| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Gastrointestinal Imaging |
1 From the Department of Radiology, Boston University Medical Center, One Boston Medical Center Place, Boston, MA 02118. Received February 23, 2004; revision requested April 15; revision received June 2; accepted July 1. Address correspondence to J.A.S. (e-mail: jorge.soto@bmc.org).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
MATERIALS AND METHODS: This study was approved by the investigational review board. Seventy-seven patients (46 men and 31 women; mean age, 51 years) underwent CT with a four–detector row scanner and endoscopic retrograde pancreatography (ERP). Section thickness was 3.2 mm, and the reconstruction interval was 3 mm. Two radiologists independently evaluated the CT data sets with picture archiving and communication system (PACS) workstations equipped with software for two- and three-dimensional postprocessing reformations; the radiologists were blinded to the clinical and ERP data. Pancreas divisum was diagnosed at CT if what the authors termed the "dominant dorsal duct sign" (the caliber of the dorsal duct was larger than that of the ventral duct) was present and if the dorsal and ventral ducts did not appear to communicate with each other at cine review of images. ERP findings were used as the standard of reference for determining the performance (sensitivity, specificity, positive and negative predictive values) of the radiologists’ CT interpretations. Interobserver agreement was measured by using 
statistics.
RESULTS: For four of the 77 patients (5%), both radiologists considered that depiction of the pancreatic duct on CT images was not sufficient to enable evaluation of ductal anatomy. These patients were excluded from further analysis. In the remaining 73 patients, ERP demonstrated pancreas divisum in 10 (14%); both observers made the correct diagnosis in nine of these patients. In addition, one radiologist had one false-positive interpretation, whereas the other radiologist had two false-positive interpretations. Thus, for observer 1, the calculated sensitivity was 90% (95% confidence interval [CI], 60%–98%) and the specificity was 98% (95% CI, 91%–100%). For observer 2, sensitivity was 90% (95% CI, 60%–98%) and specificity was 97% (95% CI, 89%–99%). Interobserver agreement was excellent ( = 0.93).
CONCLUSION: CT scans obtained with multi–detector row scanners and interpreted with PACS workstations enable depiction of pancreas divisum. This assessment is possible only when the pancreatic duct is visualized.
© RSNA, 2005
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
The definitive diagnosis of pancreas divisum is made with endoscopic retrograde pancreatography (ERP) (8,9). Magnetic resonance (MR) pancreatography has also been shown to be highly sensitive and specific for depicting pancreas divisum (10–13). MR imaging demonstrates the dorsal duct draining independently into the duodenum, superior and anterior to the ventral duct and common bile duct.
Although MR pancreatography may be useful for confirming the diagnosis when pancreas divisum is suspected, patients with unexplained abdominal pain or other symptoms of possible pancreatic origin are more likely to be studied initially with computed tomography (CT). CT, however, has not been shown to be accurate in the diagnosis of pancreas divisum (14,15). In the past, abdominal CT examinations performed with single–detector row and conventional CT scanners typically involved section collimations of 5–10 mm (14,15). With the advent of multi–detector row CT scanners, the use of thinner section collimation is now routine. The higher spatial resolution of these thin-collimation images, together with cine review of images on picture archiving and communication system (PACS) workstations affords better depiction of the anatomy of small tubular structures such as the pancreatic duct. To our knowledge, no recent studies in the radiology literature have evaluated whether this newer technology provides improved visualization of usual and variant pancreatic ductal anatomy. Thus, the purpose of this study was to retrospectively evaluate contrast material–enhanced multi–detector row CT in the depiction of pancreas divisum.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
CT Technique
All CT studies were performed with a four–detector row scanner (MX8000; Philips Medical Systems, Andover, Mass). Scans were acquired in a craniocaudal direction with the following parameters: section thickness, 3.2 mm; reconstruction interval, 3 mm; pitch, 6; tube current, 120 kVp; and 200–300 mAs. One hundred milliliters of nonionic contrast material (iohexol, Optiray 320 [milligrams of iodine per milliliter]; Mallinckrodt Imaging, Hazelwood, Mo) was injected intravenously at a rate of 3–4 mL/sec by using a power injector. The delay between the beginning of contrast material injection and image acquisition was 40 seconds for patients in whom the indication for the CT examination was suspected pancreatic disease (n = 46) and 60 seconds for patients with other indications (n = 31). At our institution, the decision to administer oral contrast material for CT of the upper abdomen varies with the specific indication, general condition of the patient, and presence of contraindications. Forty-four patients received oral contrast material (900 mL of 2.2% barium sulfate suspension [Medescan barium sulfate; Lafayette Pharmaceuticals, Lafayette, Ind]) before undergoing CT. Among these 44 patients, CT was performed to evaluate pancreatitis (n = 18), suspected biliary obstruction (n = 10), suspected choledocholithiasis (n = 2), known or suspected pancreatic tumor (n = 6), abdominal pain (n = 6), or "other" (n = 2).
Image Analysis
The CT data sets were transferred to PACS workstations (Aurora, software 6.5; Merge eFilm, Milwaukee, Wis) for analysis. CT scans were retrospectively interpreted by two radiologists (J.A.S. [observer 1] and B.C.L. [observer 2]) with 9 and 3 years of experience, respectively, in cross-sectional abdominal imaging. The radiologists interpreted the CT scans independently and without knowledge of clinical data or results of ERP. For interpretation, the radiologists used the scroll and cine mode functions available with the PACS workstations. These functions enable tracing of the pancreatic duct from the tail of the gland to the head and demonstration of the site of drainage into the duodenum.
In addition, the radiologists had the option of generating multiplanar reformations of the transverse images by using a postprocessing software package incorporated into the PACS workstation (Voxar 3D; Voxar, Framingham, Mass). This software package enables immediate, online, two- and three-dimensional reformations of data sets, which are helpful for illustrating the course of the duct through the gland and the relationship between the dorsal and ventral segments. Options for multiplanar renderings included orthogonal, oblique, and curved-linear planes. Although this option was available for all patients, the radiologists were not asked to document how often they used it or how useful they found it. Instead, we assumed that PACS-based postprocessing is part of the interpretation of the examination results and should be used freely when necessary.
Each radiologist initially reviewed the transverse CT data sets and determined whether visualization of the pancreatic duct (or ducts) was sufficient for establishing the site of drainage into the duodenum and ductal anatomy. If the duct was not adequately visualized, the radiologists were asked to provide a possible reason. The CT data sets were then evaluated for pancreatic ductal anatomy. Pancreas divisum was considered to be present if the dorsal duct could be traced from the tail and body of the gland through the anterior aspect of the head and could be seen draining into the minor papilla, which is located anterior to the common bile duct and major papilla. The ventral segment of the pancreatic duct, located in the posterior region of the head of the pancreas and draining into the duodenum together with the common bile duct, may or may not be seen. If seen, however, there should be no apparent communication between the two ducts, and the caliber of the ventral duct must be less than that of the dorsal duct. We called this the "dominant dorsal duct sign." In patients with CT findings suggestive of pancreas divisum, the radiologists recorded whether the ventral duct was visualized or not.
For well-visualized ducts, the caliber was assessed subjectively as being normal or dilated. For patients suspected of having pancreas divisum, the assessment of duct caliber comprised both the dorsal and ventral ductal systems. Differences in opinion between the two radiologists regarding duct caliber were settled by consensus. The following additional findings related to the pancreatic ducts and glandular parenchyma were also recorded, if present: calcifications, duct stricture (defined as focal narrowing with proximal dilatation), pseudocysts, focal glandular enlargement, and a fatty cleft separating the dorsal and ventral portions of the pancreas.
Standard of Reference
ERP findings were used as the standard of reference for determining pancreatic ductal anatomy and caliber. ERP was performed by one of two gastroenterologists at our institution by using a standard technique and iohexol (Optiray 320; Mallinckrodt Imaging, Hazelwood, Mo). The gastroenterologists had 10 and 3 years of experience in performing ERP examinations. Although the gastroenterologists knew the patients’ clinical information, they were not aware of the fact that the results of ERP would be possibly used for this investigation because our study was retrospective. Thus, the decision to inject contrast material into the pancreatic duct and the aggressiveness of duct injection were guided by the specific clinical situation of each patient and were not influenced by this investigation. A gastrointestinal radiologist retrospectively interpreted all direct pancreatograms; the radiologist had more than 20 years of experience in interpreting cholangiopancreatograms and was not involved in the evaluation of CT images. Initially, the radiologist assessed the quality of the pancreatograms. Patients for whom pancreatographic image quality was insufficient to allow a firm determination of ductal anatomy were excluded from further analysis.
At ERP, pancreas divisum was diagnosed on the basis of cannulation of the minor papilla and filling with contrast material of the dorsal duct, cannulation of the major papilla and filling with contrast material of the typical ramifying ("treelike") (4) ventral duct system, or cannulation of both papillae and filling with contrast material of both duct systems. The radiologist also assessed the pancreatograms for changes consistent with chronic pancreatitis: duct dilatation (subjectively determined), stricture (focal narrowing with proximal dilatation), beaded appearance, filling defects, calcifications, and communication with pseudocysts.
Clinical Information
To establish the relevance of finding pancreas divisum in our patient population, one investigator (J.A.S., B.C.L., or J.W.S.) reviewed the charts of the subgroup of patients in whom pancreas divisum was confirmed at ERP. The following data were collected: history of pancreatitis, history of chronic unexplained abdominal pain, prior pancreatic surgery or trauma, and known tumor (primary pancreatic or extrapancreatic).
Statistical Analysis
For the detection of pancreas divisum with CT, we calculated the sensitivity, specificity, and positive and negative predictive values (along with 95% confidence intervals [CIs]) for each of the two observers. In addition, we measured interobserver agreement by using statistics (16,17). According to the value, strength of agreement was classified as follows (10): slight ( value, 0–0.20), fair ( value, 0.21–0.40), moderate ( value, 0.41–0.60), very good ( value, 0.61–0.80), and excellent ( value, 0.81–1.00). We did not test for statistically significant differences between the mean age of the men and that of the women.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
Pancreas divisum was diagnosed with ERP in 10 of the 73 patients (14%). The mean age of these 10 patients was 46 years (range, 22–64 years). Seven of these patients were men and three were women. In eight of these 10 patients, only the major papilla was cannulated and the ventral duct injected. In one patient, both papillae were cannulated and both ducts injected. Finally, in one patient, only the minor papilla was cannulated and the dorsal duct injected with contrast material because a leak was suspected in the tail of the gland. Pancreas divisum was confirmed in this patient.
Both observers correctly identified the anatomic variant in nine of the 10 patients in whom pancreas divisum was confirmed at ERP. In seven of these patients, both ducts (dorsal and ventral) were identified at CT (Figs 1, 2); in the remaining two patients, only the dorsal duct was seen (Fig 3). Five of the 10 patients with pancreas divisum had evidence—including pancreatic ductal dilatation (n = 5) (Figs 1, 2), focal stricture (n = 1), and calcifications (n = 1)—of chronic pancreatitis at ERP. All of these findings were also seen at CT (Figs 1, 2). In four patients, the dorsal and ventral ducts were both dilated (Figs 1, 2), and in one patient, ductal dilatation was limited to the dorsal system. The remaining five patients had normal-caliber ducts (Fig 3). In no case did the CT scans in patients with ERP-proved pancreas divisum demonstrate focal enlargement of the head of the pancreas or a fatty cleft separating the dorsal and ventral pancreas. In one patient with pancreas divisum that was demonstrated at ERP, the pancreatic duct in the body and tail of the gland was not completely depicted on the CT images, and both interpreters failed to make the diagnosis (Fig 4). This was considered a false-negative interpretation of the CT results.
|
|
|
|
|
|
|
|
|
Statistical Results
According to the above results, observer 1 had nine true-positive, one false-negative, one false-positive, and 62 true-negative interpretations. Observer 2 had nine true-positive, one false-negative, two false-positive, and 61 true-negative interpretations. The calculated diagnostic performance parameters for observer 1 were as follows: sensitivity, 90% (95% CI, 60%–98%); specificity, 98% (95% CI, 91%–100%); positive predictive value, 90% (95% CI, 60%–98%); and negative predictive value, 98% (95% CI, 91%–100%). For observer 2, the calculated diagnostic performance parameters were as follows: sensitivity, 90% (95% CI, 60%–98%); specificity, 97% (95% CI, 89%–99%); positive predictive value, 82% (95% CI, 71%–90%); and negative predictive value, 98% (95% CI, 91%–199%). Interobserver agreement was excellent ( = 0.93).
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
Traditionally, ERP has been regarded as the preferred method for detecting pancreas divisum. At ERP, the diagnosis of pancreas divisum can be established by injecting contrast material into the ventral duct through the major papilla, by injecting into the dorsal duct through the minor papilla, or by injecting into both ducts. The ventral duct is usually short and branches off in the head of the gland in a pattern that has been termed treelike (4). Alternatively, the ventral duct may be atretic and almost nonfunctional. The dorsal duct is typically larger in caliber, an indication of its dominant function in drainage of pancreatic fluid. Despite its high accuracy for determining pancreatic ductal anatomy, ERP is an invasive test with a potential for substantial procedure-related complications. Pancreatitis may occur in up to 5% of cases (19–21); other, less frequent, complications include duodenal perforation and gastrointestinal bleeding (19).
MR pancreatography has been shown to be highly accurate in the diagnosis of pancreas divisum (10,11,22). Heavily T2-weighted MR images enable demonstration of the noncommunicating dorsal and ventral ducts as well as the independent drainage sites. When the diagnosis is suspected, the administration of exogenous secretin may increase even further the sensitivity of and diagnostic confidence in MR pancreatography (12,13). The same concept has been used to improve visualization of the pancreatic duct with ultrasonography (23) and could be of potential use for improving duct visualization at CT.
Despite the controversy surrounding the importance of pancreas divisum in clinical practice, the fact remains that patients with nonspecific or recurrent abdominal pain are studied with contrast-enhanced CT. Likewise, patients with elevated amylase levels or other clinical evidence suggestive of pancreatitis often undergo CT as the initial imaging test. Thus, information about the pancreatic duct gathered from the CT scans is useful for deciding how to further evaluate this particular subgroup of patients.
In general, CT has been considered unreliable for demonstrating the fine anatomy of the pancreatic ductal system (24). Few studies have specifically evaluated the performance of CT in the detection of pancreas divisum. Zeman et al (14) used thin-section CT and identified only five of 12 patients (42%) with a confirmed diagnosis. However, that study (published in 1988) involved nonhelical CT techniques with 5-mm collimation and 5- or 10-mm intersection gaps; hard copies were used for image interpretation. In their study, the most useful CT sign of pancreas divisum was visualization of the dual and nonfused ducts. Occasionally, a fatty cleft separating the dorsal and ventral glandular parenchyma was seen (14).
In another study that included 29 patients with pancreas divisum (15), none of the patients was noted to have such a cleft. Another finding that has been noted in CT studies of patients with pancreas divisum is focal enlargement of the head of the pancreas without a visible mass (25). Results of our study confirm the lack of reliability of these signs; the CT images did not demonstrate a fatty cleft or focal head enlargement in any of our patients.
In this study, we evaluated the utility of contrast-enhanced CT with multi–detector row technology, soft-copy interpretation with scroll and cine functions, and full postprocessing capabilities. Current technology enables the acquisition of thin-section CT images in a single breath hold, thereby decreasing the possibility of section misregistration and exclusion of portions of organs owing to variable breathing patterns. Image interpretation on PACS workstations, especially when the scroll function is used, improves the ability to visualize and follow small tubular structures through contiguous images. Finally, the possibility of performing multiplanar reformations of the transverse data sets may aid in the display of ductal anatomy and improve diagnostic confidence because demonstration of the spatial relationship of the ducts (as compared with each other) is improved. As technologic improvements continue, it is expected that PACS workstations will incorporate postprocessing software (like that used in our study).
We believe that these factors explain our improved results as compared with those of prior studies (14,15). Ductal anatomy was demonstrated on the CT scans in most patients included in our study. In patients with normal (typical) anatomy, the duct curves inferiorly and posteriorly at the neck of the pancreas toward its drainage site in the posterior and lateral aspect. The duct of Santorini may be seen as a small tubular structure connecting the main duct with the minor papilla. If seen, however, the caliber of the duct of Santorini is usually less than that of the main duct in the head of the gland. In pancreas divisum, the dorsal duct remains in the anterior aspect of the gland and drains directly into the minor papilla. The smaller ventral duct may or may not be visualized. This is the dominant dorsal duct sign, which we found to be highly suggestive of pancreas divisum in our study.
In our study, there was a single false-negative interpretation of one data set by both observers. In this patient, the pancreatic duct was not depicted well on the CT images. As technology evolves, the performance of CT in the detection of pancreas divisum (and other ductal abnormalities) may improve even further. CT scanners with 16, 32, and more detector rows are now available. These scanners enable acquisition of images with 1-mm collimation or less. The higher spatial resolution of these isotropic voxel data sets will undoubtedly facilitate even further the evaluation of tubular structures in the body (eg, the pancreatic duct). The quality of two- and three-dimensional reformations of these data sets is even better than those used in our study.
This study had several limitations. First, the total number of patients in whom pancreas divisum was confirmed was small. Because we included only patients with good-quality confirmatory pancreatograms, the prevalence of pancreas divisum was higher than what would be expected in the general population. The results of this study may be difficult to reproduce in a larger population of patients referred to undergo CT for other indications. The results of this study, however, suggest that careful attention to this type of fine detail may help detect this anatomic variant, either as an incidental finding in patients who undergo scanning for unrelated reasons or as a contributing or causal factor in patients evaluated for unexplained abdominal pain or idiopathic pancreatitis. A minority of the patients included in our study underwent pancreatography of the dorsal duct to confirm the diagnosis of pancreas divisum. Ideally, a complete pancreatogram requires injection into the dorsal and ventral ducts.
Although other conditions (eg, pancreatic tumors, postoperative changes, and trauma) may result in pancreatographic findings that mimic the changes of pancreas divisum, none of the 10 patients with pancreas divisum included in our study had any of these that explained the ERP findings. Another limitation was that we did not formally evaluate the value of adding the multiplanar reformations to the cine review of transverse images. Curved-linear (and other multiplanar) reformations are useful for illustrating and confirming findings that have already been noted at a review of the transverse source images, but they do not necessarily add diagnostic information. Reliance on multiplanar reformations exclusively may be misleading because the plane is determined by the operator, and false apparent relationships between tubular structures may be visualized.
In summary, the results of our study show that CT scans obtained with multi–detector row scanners and interpreted with PACS workstations may depict pancreas divisum. This assessment, however, is possible only when the pancreatic duct is visualized. Multiple factors related to advances in CT technology, as well as in equipment for image display and interpretation, are responsible for this improved depiction of ductal anatomy. Although the clinical importance of pancreas divisum remains controversial, increased awareness of pancreatic anatomy at CT may help in the selection of patients who should be further evaluated with MR pancreatography and aid in the planning of endoscopic diagnostic and therapeutic procedures.
|
FOOTNOTES |
|---|
Authors stated no financial relationship to disclose.
Author contributions: Guarantor of integrity of entire study, J.A.S.; study concepts, J.A.S., B.C.L.; study design, J.A.S.; literature research, J.A.S.; clinical studies, J.A.S., B.C.L., J.W.S.; data acquisition, J.A.S., B.C.L., J.W.S.; data analysis/interpretation, J.A.S., B.C.L.; statistical analysis, J.A.S., B.C.L.; manuscript preparation, definition of intellectual content, revision/review, and final version approval, J.A.S., B.C.L., J.W.S.; manuscript editing, J.A.S., B.C.L.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  H. Ogawa, S. Itoh, M. Ikeda, K. Suzuki, and S. Naganawa Intraductal Papillary Mucinous Neoplasm of the Pancreas: Assessment of the Likelihood of Invasiveness with Multisection CT Radiology, September 1, 2008; 248(3): 876 - 886. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Yu, M. A. Turner, A. S. Fulcher, and R. A. Halvorsen Congenital anomalies and normal variants of the pancreaticobiliary tract and the pancreas in adults: part 2, Pancreatic duct and pancreas. Am. J. Roentgenol., December 1, 2006; 187(6): 1544 - 1553. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Z M Zhou, C H Fang, L W Huang, S Z Zhong, B L Wang, and W Y Zhou Three dimensional reconstruction of the pancreas based on the virtual Chinese human--female number 1. Postgrad. Med. J., June 1, 2006; 82(968): 392 - 396. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. J. Mortele, T. C. Rocha, J. L. Streeter, and A. J. Taylor Multimodality Imaging of Pancreatic and Biliary Congenital Anomalies RadioGraphics, May 1, 2006; 26(3): 715 - 731. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
