<Patients>

●　We retrospectively recruited consecutive patients with AIP in multiphase contrast enhanced CT database from July 2003 to May 2010. The consecutive patients with pancreatic carcinoma and usual chronic pancreatitis were also selected in adjustment with the same ratio of the CT systems (16- or 64-slice multisection CT) in the same period. In recruiting these patients, sex and age were not considered.

●　The present study examined 55 patients with AIP (5 female and 50 male; age range 24-86 years, mean age 64 years), 50 patients with pancreatic carcinoma (19 female and 31 male; age range 46-84 years, mean age 66 years), and 50 patients with usual chronic pancreatitis (10 female and 40 male; age range 31-78 years, mean age 59 years).

<Inclusion criteria>

AIP

●　Those who met at least one of the three commonly used definite diagnostic criteria for AIP (Mayo Clinic’s HISORt criteria, revised Japanese Pancreas Society criteria, and Korean diagnostic criteria) [2,8,9]

Pancreatic carcinoma (all required)

● 　Histopathologically proven pancreatic carcinoma on surgical specimen

● 　No previous history of endoscopic retrograde cholangio pancreatography (ERCP) within 2 weeks before CT examination

Chronic pancreatitis (all required)

●　Those who met the diagnostic criteria for chronic pancreatitis (Japanese clinical diagnostic criteria for chronic pancreatitis 2009) [10]

● 　No previous history of endoscopic retrograde cholangio pancreatography (ERCP) within 2 weeks before CT examination

<CT image acquisition>

●　CT was performed using 16-slice (AIP; n=14, pancreatic carcinoma; n=13, chronic pancreatitis; n=13) and 64-slice multisection CT systems (AIP; n=41, pancreatic carcinoma; n=37, chronic pancreatitis; n=37) (Aquilion, Toshiba Medical Systems, Tokyo, Japan). Non-ionic contrast material (2-2.5ml/kg) with an iodine concentration of 300mg I/ml was injected through the peripheral venous line over a period of 30s (with an upper limit of 5 ml/s), and a 5% dextrose or saline flush was injected at a fixed rate of 5ml/s over 6 seconds immediately after contrast material injection. After unenhanced images were acquired, all patients underwent pancreatic phase and portal venous phase imaging. Delayed phase images were also acquired (AIP; n=45, pancreatic carcinoma; n=47, chronic pancreatitis; n=33). Individualized scan delays were determined using the automatic bolus-tracking method (SureStart, Toshiba). At our institution, average scan delays from the injection of contrast material to the start of pancreatic phase, portal venous phase, and delayed phase imaging were 40, 65, and 240 s, respectively, for 16-channel multisection CT; and 44, 70, and 240 s for 64-channel multisection CT.

●　Unenhanced images and contrast-enhanced images were reconstructed with a 5 mm section thickness at 5 mm intervals and a 2 mm section thickness at 2 mm intervals, respectively. In addition, reconstruction was performed with a 1 mm section thickness at 0.5 or 0.8 mm intervals for all contrast enhanced images. Coronal reformatted images were generated with a 2 mm section thickness at 2 mm intervals.

<CT image analysis>

●　Two radiologists (14 and 24 years of experience in the interpretation of abdominal images) blinded to the diagnosis and other examination findings independently evaluated whether or not the pancreatic duct wall showed intense enhancement at the non-dilatated main pancreatic duct (< 2 mm in diameter) in randomly ordered 155 patients (AIP; n = 55, pancreatic carcinoma; n = 50, and chronic pancreatitis; n = 50). In cases of disagreement between the two readers, a consensus decision was made by them. The frequency of the finding was compared among AIP, pancreatic carcinoma, and chronic pancreatitis. And the accuracy, sensitivity and specificity of the finding in AIP were calculated.

In patients demonstrating enhanced MPD sign, two radiologists evaluated additional findings in consensus.

●　With regard to the enhanced MPD and the surrounding pancreatic parenchyma, the following CT findings were assessed:

(a) Location of the enhanced MPD (head, body, and/or tail of the pancreas)

(b) Form of the enhanced MPD (solid line, dashed line)

(c) Thickness of the enhanced MPD wall

(d) Contrast-enhanced phase demonstrating the enhanced MPD (pancreatic phase, portal venous phase, and/or delayed phase)

(e) Visualization of the MPD lumen within enhanced MPD sign (completely unvisible, partially unvisible, or visible)

(f) Degree of the contrast-enhancement of pancreatic parenchyma surrounding the enhanced MPD (hypoattenuation, iso-attenuation, or hyperattenuation)

<Statistical Analyses>　

●　For data analysis, interobserver agreement was evaluated using kappa statistics. Associations were considered significant if P values are 0.05 or less. The differences between the groups were evaluated by using the Χ² test. Statistical analyses were performed using PASW statics 18 software (SPSS Inc, Chicago Ill).