Secondary Logo

Journal Logo

Strategy for Differentiating Autoimmune Pancreatitis From Pancreatic Cancer

Kamisawa, Terumi MD, PhD*; Imai, Mitsuho MD*; Yui Chen, Pong MD*; Tu, Yuyang MD, PhD*; Egawa, Naoto MD*; Tsuruta, Kouji MD, PhD; Okamoto, Atsutake MD, PhD; Suzuki, Mizuka MD; Kamata, Noriko MD, PhD

doi: 10.1097/MPA.0b013e318175e3a0
Online Articles: Original Article
Free

Objectives: It is of utmost importance that autoimmune pancreatitis (AIP) be differentiated from pancreatic cancer (PC) because some AIP cases undergo unnecessary laparotomy or pancreatic resection on suspicion of PC. This study aimed to develop an appropriate strategy for differentiating between AIP and PC.

Methods: Clinical, serological, and radiological features of 17 AIP patients forming a masslike lesion on pancreas head and 70 patients with pancreatic head cancer were compared.

Results: Numerous findings can be used to distinguish between AIP and PC, and the following are more likely in AIP: fluctuating jaundice; elevated serum IgG4 levels; delayed enhancement of the enlarged pancreas and a capsule-like low-density rim on computed tomography; long or skipped narrowed portion with side branches of the main pancreatic duct without upstream dilatation on endoscopic retrograde pancreatography, extrapancreatic lesions, such as stenosis of the intrahepatic bile duct, salivary gland swelling, and retroperitoneal mass; and responsiveness to steroid therapy.

Conclusions: In elderly male patients presenting with obstructive jaundice and a pancreatic mass, AIP should be considered in the differential diagnosis. Based on a combination of clinical, serological, and radiological findings, AIP can be differentiated from PC. An algorithm for management of patients with a masslike lesion on pancreas head is presented.

From the Departments of *Internal Medicine, †Surgery, and ‡Radiology, Tokyo Metropolitan Komagome Hospital, Tokyo, Japan.

Received for publication November 26, 2007; accepted March 14, 2008.

This study was supported by Research for Intractable Disease of the Pancreas, Minister of Labor and Welfare of Japan.

Reprints: Terumi Kamisawa, MD, PhD, Department of Internal Medicine, Tokyo Metropolitan Komagome Hospital, 3-18-22 Honkomagome, Bunkyo-ku, Tokyo 113-8677, Japan (e-mail: kamisawa@cick.jp).

Autoimmune pancreatitis (AIP) is a recently identified clinical entity of pancreatitis in which autoimmune mechanisms are suspected to be involved in the pathogenesis.

During the last 10 years, many new clinical aspects of AIP have been recognized, and AIP has become a distinct entity that is recognized worldwide. However, the precise pathogenesis or pathophysiology remains unknown. Autoimmune pancreatitis responds dramatically to steroid therapy; therefore, to avoid unnecessary surgery, an accurate diagnosis of AIP is required.1-3 The most important disease that should be differentiated from AIP is pancreatic cancer (PC). Because it is usually difficult to take adequate specimens from the pancreas, AIP is currently diagnosed based on a combination of clinical, laboratory, and imaging studies.4-7

In 2006, the Japan Pancreas Society proposed the "Clinical Diagnostic Criteria for Autoimmune Pancreatitis".4 It contained 3 items: (1) radiological imaging showing diffuse or localized enlargement of the pancreas and diffuse or segmental irregular narrowing of the main pancreatic duct; (2)laboratory data showing abnormally elevated levels of serum gamma-globulin, IgG, or IgG4, or the presence of autoantibodies; and (3) histological findings showing marked interlobular fibrosis and prominent lymphoplasmacytic infiltration in the pancreas. To make the diagnosis of AIP, criterion 1 is mandatory, and either criterion 2 or criterion 3 must be present.

However, in particular, AIP forming a masslike lesion on pancreas head is sometimes difficult to differentiate from locally advanced pancreatic head cancer. Histologically, most cases of pancreatic head mass are PC, and a few of them are AIP. When we see patients with a mass on pancreas head on imaging, who should be suspected of AIP? Then how we proceed to make differentiation between AIP and PC? To develop an appropriate strategy to differentiate between AIP and PC, clinical, serological, and radiological features of AIP and PC patients were compared.

Back to Top | Article Outline

MATERIALS AND METHODS

Study Patients

Between May 1998 and December 2007, 37 patients were diagnosed as having AIP involving the pancreatic head based on radiological (n = 37), serological (n = 37), and histological [n = 11: resection (n = 4), biopsy (n = 6), and autopsy (n = 1)] examinations. In them, 17 AIP patients forming a masslike lesion on pancreas head were enrolled in this study. There were 13 males and 4 females, and their average age was 64.2 ± 11.2 years. All patients had segmental enlargement of the pancreas head and narrowing of the main pancreatic duct. On histology, lymphoplasmacytic infiltration and fibrosis were found in the pancreas in all 7 of the patients who were assessed. Three patients had a pancreatic resection, and 3 patients had a bypass operation since PC was suspected. Eleven patients were given steroid therapy, and all of them responded well based on clinical and radiological follow-up examinations. Long-term follow-up was completed in all patients (5.0 ± 2.7 years [mean ± SD]; range, 1.5-9.5 years), and malignancy was completely ruled out from histology of the pancreas and/or the follow-up study.

During the same period as the AIP cases, 73 malignant solid tumors of the pancreatic head were resected or histologically confirmed after adequate imaging studies. Histologically, they were tubular adenocarcinoma (n = 70), malignant endocrine tumor (n = 2), and acinar cell carcinoma (n = 1). Both malignant endocrine tumor and acinar cell carcinoma could be diagnosed preoperatively based on the findings of hypervascularity and serum tumor markers such as neuron-specific enolase or elastase-1. The 70 pancreatic head cancer cases of locally advanced tubular adenocarcinoma (stage III or IVa)8 were also reviewed.

Back to Top | Article Outline

Clinical and Radiological Analysis

The following clinical factors were retrospectively assessed: age on diagnosis, sex ratio, drinking habit (drinking more than 40 g of alcohol/d for more than 7 years), smoking (more than 20 pack-years; pack years = the number of packages of cigarette per day times years of smoking), initial symptoms, and salivary gland swelling (based on physical examination and confirmed on computed tomography [CT] or ultrasonography). Serologically, immunologic findings including IgG4, autoantibodies, and tumor markers were analyzed.

The findings on the following imaging studies were reviewed by 2 radiologists and 2 endoscopists who were unaware of the clinical situation or final diagnosis. The performed imaging studies were CT (AIP: n = 17, PC: n = 70), celiac and supramesenteric angiography (AIP: n = 8, PC: n=18), endoscopic retrograde pancreatography (ERP) (AIP: n= 17, PC: n = 40), endoscopic retrograde cholangiography (ERC) (AIP: n = 17, PC: n = 40), and magnetic resonance cholangiopancreatography (MRCP) (AIP: n = 12, PC: n = 50). On CT, delayed enhancement of the enlarged pancreatic head (Fig. 1A), a capsule-like low-density rim around the pancreas (Fig. 1A), thickening of the wall of the bile duct in which abnormalities were not clearly evident on cholangiography, gallbladder wall thickening, and retroperitoneal mass (fibrosis) were assessed. On angiography, encasement of the peripancreatic arteries and deviation or obstruction of the portal vein were assessed. On ERP, the length of the narrowed portion of the main pancreatic duct, skipped narrowed lesions of the main pancreatic duct (Fig. 1B), side branches derived from the narrowed portion of the main pancreatic duct (Fig. 1B), obstruction of the main pancreatic duct, and maximal diameter of the upstream main pancreatic duct were assessed. In cases with obstructed main pancreatic duct, maximal diameter of the upstream main pancreatic duct was measured on MRCP. On ERC, stenosis or deviation of the lower bile duct and stenosis of the intrahepatic bile duct were assessed.

FIGURE 1

FIGURE 1

Back to Top | Article Outline

Statistical Analysis

Statistical analyses were performed with the Mann-Whitney test and Fisher exact probability test. Logistic regression analysis was performed in the 12 imaging factors on CT and ERCP. In all tests, P < 0.05 were considered as statistically significant.

Back to Top | Article Outline

RESULTS

Clinical Differences Between AIP and PC

Clinically, there were no differences between the 2 groups with respect to age, sex ratio, alcohol intake, and smoking habit. With respect to initial symptoms, weight loss (more than 2 kg/mo) was frequently observed in PC patients (P< 0.001), and fluctuating jaundice occurred frequently in AIP patients (P < 0.001). On physical examinations, salivary gland swelling was detected in 24% of AIP patients, but not in any of the PC patients (P < 0.001; Table 1).

TABLE 1

TABLE 1

Back to Top | Article Outline

Serological Differences Between AIP and PC

Serum IgG4 levels were frequently (P < 0.001) and significantly (P < 0.001) elevated in AIP patients. After steroid therapy, the elevated serum IgG4 levels decreased in all AIP patients. There were no differences with respect to the presence of autoantibodies or the tumor marker levels except CA19.9 between the AIP and the PC patients. Serum CA19.9 levels were significantly elevated in PC patients than in AIP patients (P < 0.001; Table 2).

TABLE 2

TABLE 2

Back to Top | Article Outline

Radiological Differences Between AIP and PC

On CT, the following were frequently detected in AIP patients: delayed enhancement of the enlarged pancreas (P <0.001); a capsule-like low-density rim around the pancreas (P < 0.001); no atrophic change of the pancreatic body or tail (P < 0.001); thickening of the wall of the bile duct in which stenosis was not clearly observed on cholangiography (P < 0.001); gallbladder wall thickening (P = 0.006); and retroperitoneal mass (P = 0.036).

There were no differences between AIP and PC patients in the angiographic findings. On ERP, skipped narrowed lesions of the main pancreatic duct (P < 0.001) and side branches derived from the narrowed portion of the main pancreatic duct (P = 0.036) were frequently detected in AIP patients, whereas obstruction of the main pancreatic duct was frequently detected in PC patients (P < 0.001). Furthermore, in AIP patients as compared with PC patients, the length of the narrowed portion of the main pancreatic duct was longer (P< 0.001) and the maximal diameter of the upstream main pancreatic duct was smaller (P < 0.001; Table 3). In respect to extrapancreatic lesions such as salivary gland swelling, retroperitoneal fibrosis, and stenosis of the intrahepatic bile duct, they were more frequently detected in AIP patients (6/17vs 0/70; P < 0.001).

TABLE 3

TABLE 3

On logistic regression analysis, there were no significant differences in these 12 imaging factors on CT and ERCP (raging from P = 0.821 to P = 0.996). Then we chose the following 6 imaging findings suggesting AIP rather than PC, which can be objectively and accurately evaluated. They were enhancement of the enlarged pancreas on CT; a capsule-like rim on CT; narrowed portion of the main pancreatic duct ≥3-cm-long on ERP; skipped lesions of the main pancreatic duct on ERP; a maximal diameter <5 mm of the upstream main pancreatic duct on ERP; and the presence of extrapancreatic lesions, such as salivary gland swelling, retroperitoneal mass, and stenosis of the intrahepatic bile duct on CT or ERC. MRCP can also be used to examine the presence of these findings assessed on ERP or ERC. All AIP patients had more than 2 positive imaging factors for AIP. Two PC patients had an elevated serum IgG4 level and showed delayed enhancement of the mass. Overall, 31 PC patients had normal serum IgG4 levels; of these, 26 had no positive imaging factors, 4 had one other positive imaging factor (a capsule-like rim or a long narrowed portion, respectively), and 1 had 2 other positive imaging factors (a long narrowed portion with a less dilated upstream main pancreatic duct; Table 4).

TABLE 4

TABLE 4

Back to Top | Article Outline

DISCUSSION

Autoimmune pancreatitis is a new entity of pancreatitis that has a presumed autoimmune etiology, and it can mimic malignancy, both clinically and radiologically.1-3 Currently, the diagnosis of AIP is based on a combination with clinical, laboratory, and imaging studies.4-6 It is of utmost importance that AIP be differentiated from PC. In North America, approximately 2.5% of pancreatoduodenectomies were done in AIP patients who were mistakenly diagnosed as having PC.9Nakazawa et al10 reported that 7 of 37 AIP patients had surgery since they were misdiagnosed as having pancreatic or bile duct cancer; 8 of our 37 patients were also misdiagnosed as having PC. Because AIP responds dramatically to steroid therapy, an accurate diagnosis of AIP can avoid unnecessary laparotomy or pancreatic resection. Therefore, to develop an appropriate strategy for differentiating between AIP and PC, the clinical, serological, and radiological features of AIP and PC patients were compared.

Clinically, AIP and PC patients frequently present with obstructive jaundice, but AIP patients sometimes have fluctuating jaundice. Serum IgG4 levels are frequently and significantly elevated in AIP patients.11,12 In the present study, elevation of serum IgG4 levels (>135 mg/dL) was detected in 71% of AIP patients and 6% of PC patients. Although the measurement of serum IgG4 levels is useful for differentiating between the 2 diseases, it should be noted that some PC patients also have elevated serum IgG4 levels.13,14 Ghazale etal14 reported that serum IgG4 levels were elevated more than 140 mg/dL in 13 (10%) of 135 PC patients and more than 280mg/dL in only 2 (1%) PC patients, and they recommended 280 mg/dL as cutoff value to differentiate AIP from PC. When we use 280 mg/dL as cutoff value, 1 of 2 false-positive PC patients in this study becomes negative, but sensitivity of AIP lowers from 71% to 47%. Raina et al15 reported that 5 (7%) of71 PC patients had serum IgG4 elevation with a mean level of 160.8 mg/dL and that an IgG4 level measuring between 135and 200 mg/dL should be interpreted cautiously and not accepted as diagnostic of AIP without further evaluation. Measurement of serum IgG4 levels is not a routine examination in general hospital because the cost of measurement of IgG4 is not covered at present by Japanese nation health insurance. Therefore, it is necessary to suggest in whom serum IgG4 should be measured in patients with a pancreatic mass.

On histology, the pancreas of AIP patients was found to have dense lymphoplasmacytic infiltration and fibrosis.1-6,16 These findings are related to several imaging features that are characteristic of AIP. On dynamic CT, a mass caused by AIP shows delayed enhancement, unlike in PC, and substantial parenchymal atrophy is lacking. As Irie et al17 reported, the presence of a capsule-like low-density rim that surrounds the pancreas, which may correspond to peripancreatic inflammation and fibrosis, strongly suggests AIP.

Histologically, the pancreatic duct is narrowed by peripancreatic nonocclusive fibrosis and lymphoplasmacytic infiltration in AIP patients.1,16 This finding is usually extensive in the pancreas, although the degree of inflammation varies by location in the same patient.1 Given these AIP changes, on ERP, the length of the narrowed portion of the main pancreatic duct is longer, and skipped narrowed lesions of the main pancreatic duct occur more frequently in AIP patients than in PC patients. Skipped narrowed lesions of the main pancreatic duct were reported in 16%10 to 26%18 of AIP cases. Although side branches from the stenotic main pancreatic duct are rarely seen in PC patients due to obstruction by invasion of cancer cells, in AIP patients, several side branches are frequently derived from the narrowed portion of the main pancreatic duct because the infiltrate is primarily subepithelial and the ductal epithelium is usually preserved.1 Obstruction of the main pancreatic duct is also rarely seen in AIP patients. In AIP patients with segmental narrowing of the main pancreatic duct, upstream dilatation of the distal pancreatic duct is less often noted than in PC patients.

Patients with AIP frequently have various extrapancreatic lesions.1,2,15,19 Because these lesions have similar histopathological findings to those found in the pancreas, including abundant infiltration of IgG4-positive plasma cells and T lymphocytes, as well as fibrosis, these extrapancreatic lesions are possibly induced by the same IgG4-related fibroinflammatory mechanisms as in AIP.1,15,20 Sclerosing cholangitis is the extrapancreatic lesion that is most frequently associated with AIP, and the stenosis is located in the lower part of the bile duct in most cases. Because histological examination in AIP patients shows extensive inflammation of the bile duct wall, wall thickening of the bile duct, even in the segment inwhich abnormalities are not clearly observed with cholangiography, is sometimes detected on imaging. Stenosis in the intrahepatic bile duct is sometimes detected in AIP patients, and this finding helps making diagnosis of AIP.10 The presence of other possible extrapancreatic lesions, such as salivary gland swelling (sclerosing sialadenitis) and a retroperitoneal mass (retroperitoneal fibrosis), also strongly suggests AIP rather than PC.

In AIP patients, obliterative phlebitis of variably sized pancreatic veins involving the portal vein and duplication of the internal elastica and intimal thickening of small arteries can be occasionally observed.1 Therefore, angiographic abnormalities that are commonly seen in PC patients are also frequently observed in AIP patients. These angiographic findings can obscure the diagnosis of AIP.

As steroid therapy is effective in AIP patients, dramatic improvement after steroid therapy confirms the diagnosis of AIP. However, empiric administration of steroids should be avoided because the facile use of steroid trials may result in delaying PC surgery, which could lead to cancer progression in some cases. Currently, the indications of steroid therapy for AIP are thought to be stenosis of the bile duct due to sclerosingcholangitis with and without obstructive jaundice, abdominal pain, and association with symptomatic extrapancreatic lesions such as retroperitoneal fibrosis.1,4

Masslike lesions on pancreas head without obvious metastatic lesions are histologically divided into neoplastic and inflammatory lesions. Most of the neoplastic lesions are pancreatic ductal cancer, and rare neoplasms, such as endocrine or acinar tumors, can be differentiated based on vascularity and tumor markers. Except for the peculiar inflammatory mass associated with alcoholic chronic calcified pancreatitis or acute pancreatitis that can be easily diagnosed clinically, all other enlarged inflammatory masses are AIP.21

Based on these findings, an algorithm for clinical management of a masslike lesion on pancreas head was developed, with special emphasis on differentiating between AIP and PC (Fig. 2). Many cases with a masslike lesion on pancreas head present with obstructive jaundice; endoscopic or percutaneous trashepatic biliary drainage is usually done initially. At that time, cytological examination of the bile or pancreatic duct should be performed as soon as possible to confirm or rule out malignancy. In cases with no positive imaging factors for AIP, surgery should be considered under the provisional diagnosis of PC. When there is at least 1 positive imaging factor for AIP, serum IgG4 levels should be measured. When serum IgG4 levels are elevated (more than 135 mg/dL), in cases with more than 2 positive imaging factors, indication of steroid therapy should be considered under the provisional diagnosis of AIP, whereas in cases with 1positive imaging factor, biopsy guided by ultrasound or endoscopic ultrasonography-guided fine-needle aspiration should be done for histological examination. Indication of steroid therapy should be considered under the provisional diagnosis of AIP in patients without evidence of cancer on histological examination, whereas those who are cancer-positive should have surgery. In patients with normal serum IgG4 levels who have more than 3 positive imaging factors, indication of steroid therapy should be considered, whereas cases with 1 or 2 positive imaging factors should be biopsied for histological examination. Because AIP responds so readily to steroid therapy, a poor response to steroid therapy suggests PC and the need for further reexamination, including laparotomy.

FIGURE 2

FIGURE 2

The criteria5 proposed by Mayo Clinic indicate that histology of the pancreas could be considered the gold standard for the diagnosis of AIP. However, pancreatic core biopsy or cytology guided with endoscopic ultrasonography can be performed easily and safely in the limited institutes in the world. This algorithm aimed at to be used by general gastroenterlogists, and suggested cases in whom biopsy of the pancreas should be performed for differential diagnosis. This algorithm, developed based on the results of this study, needs to be verified in future prospective studies because it was based on a small patient population. Furthermore, 3 cases coexisting of AIP and PC have been recently reported.22-24 We should also keep in mind that only characteristic differences in the present study cannot estimate the coexistence of AIP and PC.

In conclusion, in elderly male patients presenting with obstructive jaundice and a pancreatic mass, AIP should be considered in the differential diagnosis to avoid unnecessary surgery. In patients with a masslike lesion on pancreas head, differentiation between AIP and PC can be done according to this algorithm.

Back to Top | Article Outline

REFERENCES

1. Kamisawa T, Okamoto A. Autoimmune pancreatitis: proposal of IgG4-related sclerosing disease. J Gastroenterol. 2006;41:613-625.
2. Finkelberg DL, Sahani D, Deshpande V, et al. Autoimmune pancreatitis. N Engl J Med. 2006;355:2670-2676.
3. Hirano K, Tada M, Isayama H, et al. Long-term prognosis of autoimmune pancreatitis without and with corticosteroid treatment. Gut. 2007;56: 1719-1724.
4. Okazaki K, Kawa S, Kamisawa T, et al. Clinical diagnostic criteria of autoimmune pancreatitis: revised proposal. J Gastroenterol. 2006;41: 626-631.
5. Chari S, Smyrk TC, Levy MJ, et al. Diagnosis of autoimmune pancreatitis: the Mayo Clinic experience. Clin Gastroenterol Hepatol. 2006;4:1010-1016.
6. Kim KP, Kim MH, Kim JC, et al. Diagnostic criteria for autoimmune chronic pancreatitis revised. World J Gastroenterol. 2006;12:2487-2496.
7. Kamisawa T, Egawa N, Nakajima H, et al. Clinical difficulties in the differentiation of autoimmune pancreatitis and pancreatic carcinoma. Am J Gastroenterol. 2003;98:2694-2699.
8. American Joint Committee on Cancer. TNM Classification of Malignant Tumors, 6th ed. Philadelphia, PA: Wiley-Liss, Inc., 2002.
9. Hardacre JM, Iacobuzio-Donahue CA, et al. Results of pancreaticoduodenectomy for lymphoplasmacytic sclerosing pancreatitis. Ann Surg. 2003;237:853-859.
10. Nakazawa T, Ohara H, Sano H, et al. Difficulty in diagnosing autoimmune pancreatitis by imaging findings. Gastrointest Endosc. 2007;65:99-108.
11. Hamano H, Kawa S, Horiuchi A, et al. High serum IgG4 concentrations in patients with sclerosing pancreatitis. N Engl J Med. 2001;344:732-738.
12. Hirano K, Kawabe T, Yamamoto N, et al. Serum IgG4 concentrations in pancreatic and biliary diseases. Clin Chim Acta. 2006;367:181-184.
13. Kamisawa T, Chen PY, Tu Y, et al. Pancreatic cancer with a high serum IgG4 concentration. World J Gastroenterol. 2006;12:6225-6228.
14. Ghazale A, Chari ST, Smyrk TC, et al. Value of serum IgG4 in the diagnosis of autoimmune pancreatitis and in distinguishing it from pancreatic cancer. Am J Gastroenterol. 2007;102:1646-1653.
15. Raina A, Krasinskas AM, Greer JB, et al. Serum immunoglobulin G fraction 4 levels in pancreatic cancer. Elevations not associated with autoimmune pancreatitis. Arch Pathol Lab Med. 2008;132:48-53.
16. Kamisawa T, Funata N, Hayashi Y, et al. Close relationship between autoimmune pancreatitis and multifocal fibrosclerosis. Gut. 2003;52: 683-687.
17. Irie H, Honda H, Baba S, et al. Autoimmune pancreatitis: CT and MR characteristics. AJR Am J Roentgenol. 1998;170:1323-1327.
18. Horiuchi A, Kawa S, Hamano H, et al. ERCP features in 27 patients with autoimmune pancreatitis. Gastrointest Endosc. 2002;55:494-499.
19. Kamisawa T, Egawa N, Nakajima H, et al. Extrapancreatic lesions in autoimmune pancreatitis. J Clin Gastroenterol. 2005;39:904-907.
20. Kamisawa T, Funata N, Hayashi Y, et al. A new clinicopathological entity of IgG4-related autoimmune disease. J Gastroenterol. 2003;38:982-984.
21. Takase M, Suda K. Histopathological study on mechanism and background of tumor-forming pancreatitis. Pathol Int. 2001;51:349-354.
22. Inoue H, Miyatani H, Sawada Y, et al. A case of pancreas cancer with autoimmune pancreatitis. Pancreas. 2006;33:208-209.
23. Ghazale A, Chari S. Is autoimmune pancreatitis a risk factor for pancreatic cancer? Pancreas. 2007;35:376.
24. Fukui T, Mitsuyama T, et al. Pancreatic cancer associated with autoimmune pancreatitis in remission. Intern Med. 2008;47:151-155.
Keywords:

autoimmune pancreatitis; pancreatic cancer; IgG4; CT

© 2008 Lippincott Williams & Wilkins, Inc.