| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Original Research |
1 Department of Radiology, Mayo Clinic, 200 First St. SW, Rochester, MN
55905.
2 Department of Internal Medicine, Mayo Clinic, Rochester, MN.
Received March 23, 2007;
accepted after revision August 7, 2007.
Address correspondence to N. Takahashi.
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MATERIALS AND METHODS. Dual-phase CT scans of 74 patients (25 with autoimmune pancreatitis, 33 with pancreatic carcinoma, and 16 with a normal pancreas) were independently evaluated by three radiologists for enhancement of the pancreas; the presence of a capsule-like rim, peripancreatic strands, and pancreatic calcifications; pancreatic duct or bile duct changes; and renal involvement. The frequency of CT characteristics was compared between autoimmune pancreatitis and carcinoma. Interobserver agreement for the three reviewers for the assessment of CT characteristics was evaluated using kappa statistics.
RESULTS. Diffusely decreased enhancement of the pancreas (autoimmune
pancreatitis vs carcinoma: 28% vs 3%; p = 0.02, 
=
0.33–0.75), capsule–like rim (40% vs 9%; p = 0.009,
= 0.42–0.66), peripancreatic strands (60% vs 27%; p =
0.02, = 0.45–0.54), pancreatic calcifications (32% vs 9%;
p = 0.04, = 0.14–0.47), bile duct wall enhancement (52%
vs 6%; p = 0.0001, = 0.28–0.47), and renal involvement
(28% vs 0%; p = 0.002, = 0.32–0.74) were more frequent
in patients with autoimmune pancreatitis. Pancreatic duct dilation (24% vs
67%; p = 0.001, = 0.65–0.73) and abrupt cutoff (16% vs
55%; p = 0.003, = 0.60–0.65) were more frequent in
patients with carcinoma.
CONCLUSION. Diffusely decreased enhancement of the pancreas, a capsule-like rim, bile duct enhancement, and renal involvement are useful signs of autoimmune pancreatitis.
Keywords: autoimmune pancreatitis • CT • gastrointestinal tract • pancreas • pancreatic carcinoma
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The purpose of this study is to identify CT findings that aid in differentiating autoimmune pancreatitis from pancreatic carcinoma using dual-phase CT.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Age- and sex-matched patients with pancreatic carcinoma (n = 33; mean age, 64 years; median age, 66 years; range, 40–84 years; 29 men and four women) and with a normal pancreas (n = 16; mean age, 64 years; median age, 68 years; range, 33–81 years; 14 men and two women) who underwent dual-phase CT of the pancreas were included as control groups. These patients were randomly selected from a larger number of patients who were identified by a search of radiology reports for diagnostic codes (normal pancreas and pancreatic malignancy). Diagnosis of carcinoma was made by histologic confirmation in all patients. Patients with overt distant metastasis on CT were not included. Fifteen of the 33 patients underwent pancreatic resection (two had margin-positive resection). Diagnosis of normal pancreas was confirmed by absence of pancreatic disease at 1-year or longer CT follow-up, 2-year or longer clinical follow-up, endoscopic sonography, or identification of nonpancreatic disease that caused symptoms or signs that prompted the initial CT.
CT Protocol
The dual-phase pancreas protocol with MDCT scanners included an optional
unenhanced scan, followed by dual-phase scans through the abdomen. The
scanning delays depended on the contrast injection rate: 45- and 70-second
delays at 3 mL/s injection, 40- and 65-second delays at 4 mL/s injection, and
35- and 60-second delays at 5 mL/s injection. A total of 150 mL of IV contrast
material was used (iohexol, Omnipaque 300, GE Healthcare). Slice thickness for
contrast-enhanced images was 2–4 mm in 62 patients (autoimmune
pancreatitis [n = 19], carcinoma [n = 29], normal controls
[n = 14]) and 5–7 mm in 12 patients (autoimmune pancreatitis
[n = 6], carcinoma [n = 4], normal controls [n =
2]). Single-detector helical CT and 4-, 8-, 16-, and 64-MDCT scanners of
different manufacturers were used in eight patients (autoimmune pancreatitis
[n = 4], carcinoma [n = 2], normal controls [n =
2]), 14 patients (autoimmune pancreatitis [n = 4], carcinoma
[n = 7], normal controls [n = 3]), 25 patients (autoimmune
pancreatitis [n = 6], carcinoma [n = 11], normal controls
[n = 8]), 26 patients (autoimmune pancreatitis [n = 10],
carcinoma [n = 13], normal controls [n = 3]), and one
patient (with autoimmune pancreatitis).
CT Analysis
CT scans were independently evaluated by three radiologists without
knowledge of the final diagnosis. They were aware of the nature of the study,
age and sex of the patients, and that the cohort included patients with
autoimmune pancreatitis, pancreatic carcinoma, and a normal pancreas. CT
images were reviewed on a workstation (Advantage Windows version 4.2, GE
Healthcare). Pancreatic size was subjectively categorized as normal or
enlarged by each segment (head, body, and tail). CT attenuation (enhancement)
of the pancreas was subjectively categorized as normal or decreased by each
segment in both pancreatic and hepatic phases. The presence or absence of a
low-attenuation mass was also evaluated. When a mass was present in a segment,
the attenuation of the mass was used as the attenuation of the segment. The
readers were instructed to take the age and sex of the patients and
enhancement of other structures into consideration to determine the pancreatic
size and attenuation. The presence of a capsule-like rim of soft tissue around
the pancreas, peripancreatic strands (linear soft-tissue structures in the
peripancreatic fat), pancreatic calcifications, and pseudocysts was recorded.
The pancreatic duct was evaluated for the presence of dilation and abrupt
cutoff. Vascular involvement (celiac, superior mesenteric, or splenic artery;
and main portal, superior mesenteric, or splenic vein) was also evaluated.
Arterial involvement was defined as the presence of tumor or soft tissue
abutting at least half the circumference of the vessel. Venous involvement was
considered present if the tumor or soft tissue deformed the vein. The bile
duct was evaluated for the presence of dilation or stent or circumferential
wall enhancement. The presence of retroperitoneal fibrosis and solid renal
lesions was also evaluated. These CT characteristics were chosen from
previously described findings of autoimmune pancreatitis and pancreatic
carcinoma
[5–7,
13,
14].
Finally, readers rated the probability of their diagnoses (autoimmune pancreatitis, carcinoma, normal) between 0 and 100% for each patient, synthesizing their confidence in the presence of multiple CT findings. The sum of probabilities of the three diagnoses was 100% for each patient.
Consensus of three readers for the presence or absence of the CT findings was determined as follows. Pancreatic enlargement, decreased CT attenuation by segment, and presence of a low-attenuation mass were considered to be present when agreed on by two or three readers. The morphologic pattern of pancreatic enlargement was categorized as diffusely enlarged, focally enlarged, normal in size or atrophied, and a low-attenuation mass present using the consensus findings. Diffusely decreased enhancement of the pancreas was defined as all three segments showing decreased CT attenuation by consensus on either the pancreatic or the hepatic phase. A capsule-like rim, peripancreatic strands, vascular involvement, and bile duct and pancreatic duct changes were considered present when agreed on by two or three readers. Pancreatic calcifications, peripancreatic cyst, retroperitoneal fibrosis, and solid renal lesions were considered present when agreed on by two or three readers, or if one of the three readers rated the item present and that rating was later confirmed by conjoint reevaluation by the three radiologists.
Statistical Analysis
Frequency of CT findings was calculated on the basis of the consensus of
the three readers and was compared between patients with autoimmune
pancreatitis and those with carcinoma using the chi-square or Fisher's exact
tests, as appropriate. A p value of less than 0.05 was considered
statistically significant. Interobserver agreements of three reviewers in
terms of assessment of various CT findings were quantified using the kappa
statistic. A kappa value of 0.20 or less indicated poor agreement;
0.21–0.40, fair agreement; 0.41–0.60, moderate agreement;
0.61–0.80, good agreement; and 0.81–1.00, very good agreement.
The probability of the diagnosis rated by each radiologist was compared with the final diagnosis. A correctly assigned probability of 70% or higher was considered correct, 30–69% as indeterminate, and less than 30% as incorrect. Accuracy of diagnosing autoimmune pancreatitis and pancreatic carcinoma and overall accuracy were calculated for each radiologist. For the calculation of accuracy, an indeterminate diagnosis was considered incorrect. Presumed explanations for incorrect and indeterminate diagnoses were evaluated.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
= 0.33–0.75),
presence of a capsule-like rim of soft tissue (p = 0.009, =
0.42–0.66), peripancreatic strands (p = 0.02, =
0.45–0.54), pancreatic calcification (p = 0.04, =
0.14–0.47), bile duct wall enhancement (p = 0.0001, =
0.28–0.47), and solid renal lesions (p = 0.002, =
0.32–74) were significantly more frequently seen in patients with
autoimmune pancreatitis (Figs.
1,
2,
3,
4A,
4B). Pancreatic duct dilation
(p = 0.001, = 0.65–0.73) and abrupt cutoff of the
dilated duct (p = 0.003, = 0.60–0.65) were
significantly more frequently seen in patients with carcinoma (Figs.
5 and
6).
|
|
|
|
|
|
|
|
|
Readers correctly identified 17, 17, and 19 cases of 25 cases of autoimmune pancreatitis (68%, 68%, and 76%), ranked three, four, and two cases as indeterminate (8–16%), and were incorrect in five, four, and four cases (16–20%). Correct diagnosis was made in 29, 29, and 30 (88–91%) of 33 carcinomas and 15, 14, and 14 (88–94%) of 16 normal controls. Overall accuracy was 82%, 81%, and 85%, respectively, in this cohort. Presumed explanations for the incorrect or indeterminate diagnoses are summarized in Table 3. The main explanations for an incorrect or indeterminate diagnosis were as follows: presence of biliary dilation without pancreatic enlargement or mass (four patients, nine interpretations) and presence of a low-attenuation mass (three patients, five interpretations) in patients with autoimmune pancreatitis were mistaken for carcinoma; and presence of a subtle low-attenuation or isoattenuating carcinoma (three patients, seven interpretations) was mistaken for autoimmune pancreatitis.
|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
In our study, the useful findings of autoimmune pancreatitis included a diffusely enlarged pancreas, diffusely decreased pancreatic enhancement, a capsule-like rim, bile duct wall enhancement, pancreatic calcifications, and the presence of solid renal lesions. These findings were more commonly associated with autoimmune pancreatitis and were uncommon in pancreatic carcinoma. Of the 25 patients with autoimmune pancreatitis, 21 patients had at least one of these findings. A capsule-like rim has been reported to be present in 16–80% of patients with autoimmune pancreatitis [5–7, 13]. In our study, a capsule-like rim was present in 40% of patients with autoimmune pancreatitis and was uncommon in patients with carcinoma, and it was identified with moderate to good interobserver agreement. However, in one case, a subtle low-attenuation mass in the peripheral pancreatic head was mistaken for a capsule-like rim and given indeterminate diagnosis by two readers. Peripancreatic strands have been reported to be present in 24–35% of patients with autoimmune pancreatitis [6, 13]. In our study, the finding was present in 60% of patients, but the findings were seen in 27% of patients with carcinoma. Pancreatic duct dilation and a cutoff sign were frequently seen in patients with carcinoma but were also seen in 24% and 16% of patients with autoimmune pancreatitis, respectively.
Extrapancreatic involvement is not rare in patients with autoimmune pancreatitis, and the presence of such findings can be used to aid in differentiating autoimmune pancreatitis from carcinoma. The most commonly involved organ is the bile duct, which causes intra- and extrahepatic biliary strictures, reported in 68–90% of patients [3, 6, 13, 16–19]. According to Yang et al. [13], CT showed bile duct narrowing in 18 of 20 (90%) patients, with enhancement of the narrowed segment in 8 of 20 (44%) patients. Our study showed that bile duct wall enhancement was present in 13 of 25 (52%) patients with autoimmune pancreatitis compared with two of 33 (6%) patients with pancreatic carcinoma. The finding can be used as a reliable sign of autoimmune pancreatitis despite the presence of a biliary stent in 11 of 25 patients with autoimmune pancreatitis and 16 of 33 patients with carcinoma at the time of CT. Another important extrapancreatic feature of autoimmune pancreatitis is renal involvement [14, 20–22], which has been reported to occur in as many as 30% of patients [14]. Renal parenchymal lesions may appear as small peripheral cortical nodules, round or wedge-shaped lesions, or diffuse patchy involvement. In our study, 28% of patients with autoimmune pancreatitis had renal involvement, which can be used as a reliable secondary sign of autoimmune pancreatitis. Retroperitoneal fibrosis has been reported to occur in 3–11% of patients [6, 14, 23, 24]. In our study, the frequency of the finding was not statistically significant between autoimmune pancreatitis and carcinoma. This result was probably due to a small patient population.
Vascular involvement in patients with autoimmune pancreatitis was found in 54% at angiography [25] but was reported to be uncommon on CT [6]. Our study showed that vascular involvement in patients with autoimmune pancreatitis was relatively common on CT, with a frequency similar to that of carcinoma. Pancreatic pseudocyst and calcification of the pancreas were seen in 12% and 32% of patients with autoimmune pancreatitis in our study. These findings were uncommon on CT in one study [6]. Although such findings are more often associated with alcohol-induced chronic pancreatitis [8], the presence of pseudocysts or calcifications should not be used to exclude the possibility of autoimmune pancreatitis. Pancreatic calcifications seen in patients with autoimmune pancreatitis were typically one or two small punctate calcifications.
Our study showed that dual-phase CT of the pancreas is moderately accurate in diagnosing autoimmune pancreatitis. A correct diagnosis was made in 68–76% patients with autoimmune pancreatitis. At least one reader made incorrect or indeterminate diagnoses in 11 of 25 patients with autoimmune pancreatitis and in six of 33 patients with carcinoma. Biliary dilation with a normal-sized pancreas without a mass (n = 4) and the presence of a low-attenuation mass or focally enlarged pancreatic segment (n = 4) in patients with autoimmune pancreatitis were the most common presumed explanations for the incorrect or indeterminate diagnoses of carcinoma. Subtle low-attenuation or isoattenuation carcinoma (n = 3) was the most common presumed explanation for the incorrect or indeterminate diagnoses of autoimmune pancreatitis.
Procacci et al. [26] previously evaluated the performance of CT for the diagnosis of autoimmune pancreatitis. In their study, seven patients with autoimmune pancreatitis and 20 patients with other diseases (four with acute pancreatitis, 10 with chronic pancreatitis, and six with carcinoma) were included. Six of seven (86%) patients with autoimmune pancreatitis were correctly diagnosed, and one of 20 patients with other diseases was incorrectly diagnosed as having autoimmune pancreatitis, for an overall accuracy of 93%. The difference in the accuracy from our study probably resulted from the difference in the patient population studied. Only patients who underwent dual-phase CT of the pancreas at our institution were included in the study. Patients referred to our institution with an outside CT examination showing characteristic findings of autoimmune pancreatitis or obvious unresectable carcinoma often did not have CT repeated. We also excluded patients with obvious distant or peritoneal metastases. This might have favored the inclusion of more atypical cases of autoimmune pancreatitis (with a low-attenuation mass [n = 4, 16%] or normal-size pancreas [n = 9; 36%]) and more cases of carcinoma in an early stage (isoattenuating carcinoma [n = 5; 15%] and resectable carcinoma [n = 15; 45%]).
A limitation of our study is its retrospective nature. Only three possible diagnoses were allowed in the cohort, and the prevalence of autoimmune pancreatitis was much higher than in a clinical setting. CT protocols were variable (scan thickness and number of detector rows). It is unclear how the variability in the CT protocol affected our results.
Our study suggests the CT findings useful for the diagnosis of autoimmune pancreatitis include a diffusely enlarged pancreas, diffusely decreased pancreatic enhancement, a capsule-like rim, pancreatic calcifications, bile duct wall enhancement, and the presence of solid renal lesions. CT is moderately accurate in the diagnosis of autoimmune pancreatitis. However, a focal form of autoimmune pancreatitis is often difficult to differentiate from carcinoma.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  K. D. Bodily, N. Takahashi, J. G. Fletcher, J. L. Fidler, D. M. Hough, A. Kawashima, and S. T. Chari Autoimmune Pancreatitis: Pancreatic and Extrapancreatic Imaging Findings Am. J. Roentgenol., February 1, 2009; 192(2): 431 - 437. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S-H Moon, M-H Kim, D H Park, C Y Hwang, S J Park, S S Lee, D W Seo, and S K Lee Is a 2-week steroid trial after initial negative investigation for malignancy useful in differentiating autoimmune pancreatitis from pancreatic cancer? A prospective outcome study Gut, December 1, 2008; 57(12): 1704 - 1712. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
