIP exhibits many different CT images. Many AIP patients are elderly. Because their pancreases are atrophied to begin with, a pancreas enlarged from AIP is not clearly seen. In some cases, the pancreatic enlargement is verified only after steroid treatment by comparing the size before and after the treatment. There are cases where no abnormality other than a minor diffusely enlarged pancreas is found, partial dilatation of the main duct is pronounced, cystic lesions that appear to be pseudocysts are involved, or the pancreatic parenchyma shows obvious calcification. It must be realized that the absence of typical CT images cannot be the reason to exclude AIP from consideration [44, 53]. A recent study reported that pancreatic volumetric perfusion was attenuated in AIP patients, and improved after the steroid treatment. This suggested that the perfusion CT may be useful in evaluating therapeutic effects [56].

CQ-I-8. Can magnetic resonance cholangiopancreatography (MRCP) evaluate narrowing of the main pancreatic duct in AIP?

  • MRI images of AIP show a diffusely enlarged pancreas with distinctive characteristics, such as a low signal on T1-weighted images and a delayed enhancement pattern on dynamic MRI images. (Level of recommendation: A).

  • A “capsule-like rim” reflects strong fibrosis of the peripancreatic lesion, which is highly specific for AIP. (Level of recommendation: A).

  • At this moment, MRCP is not recommended for the accurate evaluation of the narrowing of the main pancreatic duct. (Level of recommendation: B).

Description

Like other image examinations, MR images show a diffusely or locally enlarged pancreas in cases of AIP [3]. The basic MR images used to examine AIP are T1-weighted images, T2-weighted images, and dynamic MRI; AIP lesions show a low signal on T1-weighted images. A normal pancreas shows a higher signal than the liver on T1-weighted images; therefore, a pancreas showing a lower signal than the liver is judged to be abnormal. However, since a low signal is also seen in pancreatic cancer or normal chronic pancreatitis, it is not a characteristic finding of AIP [44, 53]. The T2-weighted images may show a slightly lower signal in strong fibrosis and a slightly stronger signal in weak fibrosis [44, 53]. Meanwhile, the dynamic MR image shows a delayed enhancement pattern, as seen in the dynamic CT [44, 53]. (Refer to CQ-I-7).

Because a capsule-like rim is observed in MRI in about 36 % of patients with AIP [57], its presence can be used as a supplementary diagnostic tool for the disease. The capsule-like rim is extracted as a low signal on T2-weighted images reflecting strong fibrosis. Dynamic MR images show a delayed enhancement pattern [53, 54].

Although MRCP is useful in evaluating the narrowing of the main pancreatic duct in cases of multiple stenosis (skipped lesions), unremarkably dilated upper stream, or steroidal improvement, it is currently difficult to use MRCP images for the diagnosis of AIP [3]. However, recent significant progress in MRI technology has made it possible to extract images of the normal main pancreatic duct by 3-D MRCP without fail. Therefore, if the main pancreatic duct is not extracted by 3-D MRCP, it may be an indication of prominent stenosis. Since further image quality improvement can be expected for MRCP with the introduction of 3-Stela MRI technology, it is possible that MRCP will be used to evaluate the therapeutic effect or monitor the progress of AIP in the future [44, 53]. Recently, diffusion-weighted MRI (DWI) has been reported to be useful for detecting AIP by distinguishing it from pancreatic cancer, and for evaluating the effect of steroid therapy [5860].

CQ-I-9. What are the characteristic findings of positron emission tomography (PET) and gallium-scintigram in AIP?

  • Patients with AIP show accumulation of gallium citrate (Ga-67) and fluorine-18 fluorodeoxyglucose (FDG) in pancreatic and extra-pancreatic lesions, which disappears shortly after steroid treatment. The characteristic accumulation pattern and kinetics following the steroid treatment can be used for the diagnosis of the disease. (Level of recommendation: B).

Description

Gallium scintigraphy shows accumulation of Ga-67 in localized pancreatic lesions in patients with AIP. Previously, some such cases were diagnosed as pancreatic malignant lymphoma [61]. The accumulation of Ga-67 is found not only in pancreatic lesions but also in extra-pancreatic lesions such as in the hilar lymph nodes, lacrimal gland, or salivary gland. The accumulation is found in about 70 % of pancreatic lesions and hilar lymph nodes, and about 20 % of lacrimal/salivary glands. The accumulation reflects high disease activity and disappears quickly after steroid treatment [62]. Therefore, the distribution of Ga-67 accumulation and the kinetics after steroid treatment can be used for the diagnosis of AIP.

FDG-PET (fluorine-18 fluorodeoxyglucose positron emission tomography) is useful for the diagnosis of pancreatic cancer. However, high accumulation of FDG (90 % or higher) is also observed in patients with AIP. In these cases, the accumulation corresponds to the prominent inflammatory cell infiltration areas [6366]. FDG also accumulates in extra-pancreatic lesions such as in the salivary gland, a wide range of lymph node lesions, retroperitoneal fibrosis, and the prostate gland [6769]. Accumulated FDG in pancreatic or extra-pancreatic areas disappears quickly after steroid treatment [66]. The following two criteria are useful in distinguishing AIP from pancreatic cancer: extensive or multiple accumulations of FDG in the pancreas, or distinctive accumulation in extra-pancreatic lesions in the salivary gland, retroperitoneal fibrosis, or prostate gland [66, 67]. Although the disappearance of FDG following steroid treatment is reported to be used as a criterion to distinguish AIP from pancreatic cancer [70], a facile steroid trial should be carefully performed after a negative work-up of malignancy, because a follow-up study of FDG-PET is not supported by Japanese medical insurance.

CQ-I-10. What are the characteristic findings of endoscopic retrograde cholangiopancreatography (ERCP) in AIP?

  • ERCP shows narrowing of the main pancreatic duct characteristic of AIP. (Level of recommendation: A).

  • AIP may be associated with stenosis of the bile duct. (Level of recommendation: A).

Description

ERCP shows narrowing of the main pancreatic duct, which is characteristic of AIP. This finding is used as the basis for diagnosis [1]. Narrowing of the pancreatic duct is usually diagnosed from ERCP images. The narrowing of the pancreatic duct is “unlike the obstruction or stenosis, as the narrowing extends to certain degree and the duct diameter is smaller (narrower) than normal, with some irregularities” [1, 71] (Fig. 3a).

Fig. 3
figure 3

Pancreatogram in AIP. a Diffuse type: diffusely irregular narrowing of the main pancreatic duct is seen from the pancreas head to tail. b Segmental type: irregular narrowing of the main pancreatic duct is seen from the pancreas body to tail. c Focal type: irregular narrowing of the main pancreatic duct is seen in the pancreas head without dilation of upperstream. d Multiple narrowing type: discontinuous and multiple narrowing of the main pancreatic duct (skip lesions) are seen in the pancreatic head and body

The Clinical Diagnostic Criteria of Autoimmune Pancreatitis 2011 (JPS-2011) states that diagnosis of the disease requires pancreatic images showing “the distinctive narrowing of the main pancreatic duct”, where the narrowing may be diffuse or local. The range of narrowing varies. The typical case exhibits narrowing over one third of the entire pancreatic duct (Fig. 3b). Even when the narrowing is localized to less than one third of the entire duct, in most cases no significant dilatation is observed above the narrowed area upstream of the main duct [7275] (Fig. 3c).

There are, however, other cases where the narrowing is localized to less than one third, or the lesions are multiple-stenotic (skip lesions) (Fig. 3d) [73, 74]. If the narrowing is localized, it is necessary to consider differentiating the disease from pancreatic cancer [4, 75, 76]. Typical pancreatic duct features of AIP visible in ERCP images, such as side branch arising from narrowed portion or multiple stenosis of the main pancreatic duct, are useful for differential diagnosis from pancreatic cancer. Short narrowing images of the main pancreatic duct less than 3 cm are impossible [4, 75, 76].

About 80 % of patients with AIP show stenosis of the bile duct [7781]. Although most of the stenosis is found in the lower bile duct, it can also be detected in the extra-hepatic or intra-hepatic bile ducts [7781].

CQ-I-11. What are the characteristic histopathological findings in AIP?

  • As histopathological findings of type 1 AIP are characteristic, typical ones can be diagnosed by themselves without clinical informations (Level of recommendation: A).

  • Histopathological findings of AIP are characterized by fibrosis with strong lymphoplasmacytic infiltration, abundant infiltration of IgG4-positive plasmacytes, storiform fibrosis, obliterative phlebitis, and periductal inflammation of inflammatory cells. (Level of recommendation: A).

  • Histopathological findings of type 2 AIP are characterized by infiltration of granulocytes into the epithelium and lumen of the interlobular pancreatic duct. Infiltration of IgG4-positive plasmacytes is rarely seen (Level of recommendation: A).

Description

As histopathological findings of type 1 AIP called “lymphoplasmacytic sclerosing pancreatitis (LPSP) [21]” are characteristic, typical ones can be diagnosed by themselves without clinical information. The major histological findings are (i) prominent infiltration of lymphoplasmacytes without granulocytic infiltration and fibrosis (Fig. 4a); (ii) abundant (> 10 cells/HPF) IgG4-positive cells (Fig. 4b); (iii) storiform fibrosis (Fig. 4c); (iv) obliterative phlebitis (Fig. 4d); and (v) periductal infiltration of inflammatory cells (Fig. 4e); and the fibrosis associated with prominent infiltration of lymphocytes and plasmacytes [611, 8286]. When three of the first four (i–iv) histologic items fulfil the Japanese clinical diagnostic criteria (JPS-2011) [10, 11], or three of the last four (ii–v) fulfil the ICDC [9], definitive diagnosis can be made based on histology alone.

Fig. 4
figure 4

Histopathological and Immunohistochemical findings in AIP (LPSP: Lymphoplasmacytic sclerosing pancreatitis): a fibrosis, prominent infiltration of lymphocytes and plasmacytes (LPSP) are seen. b. Numerous IgG4-positive cells in LPSP are seen. c Storiform fibrosis. An irregularly whorled pattern of fibrosis (storiform fibrosis) with various degrees of infiltration of inflammatory cells, small fusiform cells and fibrotic changes is observed. d Obliterative phlebitis. Stenosis or obstruction of vessels with infiltration of lymphocytes and plasmacytes, and fibrosis is seen. e Circumferential inflammation of LPSP around duct epithelium with stenotic lumen

Type 1 AIP is considered to be a pancreatic manifestation of IgG4-RD [1720, 8789]. Abundant infiltration of IgG4-positive plasmacytes is characteristic, but not specific for IgG4-RD. Because a few cases have been reported in which IgG4-positive plasmacytes appear in patients with pancreatic cancer or alcoholic pancreatitis, IgG4-positive plasmacytes cannot be used as the sole basis for the diagnosis of AIP [1720, 8789]. Therefore, histopathological findings as well as infiltration of IgG4-positive plasmacytes are important in the diagnosis of IgG4-RD [1720, 8789].

The histopathological diagnosis of type 1 AIP is easy with a resected specimen, and difficult with a biopsy specimen. Although many investigators reported that diagnostic values by cytological or histological diagnosis using EUS-FNA are low, those by histological diagnosis using EUS-trucut biopsy are high [9092]. However, a recent study suggested that histological diagnosis by EUS-FNA using 22G needles was useful [93].

Type 2 AIP is idiopathic pancreatitis. It is characterized by the infiltration of neutrophils into the epithelium and/or lumen of interlobular pancreatic ducts, and is referred to as either ICDP or AIP with GEL [22, 23, 82, 85]. To make a definitive diagnosis of type 2 AIP, histopathological criteria are the gold standard due to the lack of specific images and serum markers [9]. In some cases, however, a few IgG4-positive plasmacytes (more than 10 cells/HPF in rare cases) may be observed.

CQ-I-12. How should AIP be diagnosed?

  • A comprehensive diagnosis must be performed based on pancreatic image findings, serological findings, histopathological findings, other organ involvement, and steroid treatment effects. (Level of recommendation: A).

  • The international consensus diagnostic criteria (ICDC) for AIP can differentiates between type 1 and type 2. (Level of recommendation: A).

  • In Japan, most AIP cases are type 1, as defined by the Japanese Clinical Diagnostic Criteria 2011 (JPS-2011). (Level of recommendation: A).

Description

The Japan Pancreas Society took the initiative to propose the world’s first clinical diagnostic criteria for AIP in 2002 [3]. The criteria were revised in 2006 by the joint efforts of the Japan Pancreas Society and the Research Committee for Intractable Pancreas Disease supported by the Ministry of Health and Welfare of Japan (RCIPD-MLHWJ) [4, 5]. The Japanese diagnostic criteria were designed to be as simple as possible, to be easy for general physicians as well as pancreatologists to use. After a decade of international discussion about diagnostic criteria [9496], the ICDC enabled the diagnosis of two distinctive subtypes of AIP, type 1 and type2 AIP [9]. However, the ICDC is somewhat complicated for general use. Different from western countries, extremely few cases of type 2 AIP have been confirmed in Japan [68, 10, 11]. In response to the proposed ICDC, the JPS and RCIPD-MLHWJ revised the clinical diagnostic criteria for AIP in 2011 [10, 11] (Table 3, “Appendix 1”).

Table 3 Clinical diagnostic criteria for autoimmune pancreatitis 2011

In contrast to the JPS2002 and JPS2006 criteria, JPS-2011 contains the following major diagnostic criteria [10, 11], (1) appearance of diffuse and segmental/focal type in pancreatic parenchymal CT/MRI images or ERCP duct images, (2) a single category without level 1 and 2 classifications in the ICDC (Table 4, 5, 6, 7), (3) IgG4 alone as a serum marker, (4) histopathological criteria for LPSP, (5) sclerosing cholangitis, sclerosing sialoadenitis and retroperitoneal fibrosis as typical OOIs, and (6) response to an optional steroid trial after using EUS-FNA to rule out malignancy. As in JPS-2006, the pancreatic images specific to AIP can be confirmed retrospectively after the diagnosis. Although some patients with pancreatic or biliary malignancy may show high serum levels of IgG4, measurement of serum IgG4 is very useful in the diagnosis of AIP because of higher sensitivity and values compared with other serum markers and malignancies, respectively. There are several reports showing that, if infiltration of IgG4-positive plasmacytes is observed in the biopsy of the duodenal papillary mucosa, the chance of the patient having AIP is high [97, 98]. The inflammation of the duodenal papilla is mainly extended from the pancreas and is excluded from OOIs. Diagnostic algorithms for type 1 and type 2 proposed by the ICDC (Figs. 5, 6, 7) permitted modifications depending on the local expertise.

Table 4 Diagnosis of definitive and probable type 1 AIP using ICDC
Table 5 Level 1 and Level 2 criteria for type 1 AIP
Table 6 Diagnosis of Definitive and Probable Type 2 AIP, and AIP-not otherwise specified using ICDC
Table 7 Level 1 and Level 2 Criteria for Type 2 AIP
Fig. 5
figure 5

Algorithm to diagnose type 1 AIP in subjects presenting with obstructive jaundice and/or pancreatic enlargement. This schematic drawing shows a flow to diagnose type 1 AIP with typical diffuse enlargement of the pancreas on CT/MRI (level 1 parenchymal findings)

Fig. 6
figure 6

Algorithm to diagnose type 1 AIP in subjects presenting with obstructive jaundice and/or pancreatic mass. This schematic drawing shows a flow to diagnose type 1 AIP with indeterminate or atypical findings of the pancreas on CT/MRI (level 2 parenchymal findings)

Fig. 7
figure 7

Algorithm to diagnose type 2 AIP in subjects presenting with obstructive jaundice and/or pancreatic mass. This schematic drawing shows a flow to diagnose type 2 AIP with typical/indeterminate (atypical) findings of the pancreas on CT/MRI (level 1 and 2 parenchymal findings)

CQ-I-13. Is facile steroid therapy useful for a differential diagnosis against pancreatic cancer?

  • Response to steroid treatment indicates possible AIP. However, response to steroid treatment does not exclude the possibility of the patient having pancreatic cancer. (Level of recommendation: B).

  • In the cases of segmental/focal swelling or tumor-forming pancreas, facile steroid treatment should be performed only after a negative work-up for malignancy using ERCP or EUS-FNA cytology. (Level of recommendation: B).

In contrast the Mayo’s [94] and Korean [95] criteria, the effect of steroid treatment on the pancreas and OOIs are excluded from the Japanese diagnostic criteria. This decision is based on the following reasons: (1) autoimmune hepatitis is the only autoimmune disease that uses the effect of steroid treatment as a diagnostic criteria; (2) the clinical significance is different for the case of autoimmune hepatitis requiring differentiation from chronic hepatitis of other pathogenesis and the case of AIP requiring differentiation from pancreatic or bile duct cancer; (3) no evidence exists to show that the use of steroids does not affect the success of an operation or the long-term prognosis; (4) there is a danger that therapeutic diagnosis by steroid administration may be used as an easy solution to differentiate AIP from malignant tumors such as pancreatic cancer; (5) the standards were established for not only pancreas specialists but also general gastrointestinal internists and general practitioners; (6) in Japan, the objective of the diagnostic criteria is not so much to find AIP but rather to eliminate the misdiagnosis of diseases with malignant tumors as often as possible; and (7) there have been reports of AIP associated with pancreatic cancer [35]. The ICDC [9] include response to steroid treatment following the Korean [95], and Asian criteria [96] in the criteria proposed jointly by Japan and South Korea in 2008. The JPS-2011 [10, 11] states that if the possibility of pancreatic cancer is excluded by a reliable exclusive-diagnosis using EUS-FNA or a similar test, the effect of steroid treatment may be used as diagnostic criteria. Meanwhile, there have been reports of pancreatic cancers associated with AIP [3]. (Refer to Treatment, Prognosis CQ-IV-10, 11.) If a patient responds to steroid treatment, it may suggest that he/she has AIP; however, since it does not exclude malignant tumors, such as pancreatic cancer or deny the association of pancreatic cancer, a simplistic diagnostic treatment must be avoided.