|
|
||||||||
Abdominal Imaging |
1 The Russell H. Morgan Department of Radiology and Radiological Science, Johns
Hopkins Hospital, 600 N Wolfe St., Baltimore, MD 21287.
2 Department of Pathology, Johns Hopkins Hospital, Baltimore, MD.
Received September 29, 2003;
accepted after revision February 16, 2004.
Address correspondence to S. Kawamoto
(skawamo1{at}jhmi.edu).
Abstract
|
|
|---|
MATERIALS AND METHODS. Five patients with LPSP were evaluated. Morphologic features of the pancreas on CT scans, including the size of the pancreas, presence or absence of a mass, segmental difference of contrast enhancement, pancreatic duct, major pancreatic vasculature, and biliary tract, were retrospectively evaluated and correlated with histopathology. The degree of contrast enhancement of the pancreas was compared in 10 patients without LPSP, who were scanned with the same protocol.
RESULTS. CT scans showed diffuse (n = 2) or focal (n = 3) enlargement of the pancreatic head. The normal lobular appearance of the pancreas was effaced, and the gland appeared featureless in the involved region. Enlarged areas showed an enhancement pattern similar to that of the rest of the pancreas, and no segmental difference of contrast enhancement was identified. Pancreatic duct dilatation was not seen in any patient. Thickening and contrast enhancement of the common bile duct wall (n = 4) and gallbladder wall (n = 3) were observed and were pathologically correlated with inflammatory infiltrate and fibrosis of the common bile duct (n = 3) and gallbladder (n = 1).
CONCLUSION. When these findings are encountered, further evaluation with serologic tests or biopsy may aid in the diagnosis of LPSP.
|
|
|---|
LPSP is a distinctive form of chronic pancreatitis characterized by a mixed
inflammatory infiltrate that centers on the pancreatic ducts
[3]. LPSP is also variously
termed "nonalcoholic duct destructive chronic pancreatitis"
[6], "sclerosing
pancreatitis" or "sclerosing pancreatic cholangitis," and
"autoimmune pancreatitis"
[7,
8]. The pathogenesis of LPSP is
unknown, but it is hypothesized to be a form of autoimmune pancreatitis.
Yoshida et al. [7] reported
unique features of this disorder, including increased serum
-globulin,
presence of autoantibodies, absence of acute attacks of pancreatitis, diffuse
enlargement of the pancreas, and diffuse irregular narrowing of the main
pancreatic duct on endoscopic retrograde pancreatography. Occasional
association with other autoimmune diseases such as Sjögren's disease,
primary sclerosing cholangitis, and ulcerative colitis is made
[7]. Most important, LPSP
responds to steroid therapy, and pancreatic enlargement and diffuse narrowing
of the pancreatic duct may resolve without surgical intervention
[79].
Pathologically, the pancreas reveals dense lymphoplasmacytic infiltrates centered around the pancreatic ducts. A distinctive venulitis is also seen, and the inflammatory process also involves the biliary tree. Although LPSP shares the fibrosis and parenchymal atrophy that is characteristic of other forms of chronic pancreatitis such as ethanol-associated pancreatitis, LPSP lacks the parenchymal calcifications, pseudocysts, and fat necrosis that are commonly present in these other disorders [3, 10]. The morphologic alternations produced by LPSP may simulate malignancy: Masslike enlargement of the pancreatic head or irregular narrowing of the pancreatic duct mimics pancreatic cancer, and stricture of the common bile duct simulates primary bile duct malignancy.
We recently encountered five patients with LPSP who were referred to the radiology department for CT examination of suspected pancreatic malignancy and were treated with pancreaticoduodenectomy. The purpose of this article is to describe the CT features of LPSP, to correlate CT findings with histopathology, and to raise the consciousness of radiologists about the existence of an entity that is readily confused with periampullary malignancy.
|
|
|---|
Abdominal CT examinations, histopathologic results, and clinical records of five patients with LPSP after pancreaticoduodenectomy were reviewed. All patients were referred for abdominal CT examination for suspected pancreatic cancer from April 2002 to June 2003. Our study included three male and two female patients who were 5171 years old (mean age, 61.0 years).
Clinically, all patients presented with obstructive jaundice. Of five patients, four had weight loss, three had abdominal pain, three had diarrhea, and one had nausea and vomiting. Two patients had elevated cancer antigen (CA) 19-9 levels. Pancreaticoduodenectomy was performed in the presence of a presumptive clinical diagnosis of malignant neoplasm of the pancreas, periampullary region, or common bile duct. None of these patients was clinically suspected to have autoimmune pancreatitis; therefore, none had been evaluated for serum autoimmune markers or received a trial of steroid therapy.
CT Examination
Four patients underwent CT examinations performed on a Somatom Volume Zoom
scanner (Siemens Medical Solutions), and one, on a Sensation 16 scanner
(Siemens Medical Solutions). The scans were obtained with a dual-phase
acquisition that consisted of arterial phase images that began 25 sec after
the initiation of contrast material injection and portal venous phase that
began at approximately 55 sec after injection of 120 mL of iohexol (Omnipaque
350, Amersham Health) through a peripheral line at 3 mL/sec. For imaging of
the pancreas and creating vascular maps, 4 x 1 mm or 16 x 0.75 mm
detector collimation was used. For the diagnostic interpretation, axial images
were reconstructed with a slice thickness and increment of 4 or 5 mm for
arterial phase and venous phase imaging. Three-dimensional reconstruction was
also reviewed on a freestanding Onyx Infinite Reality (Silicon Graphics) or
Leonardo (Siemens Medical Solutions) workstation using a 1.25- or 0.75-mm
slice thickness and a 1- or 0.75-mm reconstruction increment. Each patient
ingested 1,000 mL of water over a 20-min period before scanning began. Other
scanning parameters included 120 kVp, 125 mA, 1.25 collimation, and 0.5-sec
rotation speed. All imaging data were reconstructed with the body soft-tissue
algorithm.
Image Analysis
CT findings were retrospectively evaluated. The sizes of the head, body,
and tail of the pancreas were subjectively assessed as small, normal, or
enlarged. Quantitative assessment of Hounsfield units of the pancreas was
determined by manually identifying the region of interest in the same area of
the head, body, and tail of the pancreas on arterial and portal venous phases.
The degree of contrast enhancement of each segment was calculated as
![]() |
|
|
|---|
All patients were referred to our hospital with suspected neoplasm of the pancreatic head or periampullary region on CT (n = 5) or ERCP (n = 3) performed at another institution. ERCP of three patients showed stricture of the distal common bile duct. Two patients underwent endoscopic retrograde cholangiography or ERCP at our institution, and two patients underwent endoscopic sonography at our institution. ERCP of one patient showed a malignant-appearing stricture of the pancreatic duct in the head of the pancreas as well as a 1-cm-long tight stricture of the common bile duct approximately 2 mm in diameter in the area of the pancreatic head with moderate dilatation above it. Endoscopic retrograde cholangiography of the other patient showed a long, smooth stricture in the mid and lower common bile duct with dilatation above the stricture (Fig. 1A, 1B, 1C). Scattered foci of stenosis in the right hepatic ducts were also seen. The pancreatic duct was not cannulated in this patient. Results of endoscopic sonography of one patient were suspicious for a 1.6 x 2.2 cm mass in the pancreatic head with a 1.0 x 0.6 cm peripancreatic lymph node. Fine-needle aspiration of the mass under endoscopic sonography showed rare fragments of atypical pancreatic duct epithelium. Fine-needle aspiration of the lymph node was negative for neoplasm. On endoscopic sonography, the other patient also was suspected of having a 3 x 3 cm mass in the pancreatic head and thickening of the distal common bile duct. Fine-needle aspiration of this mass under endoscopic sonography showed fragments of ductal epithelium and fibrous tissue.
|
|
|
Summary of CT Findings
CT findings of the five patients are summarized in
Table 1. Two patients had
diffuse enlargement of the pancreas (Figs.
1A,
1B,
1C and
2A,
2B,
2C,
2D,
2E,
2F), and three had more focal
enlargement of the pancreatic head (Fig.
3A,
3B,
3C). However, none of the five
patients showed a discrete mass or segmental difference of contrast
enhancement. The usual lobular appearance of the pancreas was effaced, and the
gland appeared featureless and had a relatively straight border in three
patients. A typical capsulelike rim of low attenuation reported in the
literature [11] was seen in
two patients. Contrast enhancement of the pancreas was increased on the venous
phase compared with the arterial phase in four patients and decreased in one
patient, and average degree of contrast enhancement was 11.9% ± 18.5%.
These results were compared with those of 10 patients without LPSP who were
scanned with the same protocol and showed an increased degree of contrast
enhancement on the venous phase compared with the arterial phase in eight
patients and decreased in two patients; average degree of contrast enhancement
was 14.9% ± 17.6% (p = 0.79). Therefore, the degree of
contrast enhancement was not significantly different in patients with LPSP in
our study population. In one patient, the splenic vein was minimally narrowed
by an enlarged pancreatic tail and adjacent inflammation but was patent. Small
lymph nodes up to 1 cm were seen in the porta hepatis, portocaval region, or
peripancreatic region in two patients (Fig.
2A,
2B,
2C,
2D,
2E,
2F). None of these five
patients showed dilatation of the main pancreatic duct or calcification of the
pancreas. Thickening and contrast enhancement of the common bile duct were
seen in four patients (Figs.
1A,
1B,
1C and
2A,
2B,
2C,
2D,
2E,
2F). Three of them had mixed
infiltrates and fibrosis consistent with common bile duct disease associated
with LPSP, but in one patient the specimen was unavailable for review. The
fifth patient had pathologic evidence of common bile duct involvement, but CT
assessment was compromised by pneumobilia. Mild to moderate gallbladder wall
thickening and contrast enhancement were seen in four patients, and two
patients had mixed infiltrates and fibrosis consistent with gallbladder
disease associated with LPSP (Fig.
2A,
2B,
2C,
2D,
2E,
2F). Three patients also had
cholelithiasis.
|
|
|
|
|
|
|
|
|
|
Surgery and Pathology Findings
Although these findings did not have the classic appearance of pancreatic
cancer, the patients were clinically thought to have resectable pancreatic
cancer. Specifically, in one patient, the diffuse enlargement of the gland was
considered indicative of pancreatitis, probably secondary to a small tumor
obstructing the pancreatic duct. In the other two patients, a mass was
suspected in the pancreatic head on endoscopic sonography. Fine-needle
aspiration under sonographic guidance was performed in these two patients, and
the specimen from one of these two patients showed fragments of atypical
pancreatic duct epithelium. All patients underwent
pancreaticoduodenectomy.
At surgery, the pancreas was enlarged and extremely firm in all patients. Three patients with more focal enlargement of the pancreatic head had a hard mass in the head of the pancreas, but the remaining body and tail of the pancreas were also firm and hard in texture. Dense adherence between the pancreas and the structures of the retroperitoneum, the adjacent superior mesenteric vein, and the superior mesenteric artery was described in three patients.
Pathologically, the pancreas showed features typical of LPSP, including mixed inflammatory cell infiltrates centered around the pancreatic ducts (Fig. 2A, 2B, 2C, 2D, 2E, 2F) and venulitis, in all patients. One patient who had fragments of atypical pancreatic duct epithelium at fine-needle aspiration under endoscopic sonography had LPSP and focal pancreatic intraepithelial neoplasia [12], but no carcinoma was identified. The common bile duct showed dense mixed infiltrates and fibrosis in four patients (Fig. 2A, 2B, 2C, 2D, 2E, 2F). The gallbladder also showed mild mixed infiltrates and fibrosis in two patients (Table 1).
Four patients were doing well without abdominal pain after operation. One patient had occasional abdominal pain but has gained weight.
|
|
|---|
The clinical presentation of LPSP and pancreatic cancer can be very similar. Hardacre et al. [11] reported that no statistically significant differences in the rates of abdominal pain, weight loss, jaundice, or preoperative carcinoembryonic agent or CA 19-9 levels. Jaundice was seen in 84% of patients with LPSP who underwent pancreaticoduodenectomy [13]. Although a focal mass is the primary CT finding of a pancreatic adenocarcinoma, not all carcinomas present as masses. Differentiation between carcinoma and inflammatory disease can be difficult because the pancreas may be diffusely enlarged by a carcinoma-induced pancreatitis. In a certain percentage of cases, only surgically obtained histologic material will resolve this question. Because of recent improvement in the safety of pancreaticoduodenectomy and because of false-negative results from needle biopsy, particularly in patients who have the smallest lesions and the best likelihood of recovery, surgeons make a decision to proceed directly to resection without performing biopsy [2].
CT features of this form of pancreatitis have previously been reported in the literature [6, 9, 11]. Among 15 of these reported cases, 11 patients showed a diffusely enlarged pancreas and four patients had more focal enlargement. One patient had small calcifications in the head. Common bile duct stenosis or dilatation was described in six patients. The pancreatic duct is usually not dilated, but mild upstream dilatation of the main pancreatic duct was reported in four patients. Splenic vein occlusion was described in one patient, but no other vascular involvement was reported. A capsulelike low-attenuation rim that showed subtle delayed enhancement was described in four patients. Involved pancreatic parenchyma may show zones of decreased contrast enhancement with or without discrete margins on arterial or early phase images. When delayed phase images were acquired, the lesion showed increased contrast enhancement in seven of eight reported cases. However, the lesion may remain hypodense on delayed phase compared with relatively normal parenchyma [15]. Mass or masslike enlargement has been reported more commonly in the pancreatic head but may be seen in the body or tail [6]. The existence of multiple masses has also been reported [16]. In our five patients, no discrete mass or discrete area of different contrast enhancement was observed even with focal enlargement. No pancreatic duct dilatation was seen in our five patients.
Patients with LPSP have distal common bile duct fibrosis, strictures, and inflammationfeatures that overlap somewhat with primary sclerosing cholangitis. Abraham et al. [17] reported that among 20 patients with LPSP who were treated with pancreaticoduodenectomy, 12 (60%) had moderate or marked inflammatory infiltrates in the extrapancreatic common bile duct and 84.2% had moderate to marked inflammatory infiltrates in the intrapancreatic common bile duct. Most (60%) of the gallbladders in LPSP also contained moderate or marked inflammatory infiltrates and lymphoid nodules [17]. Abraham et al. concluded that gallbladder involvement in LPSP, similar to primary sclerosing cholangitis, is most commonly part of a spectrum of biliary tract disease that includes frequent distal strictures, fibrosis, and inflammation of the extrapancreatic or intrapancreatic common bile duct. Thickening of the gallbladder wall on CT scans has been described in a reported case [18].
Several CT features can be useful in differentiating LPSP from pancreatic cancer. The pancreatic duct usually shows diffuse or focal irregular narrowing in LPSP without significant dilatation, but it is usually dilated in patients with pancreatic ductal adenocarcinoma. Major pancreatic vascular involvement is uncommon in LPSP compared with pancreatic adenocarcinoma, although cases have been reported to involve the peripancreatic vessels [9]. One of our five patients also showed mild splenic vein narrowing. Atrophy of the pancreas proximal to duct obstruction is often seen in patients with pancreatic cancer, but LPSP usually is not manifest in detectable atrophy. Acinar parenchyma becomes atrophic in LPSP but is replaced with fibrous tissue [3], which explains why overall size of the gland does not change.
When the characteristic features of LPSP are observed on CT scans, the diagnosis of LPSP can be suggested. Procacci et al. [15] reported that dynamic CT findings can aid in making a correct diagnosis of autoimmune pancreatitis with 92.5% accuracy. They evaluated seven patients with proven autoimmune pancreatitis and 20 patients with other pancreatic diseases (acute or chronic pancreatitis and adenocarcinoma of the pancreas) using the following criteria: focal or diffuse enlargement; net hypodensity compared with the spleen and adjacent unaffected parenchyma on early (or arterial), late (or venous), or delayed contrastenhanced images; possible identification of a hypodense capsulelike rim; absence of calcification; possible stenosis or obstruction of the common bile duct at the pancreatic head; and possible stenosis or upstream dilatation of the main pancreatic duct. In these patients, immunologic tests for elevated IgG4 level may be helpful to confirm the diagnosis [19].
Findings of LPSP using other imaging techniques have been reported. On sonographic examination, the pancreas appears hypoechoic and diffusely enlarged with a so-called sausagelike appearance [8, 9] or as a hypoechoic mass in the affected site [6]. MRI may reveal diffuse pancreatic enlargement with hypointensity on T1-weighted images. The low-attenuation capsulelike rim described on CT is hypointense on T2-weighted images and shows delayed enhancement, suggesting fibrous tissue rather than a fluid collection or phlegmon [12]. ERCP may show diffuse narrowing and irregularity of the main pancreatic duct [8, 9] with nonvisualization of secondary branches. Gallium scintigraphy [20] and FDG PET [21] have been reported to show increased uptake in the affected site of the pancreas, which can be confused with malignancy.
In conclusion, LPSP is a unique clinical entity that is becoming more recognized in clinical practice. Because it responds to steroid therapy, early recognition is important to preclude surgery. CT features suggestive of LPSP include diffuse or focal enlargement of the pancreas, a rim of low attenuation surrounding the pancreas, absence of atrophy of the pancreas, or absence of significant pancreatic duct dilatation. Thickening and contrast enhancement of the common bile duct and the gallbladder may be observed, reflecting inflammatory disease of these structures associated with LPSP. When these CT findings are encountered, further evaluation with serologic tests or biopsy is justified.
|
|
|---|
This article has been cited by other articles:
![]() |
N. Takahashi, J. G. Fletcher, J. L. Fidler, D. M. Hough, A. Kawashima, and S. T. Chari Dual-Phase CT of Autoimmune Pancreatitis: A Multireader Study Am. J. Roentgenol., February 1, 2008; 190(2): 280 - 286. [Abstract] [Full Text] [PDF] |
||||
![]() |
S. Kawamoto, S. S. Siegelman, R. H. Hruban, and E. K. Fishman Lymphoplasmacytic Sclerosing Pancreatitis (Autoimmune Pancreatitis): Evaluation with Multidetector CT RadioGraphics, January 1, 2008; 28(1): 157 - 170. [Abstract] [Full Text] [PDF] |
||||
![]() |
S. Kim, N. K. Lee, J. W. Lee, C. W. Kim, S. H. Lee, G. H. Kim, and D. H. Kang CT Evaluation of the Bulging Papilla with Endoscopic Correlation RadioGraphics, July 1, 2007; 27(4): 1023 - 1038. [Abstract] [Full Text] [PDF] |
||||
![]() |
D. Brennan and I. Pedrosa Lymphoplasmacytic Sclerosing Pancreatitis Am. J. Roentgenol., November 1, 2005; 185(5): 1367 - 1368. [Full Text] [PDF] |
||||
![]() |
S. Kawamoto and E. K. Fishman Reply Am. J. Roentgenol., November 1, 2005; 185(5): 1368 - 1368. [Full Text] [PDF] |
||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |