| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Gastrointestinal Imaging |
1 From the Department of Radiology, Thomas Jefferson University Hospital, 1096 Main Bldg, 132 S 10th St, Philadelphia, PA 19107 (X.M.Z., D.G.M., M.D., G.A.H., L.P.); and Department of Radiology, Affiliated Hospital of North Sichuan Medical College, Sichuan, People’s Republic of China (X.M.Z.). Received April 19, 2001; revision requested May 23; revision received August 10; accepted September 28. Address correspondence to D.G.M. (e-mail: donald.mitchell@mail.tju.edu).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
MATERIALS AND METHODS: Single-shot fast SE images of the pancreas were obtained in 1,444 patients. The images were reviewed for presence of pancreatic cysts, which were classified as simple and nonsimple types, and cyst diameters were measured. 
statistic, Fisher exact, McNemar Q, and Pearson product moment correlation tests were performed.
RESULTS: Two hundred eighty-three (19.6%) patients had at least one pancreatic cyst. The prevalence of pancreatic cysts increased with age (r = 0.96). The percentages of male and female patients with pancreatic cysts (20.4% vs 18.8%) were not significantly different. Two hundred seventy-one (18.8%) patients had simple cysts, and 147 (10.2%) had nonsimple cysts. Of 283 patients with pancreatic cysts, 158 (55.8%) had only one pancreatic cyst. The number of patients with multiple cysts increased with age after 70 years. Four hundred fifteen (83.8%) cysts were 10 mm in diameter or smaller; 56 (11.3%), 11–20 mm in diameter; and 24 (4.9%), 21 mm in diameter or larger. Sixteen (5.7%) patients with pancreatic cysts had malignant pancreatic tumors, and 75 (26.5%) patients had pancreatitis.
CONCLUSION: The prevalence of pancreatic cysts at single-shot fast SE MR imaging—especially cysts with a diameter smaller than 10 mm—is similar to that of pancreatic cysts at autopsy and higher than that of pancreatic cysts at transabdominal ultrasonography. Prevalence is especially high in patients with pancreatitis.
© RSNA, 2002
Index terms: Pancreas, cysts, 770.312 • Pancreas, MR, 770.121411, 770.121416, 770.12143 • Pancreas, neoplasms, 770.312, 770.321 • Pancreatitis, 770.291
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
The single-shot fast spin-echo (SE) magnetic resonance (MR) imaging sequence used with the half-Fourier method has enabled faster imaging of the pancreaticobiliary tree with minimal motion artifact (5). It has also been effective in depicting other fluid-filled structures, such as cysts (6).
We have observed that small pancreatic cysts are depicted on single-shot fast SE images more commonly than has been previously reported (1). The purpose of our study was to evaluate the single-shot fast SE MR imaging sequence for the depiction of pancreatic cysts in a large number of patients and to analyze cyst prevalence with respect to patient age and sex and other clinical information. Herein we report our experiences in this evaluation.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
The clinical indication for the MR examination often was not apparent at our retrospective review. Most of the 323 patients who underwent MR cholangiopancreatography were examined for determination of the cause of jaundice or biliary obstruction, evaluation of acute or chronic pancreatitis, or evaluation of indeterminate findings of the pancreas noted at prior computed tomographic (CT) or ultrasonographic (US) examination. Most of the 1,121 patients who did not undergo MR cholangiopancreatography were examined for reasons unrelated to the pancreas or the biliary system, such as suspected hepatic mass, diffuse liver disease, or evaluation of other abdominal organs.
MR Imaging Technique
All MR examinations had been performed by using a 1.5-T system (Signa Horizon; GE Medical Systems, Milwaukee, Wis) with a phased-array torso coil. Coronal T2-weighted single-shot fast SE MR images were obtained during two or more breath holds with the following parameters: 
/90–100 or 180–190 (repetition time msec/echo time msec), 5–7-mm section thickness, no intersection gap, 256 x 192 matrix, a half signal acquired, and 32 x 32-cm field of view. Transverse T2-weighted single-shot fast SE MR images were obtained during one or two breath holds with the following parameters: /90–100 or 180–190, 5-mm section thickness, no intersection gap, 256 x 160 matrix, a half signal acquired, and 32 x 24-cm field of view. Radial oblique slab single-shot fast SE MR images were obtained for MR cholangiopancreatography with the following parameters: /700, fat saturation, 40-mm section thickness, 256 x 192 matrix, a half signal acquired, and 24 x 24-cm field of view.
Additional MR imaging sequences, such as transverse T2-weighted fast SE, two-dimensional T1-weighted spoiled gradient-echo both in and out of phase, and dynamic contrast material–enhanced three-dimensional spoiled gradient-echo, also were performed, but only the single-shot fast SE images were used to count cysts.
Image Interpretation
The original MR imaging data were loaded onto a computer workstation (Cannon; Sun Microsystems, Palo Alto, Calif) and viewed by using proprietary software (Rational Imaging; Intuitive Software, West Hills, Calif). Two radiologists (X.M.Z., M.D.) with more than 6 years of body MR imaging experience retrospectively and individually reviewed the coronal and transverse T2-weighted single-shot fast SE MR images and MR cholangiopancreatographic images (if obtained) to find pancreatic cysts. When more than one MR imaging examination was performed in a patient, only the most recent study was analyzed.
Pancreatic cysts were divided into two groups: simple cysts and nonsimple cysts. A simple pancreatic cyst was considered to be round or oval without lobules or septa and sharply demarcated from the surrounding parenchyma, have a smooth thin wall, and have homogeneous signal intensity similar to that of simple fluid. A cyst with any other features was considered to be a nonsimple cyst. We distinguished cysts from the pancreatic duct by paging through the images to separately identify the pancreatic duct. The cysts were also classified according to their diameter, their number, and patient age and sex. In this study, we defined a small cyst as that having a diameter of less than 10 mm.
On both coronal and transverse images, we differentiated the three parts of the pancreas according to the conventional anatomic classification system (7). The head of the pancreas, including the uncinate process, was distinguished from the body of the pancreas by the sagittal plane defined by the superior mesenteric artery. The tail of the pancreas was defined as the part of pancreas that extended into the splenic hilum.
The pancreatic duct was noted as being normal, narrow, or dilated, or as communicating with a cyst. The pancreatic duct was considered to be dilated when its diameter was greater than 3 mm. The duct was considered to be narrow only when there was dilatation peripheral to an area of narrowing. If the pancreas showed cysts, associated disease, including pancreatic mass, pancreatitis, extrapancreatic pseudocyst, liver cyst, and/or renal cyst, was noted and the medical record was reviewed for a history of autosomal dominant polycystic kidney disease (ADPKD), von Hippel–Lindau disease, cystic fibrosis, and/or pancreatic cancer.
Statistical Analyses
Results were documented as the mean of the values obtained by the two observers plus or minus the 95% CI. statistics were calculated to measure agreement between the observers’ findings at qualitative analysis. Agreement was interpreted as follows: value greater than 0.81, very good agreement; value of 0.61–0.80, good agreement; value of 0.41–0.60, moderate agreement; and value lower than 0.41, poor agreement. Nonparametric tests were performed to analyze differences between groups; the 
2 test and Fisher exact test (one cell with a value < 5) were used to analyze differences related to, respectively, the sex and age of the patients. A P value of less than .05 indicated a statistically significant difference.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
= 0.68) and moderate with regard to the prevalence of small pancreatic cysts ( = 0.58). Also, there were no significant differences between the two radiologists with regard to total number of cysts (479 vs 510; 2 = 0.72; P > .05) and number of small cysts (414 vs 415; 2 = 0.001; P > .05).
Two hundred eighty-three (19.6%; 95% CI: 18.8%, 20.4%) of the 1,444 patients had at least one pancreatic cyst (Table 1). The prevalence of pancreatic cysts increased with age (r = 0.96) (Fig 1). The prevalence of pancreatic cysts was 4.8% (in three of 63 patients; 95% CI: 3.7%, 5.9%) before age 30 years, which was significantly lower than the prevalence after age 30 years (in 280 of 1,381 patients) (P = .001, Fisher exact test). The numbers of male (20.4%; 95% CI: 19.2%, 21.6%) and female (18.8%; 95% CI: 17.7%, 19.9%) patients with pancreatic cysts (Table 2) were not statistically different (2 = 0.6; P > .05). The distributions of cysts in the head, body, and tail of the pancreas were 32.1% (95% CI: 30.2%, 34.0%), 45.7% (95% CI: 43.5%, 47.9%), and 22.2% (95% CI: 20.7%, 23.7), respectively (P < .01) (Table 3).
|
|
|
|
2 = 43.0, P < .001). The prevalences of both types of cysts increased with age (Table 1). The number of male patients with simple cysts (100 [14.2%] of 705; 95% CI: 13.3%, 15.1%) was not significantly different from the number of female patients with simple cysts (84 [11.4%] of 739; 95% CI: 10.5%, 12.1%) (2 = 2.6; P > .05). The numbers of male (73 [10.4%] of 705; 95% CI: 9.7%, 11.1%) and female (70 of [9.5%] 739; 95% CI: 8.9%, 10.1%) patients with nonsimple cysts were almost the same.
|
|
70 years; Fisher exact test, P < .001).
|
|
|
|
Nineteen of the 283 patients with pancreatic cysts had pancreatic tumors that were confirmed at pathologic analysis. Two of these 19 patients had islet cell tumor, one had microcystic adenoma, and 16 (5.7%; 95% CI: 5.1%, 6.3%) (age range, 37–84 years; average age, 65.1 years) had malignant pancreatic neoplasms. The 16 patients with pancreatic malignancies included 12 (4.2%; 95% CI: 3.7%, 4.7%) patients with ductal cell adenocarcinoma in the head of the pancreas and postobstructive pancreatitis (Table 7), two with intraductal papillary mucinous carcinomas in the head, one with malignant islet cell neuroendocrine carcinoma, and one with malignant spindle cell neoplasm in the body.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
We classified pancreatic cysts as simple or nonsimple types according to their signal intensities on single-shot fast SE MR images. Our category of simple cysts included true cysts, but it may have also included some pseudocysts and some cystic tumors. We were not able to verify that all lesions that appeared to be simple cysts were truly simple. Nonsimple cysts are mainly pseudocysts and cystic neoplasms (9–13).
In this study, 283 (19.6%; 95% CI: 18.8%, 20.4%) of the 1,444 patients had at least one pancreatic cyst. The prevalence of pancreatic cysts increased with age; these findings were similar to those in a study of 300 consecutive autopsies, in which prevalences of 8.6%, 18.9%, 30.4%, and 35.3% were observed in the age groups of 69 years or younger, 70–79 years, 80–89 years, and 90 years or older, respectively (3). These prevalences are slightly lower than those in the same age groups in our study (Table 1). One reason for this could be that only small pancreatic cysts (diameter < 20 mm) were included in the other study. We also speculate that our study population included more patients with pancreatic disease who were referred for this reason, although we are not able to verify this.
The prevalence of pancreatic cysts determined with MR imaging in our study was much higher than that determined with US. In the report by Ikeda et al (1), cystic lesions were found in 271 (0.21%) patients at screening US in 130,951 subjects. Most lesions detected at US were larger than 10 mm in diameter (1), partly because the retroperitoneal location of the pancreas often makes imaging difficult; some cystic lesions may be obscured by gas or mistaken as a solid mass (14).
We observed no sex-related predominance in the prevalence of pancreatic cysts (P > .05), unlike Ikeda et al (1), who found pancreatic cysts more frequently in female patients (P < .01). We are unable to explain these differences; note, however, that Ikeda et al used screening US, and, thus, subcentimeter cysts represented a smaller proportion of the cysts detected in their series than they did in our study. We also cannot determine whether their screening population differed from our population of patients who were referred for abdominal MR imaging. Mucinous cystic neoplasms are far more common in female subjects than they are in male subjects (male-to-female ratio, 1:9) (10), and serous cystic tumors occur more than twice as often in female subjects (11). However, intraductal papillary mucinous tumors have been reported to be more common in male subjects (12). For pseudocysts, an almost equal sex distribution is observed when these cysts are secondary to non–alcohol-related pancreatitis, whereas there is a strong male predominance with alcohol-related pancreatitis (13).
Our results with MR imaging are more similar to those obtained at pathologic analysis than are the results with US. In the report by Kimura et al (2), 177 of 186 small cysts found at 300 autopsies were examined histologically. The largest cyst in their series was 12 mm in diameter. Of 177 small cysts, 47.5% had a normal epithelium, 32.8% were papillary hyperplasias without atypia, 16.4% were atypical hyperplasias, and 3.4% were carcinomas in situ. These authors suggested that small cystic lesions that exhibit either atypical hyperplasia or carcinoma in situ begin to infiltrate when they are 4–5 mm in diameter and then develop into the typical ductal cell carcinoma.
Epithelial proliferative lesions, such as papillary or atypical hyperplasia of the pancreatic duct, have been considered to have a role in the pathogenesis of ductal cell carcinoma of the pancreas (15). Carcinoma in situ may precede the development of invasive carcinoma (16,17). Kozuka et al (18) suggested that these lesions change from nonpapillary hyperplasia to papillary and atypical hyperplasia and then to carcinoma and that this process may reflect the pathogenesis of ductal cell carcinoma of the pancreas. In our study, small pancreatic cysts were seen incidentally in 16.1% (±0.7) of patients; 60.1% (±6.3) of the patients with small cysts were older than 60 years (Table 6). Although we could not determine the histologic features of these small cysts, according to the report of Kimura et al (2), almost half of them may be epithelial proliferative lesions.
Proliferative cysts can be related to cystic tumors in addition to pancreatic ductal cell cancer (15,16,18,19). Cystic tumors represent 5%–15% of pancreatic cysts (10). The most common cystic neoplasms of the pancreas include serous cystic neoplasms (38.0%), mucinous cystic neoplasms (9.7%), intraductal papillary mucinous tumors (30.6%), and solid pseudopapillary tumors (11.9%) (20). Serous cystic tumors are clearly benign in the majority of cases and do not communicate with the pancreatic ductal system (11). Mucinous cystic neoplasms are commonly malignant, but patients with these neoplasms have a better prognosis after resection than do those with ductal cell adenocarcinoma (10,21,22). Intraductal papillary mucinous tumor is a relatively new category of pancreatic tumors that was created in the middle 1990s (12). Intraductal papillary mucinous tumors are composed of an intraductal proliferation of mucin-producing cells that often form papillary projections and frequently are associated with copious intraluminal mucin secretion that leads to cystic dilatation of the ductal system. Histologically, the proliferation may show a spectrum of cytoarchitectural atypia ranging from none (ie, adenoma) to borderline and to marked (carcinoma in situ), and in some cases (in approximately 35% of patients), the proliferation may be associated with invasive carcinoma (12). We speculate that some of the small pancreatic cysts in our series may have been premalignant or early malignant lesions.
In the United States, the incidence of pancreatic cancer is 11.7–13.7 per 100,000 male patients and 7.2–11.9 per 100,000 female patients, and about 1% of the entire population will eventually develop pancreatic cancer (23). In our study, 12 (4.2%; 95% CI: 3.7%, 4.7%) of 283 patients with pancreatic cysts had pancreatic ductal cell adenocarcinoma. The much higher prevalence of pancreatic cancer in the patients with pancreatic cysts in our series, as compared with the overall incidence of pancreatic cancer, indicates that pancreatic cysts may be related to pancreatic cancer.
Congenital pancreatic cyst is an intrapancreatic cystic lesion that does not communicate with the ductal system and is lined by a single layer of flat epithelium (13). It may manifest as a single cyst or as multiple cysts involving one portion of the pancreas or the entire pancreas (24). Data in previously published literature indicate that congenital pancreatic cysts account for less than 1% of all pancreatic cysts and usually occur in children younger than 2 years (25). To our knowledge, only 10 cases of solitary true cysts of the pancreas in adults have been reported in the English-language literature, and most of them were larger than 2 cm in diameter (26). However, the report of Kimura et al (2) suggests that almost half of small cysts are congenital epithelial cysts. Our series included a larger proportion of subcentimeter cysts than that noted in prior imaging series because of the high sensitivity of single-shot fast SE MR imaging for the detection of small cysts. We therefore believe that it is possible that many of the tiny cysts detected in our series were congenital epithelial cysts, as described by Kimura et al (2); however, we are not able to confirm this.
Pseudocysts, which are reported to be the most common type of pancreatic cystic lesions (8,27,28), develop subsequently to an episode of severe acute pancreatitis. About 10% of patients with acute pancreatitis have clinically relevant pseudocysts, and more than half of the patients with alcohol-related chronic pancreatitis develop pseudocysts (13). In our study, 75 (26.5%; 95% CI: 24.6%, 29.2%) of the patients with pancreatic cysts, including eight patients with extrapancreatic pseudocysts, had a history of pancreatitis. We believe that most of the pancreatic cysts in these 75 patients were pseudocysts.
One hundred forty-seven (51.9%; 95% CI: 49.0%, 54.8%) of the 283 patients with pancreatic cysts had at least one renal cyst. These results may be related to the large number (62.5%) of patients with renal cysts identified with single-shot fast SE MR imaging rather than to an association between these two entities (6).
ADPKD is a common hereditary disorder characterized by the progressive formation and enlargement of renal cysts and by the involvement of other organs, such as the liver, spleen, and pancreas. Pancreatic cysts were reported in 5.7% of 173 patients with ADPKD identified at US and in about 9.0% of patients with ADPKD who were older than 30 years (29). In our study, three patients (1.1% ± 0.1) with pancreatic cysts had ADPKD: Two patients each had two cysts, and one patient had more than 10 cysts.
Other systemic diseases that commonly manifest with multiple pancreatic cysts are von Hippel–Lindau disease (30) and cystic fibrosis (31). However, among the nine patients with more than 10 cysts in our study, none had a history of either of these diseases, although five had pancreatitis. We therefore believe that pancreatitis was the primary cause of multiple pancreatic cysts in our series.
We did not obtain independent confirmation that all of the simple cysts detected on heavily T2-weighted single-shot fast SE images were truly simple. Higher spatial resolution images might have depicted septa or wall irregularities that would have led us to categorize these as nonsimple cysts. However, our findings included the sizes and prevalences of apparent simple cysts and their associations with age, sex, and presence of associated disease.
In summary, the prevalence of pancreatic cysts detected with single-shot fast SE MR imaging, especially those with a diameter of less than 10 mm, was similar to that of pancreatic cysts detected at autopsy (2) and higher than that of pancreatic cysts detected at transabdominal US. The prevalence of these cysts was especially high in patients with pancreatitis.
|
FOOTNOTES |
|---|
Author contributions: Guarantors of integrity of entire study, X.M.Z., D.G.M.; study concepts and design, X.M.Z., D.G.M.; literature research, X.M.Z., D.G.M.; clinical and experimental studies, X.M.Z., M.D., G.A.H.; data acquisition, X.M.Z., M.D., G.A.H.; data analysis/interpretation, X.M.Z., D.G.M., L.P.; statistical analysis, X.M.Z., L.P.; manuscript preparation, X.M.Z., D.G.M., G.A.H.; manuscript definition of intellectual content, X.M.Z., D.G.M.; manuscript editing and revision/review, all authors; manuscript final version approval, X.M.Z., D.G.M.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  T. A. Laffan, K. M. Horton, A. P. Klein, B. Berlanstein, S. S. Siegelman, S. Kawamoto, P. T. Johnson, E. K. Fishman, and R. H. Hruban Prevalence of Unsuspected Pancreatic Cysts on MDCT Am. J. Roentgenol., September 1, 2008; 191(3): 802 - 807. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. D'Onofrio, A. J. Megibow, N. Faccioli, R. Malago, P. Capelli, M. Falconi, and R. P. Mucelli Comparison of Contrast-Enhanced Sonography and MRI in Displaying Anatomic Features of Cystic Pancreatic Masses Am. J. Roentgenol., December 1, 2007; 189(6): 1435 - 1442. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. S. Katz, D. M. Friedel, D. Kho, N. Georgiou, and J. J. Hines Relative Accuracy of CT and MRI for Characterization of Cystic Pancreatic Masses Am. J. Roentgenol., September 1, 2007; 189(3): 657 - 661. [Full Text] [PDF]
|
|
|
|
|
|
S. J. Handrich, D. M. Hough, J. G. Fletcher, and M. G. Sarr The Natural History of the Incidentally Discovered Small Simple Pancreatic Cyst: Long-Term Follow-Up and Clinical Implications Am. J. Roentgenol., January 1, 2005; 184(1): 20 - 23. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
