- Email: [email protected]
Macrocystic pancreatic lesions: Differentiation of benign from premalignant and malignant cysts by CT
European Journal of Radiology 71 (2009) 122–128
Macrocystic pancreatic lesions: Differentiation of benign from premalignant and malignant cysts by CT Seong Hyun Kim ∗ , Jae Hoon Lim, Won Jae Lee, Hyo K. Lim Department of Radiology and Center for Imaging Science and Samsung Medical Center, Sungkyunkwan University School of Medicine, 50 Ilwon-dong, Gangnam-gu, Seoul 135-710, South Korea Received 3 January 2008; received in revised form 9 March 2008; accepted 12 March 2008
Abstract Objective: To assess useful CT features for differentiating benign from premalignant and malignant macrocystic pancreatic lesions. Methods: Seventy-four patients with pathologically proven macrocystic pancreatic lesions were enrolled: 17 benign cysts (macrocystic serous cystadenoma, n = 12; congenital cyst; n = 5) and 57 premalignant and malignant cysts (mucinous cystic neoplasm, n = 28; intraductal papillary mucinous neoplasm of branch duct type, n = 20; tumor with cystic change, n = 9). Size, location, shape (lobulated, round or oval, or complex cystic with tubular cyst), wall thickness (thin, ≤1 mm; thick, >1 mm), internal surface (smooth or irregular), and other findings were analyzed with multiphasic CT with thin-section (2.5–3 mm) images. CT features between two groups were compared using univariate and multivariate stepwise logistic regression analyses. Results: On univariate analysis, the differences for the shape (p = 0.007), wall thickness (p = 0.011), and internal surface (p = 0.012) between benign and premalignant and malignant cysts were significant. A lobulated shape, a thin wall and a smooth internal surface were more frequent in benign cysts, whereas a round or oval shape or a complex cystic shape with tubular cyst, a thick wall and an irregular internal surface were more frequent in premalignant and malignant cysts. On multivariate analysis, the shape (p = 0.002) and wall thickness (p = 0.025) were significant CT features for differentiating benign from premalignant and malignant cysts. Conclusion: Shape and wall thickness are the main CT features for differentiating benign from premalignant and malignant macrocystic pancreatic lesions. © 2008 Elsevier Ireland Ltd. All rights reserved. Keywords: Pancreas; Pancreas neoplasms; Pancreas cysts; Pancreaticobiliary imaging; CT
1. Introduction Cystic lesions of the pancreas are being increasingly recognized because of frequent use of advanced imaging techniques. Of these, congenital cyst, serous cystadenoma (SCA) with macrocystic variant, mucinous cystic neoplasm (MCN), intraductal papillary mucinous neoplasm (IPMN) of the branch duct type and tumor with cystic change which includes solid pseudopapillary tumor, neuroendocrine tumor and pancreatic ductal adenocarcinoma all may have macrocystic morphologic features that overlap on imaging despite of different malignant potential [1–15]. Benign cysts such as congenital cyst and macrocystic SCA can clinically be managed in a non-surgical manner unless the lesions are symptomatic. In contrast, MCN, IPMN of the
∗
Corresponding author. Tel.: +82 2 3410 2548; fax: +82 2 3410 2559. E-mail address: [email protected] (S.H. Kim).
0720-048X/$ – see front matter © 2008 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.ejrad.2008.03.012
branch duct type and tumor with cystic change are premalignant or malignant, and so they warrant surgical resection [1–11]. CT is still the preferred imaging modality for both initial detection and characterization of these macrocystic pancreatic lesions. Thus, it is crucial to identify the useful CT features for making the differentiation of benign cysts from premalignant and malignant cysts so as to preclude unnecessary surgery because the morbidity after surgical management remains high and the outcomes of loosing pancreatic tissue may be serious. Although a few investigators [5,6] have reported some useful CT features for differentiating macrocystic SCA from mucinous cystadenoma and IPMN of the branch duct type, to the best of our knowledge, no reports are available concerning the usefulness of CT with thin-section images for differentiating benign from premalignant and malignant macrocystic lesions of the pancreas in a large population, along with the pathological proof. The purpose of this study was to assess the useful CT features for differentiating benign from premalignant and malignant macro-
S.H. Kim et al. / European Journal of Radiology 71 (2009) 122–128
cystic lesions of the pancreas by using multiphasic helical CT with thin-section images. 2. Materials and methods 2.1. Patients Institutional review board approval was obtained for this retrospective study and informed consent was waived. Between January 1999 and March 2007, we searched for all the pathologically proven macrocystic pancreatic lesions including pancreatic tumor with cystic or necrotic change more than 90%, which were equal to or larger than 2 cm in diameter, via a review of the archives of our departments of surgery, pathology, and radiology. We found a total of 253 consecutive patients who underwent pancreatic resection for pancreatic cystic lesions during this period. Of these, we excluded the patients (n = 13) with pseudocyst, which had clinical and laboratory evidence of pancreatitis and/or imaging evidence of pancreatic and peripancreatic inflammation, pancreatic ductal calcification or atrophy. The patients with typical microcystic SCA with multiple cysts smaller than 2 cm in diameter and IPMNs with the main duct type or the combined type were also excluded (n = 74). Of the remaining 166 patients, 92 patients were excluded because either CT was not performed or thick-section (thicker than 5 mm) CT images were performed. The remaining 74 patients were finally enrolled. The final diagnosis were macrocystic SCA in 12, simple cyst in two, lymphoepithelial cyst in three, MCN proven as adenoma in 22, borderline in two, and carcinoma in four, IPMN of the branch duct type proven as adenoma in 12, borderline in five, and carcinoma in three, and tumor with cystic change in nine, including solid pseudopapillary tumor in six, well differentiated neuroendocrine tumor in two, and ductal adenocarcinoma in one. We divided the patients into those with benign cysts including congenital cyst and macrocystic SCA (age range, 34–79 years; mean age, 52 years; 5 men, 12 women) and those with premalignant or malignant cysts (n = 57) including MCN, IPMN of the branch duct type and tumor with cystic change (age range, 19–73 years; mean age, 51 years; 19 men, 38 women). 2.2. CT examination All patients underwent contrast-enhanced CT of the upper abdomen. Different CT machines were used during this period; a single-detector helical CT scanner (HiSpeed Advantage; GE healthcare) (n = 15) and MDCT scanners with four (LightSpeed QX/I; GE healthcare) (n = 16), eight (LightSpeed Ultra 8; GE Healthcare) (n = 11), 16 (LightSpeed 16; GE Healthcare) (n = 16), and 40 (Brilliance 40; Philips Medical Systems, Eindhoven, Netherlands) (n = 16) detectors, respectively. The scanning parameters were 120 kVp and 175–189 mAs. After unenhanced CT scanning, a total of 120 mL of nonionic iodinated contrast material (Ultravist 300 [300 mg I/mL iopromide], Schering AG, Berlin, Germany) was administered intravenously with an automatic injector at a rate of 3–4 mL/s. The images were obtained at 35 s, 55 s and 70 s after the initiation of the contrast material injection, which represent the arterial, pancreatic and
123
portal venous phases, respectively. With the single-detector helical CT scanner, the arterial and pancreatic phase images were obtained in a craniocaudal direction with 3 mm slice thickness and a 3 mm reconstruction interval. With the MDCT scanner, the arterial and pancreatic phases images were obtained in a craniocaudal direction with 2.5 mm slice thickness and a 2.5 mm reconstruction interval. The unenhanced and portal venous phase images were obtained with 5 mm slice thickness and a 5 mm reconstruction interval. 2.3. Image analysis All CT images were interpreted on 1536 × 2048 (3 M pixel) monitor (CCL350i, TOTOKU, Japan) of a picture archiving and communication system (PACS) (GE Medical Systems Integrated Imaging Solutions, Mt Prospect, IL, USA) with stack-view capability. Unenhanced and contrast-enhanced images were evaluated together. Subjective visual criteria were used rather than density measurements. Two abdominal radiologists (7- and 27-year experience, respectively), who were blinded to the specific diagnosis and clinical information, reviewed the CT images by consensus to define the following morphological features of the lesions; maximum dimension (measured in centimeters), location (head, neck, body or tail), shape, wall thickness (measured in millimeters), content within the lesion (homogeneous or inhomogeneous), internal surface (smooth or irregular), presence or absence of calcification in the wall and/or septum, septa, mural nodule, dilatation of the main pancreatic duct, peripancreatic fat infiltration, peripancreatic enlarged lymph node or vascular involvement as based on the reports [3,5–23]. The presence of calcification was determined on the unenhanced images. For the size of the cysts, we used 3 cm in diameter as a cut-off value for being relatively small or large [16]. The shape of the lesion was categorized as lobulated, round or oval, and complex cystic with tubular cyst based on the reports [5,6]. The lobulated shape was defined as the presence of rounded contours that could not be described as a well-defined border of the same circle. The round or oval shape was defined as a round or oval configuration with a broad well-defined border. The complex cystic shape with tubular cyst was defined as any shaped cystic lesions with at least one tubular cyst. A tubular cyst was defined as one having a conic section, in which the plane of the tubular cyst was not parallel to the axis, base, or generatrix of the intersection cone as based on the report [6]. The wall of the cyst was measured with using exclusively thin-section (2.5–3 mm) images obtained during the arterial or pancreatic phase by one of the reviewers, with using both a compass and a digital caliper on the PACS. The wall was considered thin if the wall thickness was 1 mm or less, and it was considered thick if it was thicker than 1 mm in diameter for at least a quarter of the lesion circumference as based on the reports [22,23]. If the wall of the cyst was not discriminated, the wall was considered to be imperceptible and included in the group of thin wall [5]. Any irregularity of the internal surface was classified as irregular. Peripancreatic lymphadenopathy was diagnosed when one or more peripancreatic lymph nodes with a short-axis diameter of 10 mm or more were identified. The peripancreatic fat
124
S.H. Kim et al. / European Journal of Radiology 71 (2009) 122–128
infiltration was defined if an increased attenuation relative to that of subcutaneous fat was identified. The vascular involvement was defined when the occlusion of any nearby veins with engorged collaterals or irregular narrowing or apparent encasement of any nearby arteries was seen. 2.4. Statistical analysis The mean maximum dimensions of two groups were compared using Mann–Whitney U test. For univariate analysis, statistical differences of the CT features of two groups were analyzed with the Chi-square test and Fisher’s exact test. A multivariate stepwise logistic regression model was used to determine the most significant CT features differentiating two groups. Significant differences were defined as p values less than 0.05. All statistical analyses were performed with the SPSS software package (version 11.5; SPSS, Chicago III). 3. Results 3.1. Univariate analysis Table 1 summarizes the different CT features observed in two groups. The benign cysts ranged in maximum dimension from 2.0 to 9.5 cm (mean ± standard deviation, 4.3 ± 2.2 cm)
Fig. 1. A 57-year-old woman with a macrocystic serous cystadenoma in the head of the pancreas. CT image at pancreatic phase shows a lobulated cystic lesion (arrow) with the imperceptible wall, homogeneous content, and smooth internal surface in the head of the pancreas. Note a thin septum (arrowhead).
and this was similar to premalignant and malignant cysts (range, 2.0–12 cm; mean ± standard deviation, 4.7 ± 2.9 cm) (p = 0.969). The lobulated shape was more frequently seen in benign cysts (Figs. 1 and 2) than in premalignant and malignant cysts (Figs. 3 and 4). In contrast, the round or
Table 1 CT features of benign and premalignant and malignant cysts of the pancreas CT feature
Benign cyst (n = 17)
Premalignant and malignant cyst (n = 57)
Size ≤3 cm >3 cm
7 (41) 10 (59)
25 (44) 32 (56)
Location Head Body or tail
9 (53) 8 (47)
17 (30) 40 (70)
Shape* Lobulated Round or oval Complex cystic with tubular cyst
9 (52.9) 8 (47.1) 0
12 (21.1) 30 (52.6) 15 (26.3)
p 0.845
0.080
0.007
Wall thickness* ≤1 mm (thin) >1 mm (thick)
14 (82) 3 (18)
27 (47) 30 (53)
Content Homogeneous Inhomogeneous
15 (88) 2 (12)
39 (68) 18 (32)
Internal surface Smooth Irregular
15 (88) 2 (12)
31 (54) 26 (46)
Calcification Septa Mural nodule Dilatation of main pancreatic duct Peripancreatic fat infiltration Peripancreatic lymphadenopathy Vascular involvement
3 (18) 11 (65) 1 (6) 2 (12) 0 0 0
16 (28) 48 (84) 11 (19) 13 (23) 6 (11) 4 (7) 7 (12)
0.011
0.131
0.012
0.533 0.095 0.274 0.496 0.326 0.568 0.192
Note: Numbers in parentheses are percentages. * On multivariate stepwise logistic regression analysis, shape (p = 0.002; 95% confidence interval [1.460, 5.805]) and wall thickness (p = 0.025; 95% confidence interval [1.199, 15.826]) were statistically significant.
S.H. Kim et al. / European Journal of Radiology 71 (2009) 122–128
Fig. 2. A 53-year-old woman with a macrocystic serous cystadenoma in the tail of the pancreas. CT image at pancreatic phase shows a lobulated cystic lesion (long arrow) with the thin wall (short arrow) in the tail of the pancreas. Note septa and homogeneous content, smooth internal surface.
oval shape (Figs. 5–7) was more frequently seen and the complex cystic shape with tubular cyst (Fig. 8) was exclusively seen in premalignant and malignant cysts as compared with benign cysts (Figs. 9–11). The differences of the shape between two groups were significant (p = 0.007). The wall of the cysts were considered thick (>1 mm) in three benign cysts (Figs. 9 and 11) and in 30 premalignant and malignant cysts (Figs. 5–7), and this difference was significant (p = 0.011). The smooth internal surface was more frequently seen in benign cysts (Figs. 1, 2, 9–11) than in premalignant and malignant cysts (Figs. 3 and 4) (p = 0.012). No significant differences were found between two groups with respect to a size difference with 3 cm being used as cut-off value (p = 0.845), location (p = 0.080), content (p = 0.131), and the presence of calcification (p = 0.533), septa (p = 0.095), mural nodule (p = 0.274), and dilatation of the main pancreatic duct (p = 0.496). Peripancre-
Fig. 3. A 73-year-old woman with a mucinous cystadenoma of the pancreas. CT image at arterial phase shows a lobulated cystic lesion (arrow) with the imperceptible wall and smooth internal surface, which especially mimics such a benign cyst as a macrocystic serous cystadenoma, in the body of the pancreas.
125
Fig. 4. A 69-year-old woman with an intraductal papillary mucinous adenoma of the branch duct type in the pancreas. CT image at pancreatic phase shows a lobulated cystic lesion (arrow) with the imperceptible wall and smooth internal surface, which especially mimics such a benign cyst as a macrocystic serous cystadenoma, in the head of the pancreas.
atic abnormalities such as fat infiltration, lymphadenopathy and vascular involvement (Figs. 6 and 7) were only seen in six, four, and seven premalignant and malignant cysts, respectively, but no statistical significance was found compared with benign cysts. 3.2. Multivariate analysis Of three significant CT features on univariate analysis, multivariate stepwise logistic regression analysis showed that the shape (p = 0.002; 95% confidence interval [1.460, 5.805]) and wall thickness (p = 0.025; 95% confidence interval [1.199, 15.826]) were statistically significant CT features for differentiating benign from premalignant and malignant cysts.
Fig. 5. A 26-year-old woman with a mucinous cystadenoma of the pancreas. CT image at pancreatic phase shows a septated cystic lesion with round shape (arrow), surrounded by the thick wall in the head of the pancreas. The internal surface appears irregular (arrowheads).
126
S.H. Kim et al. / European Journal of Radiology 71 (2009) 122–128
Fig. 6. A 28-year-old woman with a mucinous cystadenocarcinoma of the pancreas. CT image at pancreatic phase shows a septated cystic lesion of the oval shape (long black arrow) with the thick wall and irregular internal surface (long white arrows) in the tail of the pancreas. The inhomogeneous content and calcifications (short arrows) in the septum and the wall are shown. Note perigastric engorged veins (arrowheads) due to the occlusion of the splenic vein (not shown).
4. Discussion The main clinical problem for the macrocystic pancreatic lesions is not detecting them, but rather, it is characterizing them, followed by deciding on the most appropriate therapeutic plan. However, neither imaging studies nor even cyst aspiration gives the physician the precise information to differentiate each benign, premalignant and malignant cystic lesion of the pancreas [4]. Therefore, it is clinically valuable to identify the useful CT features for making the differentiation of benign macrocystic lesions that can be closely followed unless the lesions are symptomatic, from premalignant and malignant macrocystic lesions that warrant surgical resection. Pseudocysts are the most common pancreatic cystic lesions [8]. Although pseudocysts and cystic neoplasms of the pancreas
Fig. 7. A 19-year-old woman with a solid pseudopapillary tumor of the pancreas. CT image at pancreatic phase shows a round cystic lesion (long arrow) with the thick wall and irregular internal surface (short arrows) in the tail of the pancreas. Note a solid component (asterisk), calcifications (black arrowheads) in septa, and perigastric engorged veins (white arrowheads).
Fig. 8. A 72-year-old man with an intraductal papillary mucinous adenoma of the branch duct type of the pancreas. (a) CT image at pancreatic phase shows a cystic lesion (arrow) with septa (arrowhead) and imperceptible wall, which is formed by cystically dilated branch duct in the head of the pancreas. Note mild dilatation of the main pancreatic duct. (b) The Coronal reformatted CT image shows a complex cystic shape with tubular cyst (arrowheads) with the irregular internal surface (short arrow) and mural nodule (long arrow).
Fig. 9. A 50-year-old woman with a macrocystic serous cystadenoma in the pancreas. CT image at pancreatic phase shows a round cystic lesion (arrow) with the thick wall, mimicking premalignant or malignant cyst, in the tail of the pancreas.
S.H. Kim et al. / European Journal of Radiology 71 (2009) 122–128
Fig. 10. A 36-year-old woman with a simple cyst of the pancreas. CT image at pancreatic phase shows a round cystic lesion (arrow) with the imperceptible wall and smooth internal surface in the body of the pancreas.
can mimic each other on imaging [24], a careful clinical history of pancreatitis with laboratory clues (amylase or lipase), and/or imaging evidence of pancreatic and peripancreatic inflammation, pancreatic ductal calcification or atrophy may allow diagnosis of a cystic lesion as a pseudocyst [3,8,9,12,18]. In our study, we excluded 13 patients with pseudocysts because the patients had clinical and laboratory evidence of pancreatitis and/or the imaging evidence of pancreatic and peripancreatic inflammation, pancreatic ductal calcification or atrophy. Instead, the patients with benign congenital cysts such as simple cyst and lymphoepithelial cyst, and premalignant or malignant IPMNs of the branch duct type were included in the study, and all of these may mimic other macrocystic lesions of the pancreas [1–3,6,9,10,13,19–21]. The present study showed that the shape and wall thickness were the main CT features for differentiating benign from premalignant and malignant macrocystic lesions of the pancreas.
127
Some investigators reported that the shape of macrocystic pancreatic neoplasms was useful for differentiating macrocystic SCAs, MCNs and IPMNs of the branch duct type from each other [5,6]. In our study, the lobulated shape was a common feature in benign cysts and the complex cystic shape with tubular cyst was only seen in premalignant and malignant cysts. We believe that these results are due to a large proportion of macrocystic SCA within benign cysts and due to premalignant or malignant IPMNs of the branch duct type which appear as multiple separate cysts with each cystic locule representing a dilated, tortuous pancreatic duct resulting in a tubular cyst [2,3,5,6,10–12]. In our study, a round or oval shape was slightly more frequent in premalignant and malignant cysts than in benign cysts. We think that this result may be influenced by MCNs and tumors with cystic change [5,6,13,19–21]. Some investigators have reported that benign cysts such as macrocystic SCAs had a thin wall that measured 0.1 cm in thickness [22,23] and no predominant enhancement of the wall of these lesions was seen [5,7]. In our study, we often found the wall of benign cysts appeared to be imperceptible, equal to or less than 1 mm in diameter, as compared with premalignant and malignant cysts being thick-walled on thin-section CT. MCNs, IPMNs of the branch duct type and tumors with cystic change may more frequently have an irregular internal surface compared to macrocystic SCAs and congenital cysts because MCNs and IPMNs of the branch duct type may have papillary excrescences into the cystic lumen, and tumor with cystic change may frequently be caused by necrosis or hemorrhage; moreover, all these lesions may have mural nodule or a solid component [3,6,9,10,12,13,15]. Our results showed that the presence or absence of irregular internal surface was a helpful CT feature for differentiating benign from premalignant and malignant cysts. CT features such as size, location, calcification, septa and mural nodule (or a solid component) may be helpful in differentiating benign and malignant cystic neoplasms [3,9,10,14,16,17]. However, in our study, these CT features were not statistically different between benign and premalignant and malignant cysts. The imaging overlap between cystic lesions or less frequent occurrence of these findings may be possible reasons on these results. In our study, peripancreatic abnormalities such as fat infiltration, lymphadenopathy and vascular involvement were only seen in premalignant and malignant cysts. We think that these CT features may be of additional value for differentiating benign from premalignant and malignant cysts. Our study has some limitations. First, we retrospectively investigated CT features that could facilitate differentiation between benign and premalignant and malignant macrocystic pancreatic lesions that were pathologically proven, which might be a bias in case selection. Second, the enrolled patients were collected over 8 years. Although only the CT images of optimal quality were included, the imaging parameters did vary. 5. Conclusion
Fig. 11. A 50-year-old man with a lymphoepithelial cyst of the pancreas. CT image at pancreatic phase shows an oval cystic lesion (arrow) with the thick wall and septum (arrowhead), mimicking a premalignant or malignant cyst, in the body of the pancreas.
Our results show that the shape and wall thickness are significant independent CT features for differentiating benign from premalignant and malignant macrocystic pancreatic lesions, in
128
S.H. Kim et al. / European Journal of Radiology 71 (2009) 122–128
which a lobulated shape and a thin wall (≤1 mm) are more frequently seen in benign macrocystic pancreatic lesions and a round or oval shape or a complex cystic shape with tubular cyst and a thick wall (>1 mm) are more frequently seen in premalignant and malignant macrocystic pancreatic lesions. References [1] Sakorafas GH, Sarr MG. Cystic neoplasms of the pancreas; what a clinician should know. Cancer Treat Rev 2005;31(7):507–35. [2] Klimstra DS. Cystic, mucin-producing neoplasms of the pancreas: the distinguishing features of mucinous cystic neoplasms and intraductal papillary mucinous neoplasms. Semin Diagn Pathol 2005;22(4):318–29. [3] Demos TC, Posniak HV, Harmath C, Olson MC, Aranha G. Cystic lesions of the pancreas. AJR 2002;179(6):1375–88. [4] Spinelli KS, Fromwiller TE, Daniel RA, et al. Cystic pancreatic neoplasms: observe or operate. Ann Surg 2004;239(5):651–9. [5] Cohen-Scali F, Vilgrain V, Brancatelli G, et al. Discrimination of unilocular macrocystic serous cystadenoma from pancreatic pseudocyst and mucinous cystadenoma with CT: initial observations. Radiology 2003;228(3):727–33. [6] Kim SY, Lee JM, Kim SH, et al. Macrocystic neoplasms of the pancreas: CT differentiation of serous oligocystic adenoma from mucinous cystadenoma and intraductal papillary mucinous tumor. AJR 2006;187(5):1192–8. [7] Khurana B, Mortele KJ, Glickman J, Silverman SG, Ros PR. Macrocystic serous adenoma of the pancreas: radiologic–pathologic correlation. AJR 2003;181(1):119–23. [8] Curry CA, Eng J, Horton KM, et al. CT of primary cystic pancreatic neoplasms: can CT be used for patients triage and treatment? AJR 2000;175(1):99–103. [9] Sahani DV, Kadavigere R, Saokar A, Fernandez-del Castillo C, Brugge WR, Hahn PF. Cystic pancreatic lesions: a simple imaging-based classification system for guiding management. Radiographics 2005;25(6):1471– 84. [10] Kawamoto S, Horton KM, Lawler LP, Hruban RH, Fishman EK. Intraductal papillary mucinous neoplasm of the pancreas: can benign lesions be differentiated from malignant lesions with mulidector CT? Radiographics 2005;25(6):1451–70.
[11] Chiu SS, Lim JH, Lee WJ, et al. Intraductal papillary mucinous tumour of the pancreas: differentiation of malignancy and benignancy by CT. Clin Radiol 2006;61(9):776–83. [12] Kim YH, Saini S, Sahani D, Hahn PF, Mueller PR, Auh YH. Imaging diagnosis of cystic pancreatic lesions: pseudocyst versus nonpseudocyst. Radiographics 2005;25(3):671–85. [13] Park M, Kim KW, Lim JS, et al. Unusual cystic neoplasms in the pancreas: radiologic–pathologic correlation. J Comput Assist Tomogr 2005;29(5):610–6. [14] Procacci C, Carbognin G, Accordini S, et al. CT features of malignant mucinous cystic tumors of the pancreas. Eur Radiol 2001;11(9):1626–30. [15] Buetow PC, Rao P, Thompson LD. Mucinous cystic neoplasms of the pancreas: radiologic–pathologic correlation. Radiographics 1998;18(2):433–49. [16] Sahani DV, Saokar A, Hahn PF, Brugge WR, Fernandez-del Castillo C. Pancreatic cysts 3 cm or smaller: how aggressive should treatment be? Radiology 2006;238(3):912–9. [17] Procacci C, Graziani R, Bicego E, et al. Serous cystadenoma of the pancreas: report of 30 cases with emphasis on the imaging findings. J Comput Assist Tomogr 1997;21(3):373–82. [18] Procacci C, Biasiutti C, Carbognin G, et al. Characterization of cystic tumors of the pancreas: CT accuracy. J Comput Assist Tomogr 1999;23(6):906–12. [19] Fiamingo P, Veroux M, Gringeri E, et al. True solitary pancreatic cyst in an adult: report of a case. Surg Today 2005;35(11):979–83. [20] Kim YH, Auh YH, Kim KW, Lee MG, Kim KS, Park SY. Lymphoepithelial cysts of the pancreas: CT and sonographic findings. Abdom Imaging 1998;23(2):185–7. [21] Shinmura R, Gabata T, Matsui O. Lymphoepithelial cyst of the pancreas: case report with special reference to imaging–pathologic correlation. Abdom Imaging 2006;31(1):106–9. [22] Huh JR, Chi JG, Jung KC, Choe KJ, Yoon YB. Macrocystic serous cystadenoma of pancreas: a case report. J Korean Med Sci 1994;9(1):78–85. [23] Fujiwara H, Ajiki T, Fukuoka K, Mitsutsuji M, Yamamoto M, Kuroda Y. Macrocystic serous cystadenoma of the pancreas. J Hepatobiliary Pancreat Surg 2000;7(1):92–6. [24] Scott J, Martin I, Redhead D, Hammond P, Garden OJ. Mucinous cystic neoplasms of the pancreas: imaging features and diagnostic difficulties. Clin Radiol 2000;55(3):187–92.
Macrocystic pancreatic lesions: Differentiation of benign from premalignant and malignant cysts by CT Seong Hyun Kim ∗ , Jae Hoon Lim, Won Jae Lee, Hyo K. Lim Department of Radiology and Center for Imaging Science and Samsung Medical Center, Sungkyunkwan University School of Medicine, 50 Ilwon-dong, Gangnam-gu, Seoul 135-710, South Korea Received 3 January 2008; received in revised form 9 March 2008; accepted 12 March 2008
Abstract Objective: To assess useful CT features for differentiating benign from premalignant and malignant macrocystic pancreatic lesions. Methods: Seventy-four patients with pathologically proven macrocystic pancreatic lesions were enrolled: 17 benign cysts (macrocystic serous cystadenoma, n = 12; congenital cyst; n = 5) and 57 premalignant and malignant cysts (mucinous cystic neoplasm, n = 28; intraductal papillary mucinous neoplasm of branch duct type, n = 20; tumor with cystic change, n = 9). Size, location, shape (lobulated, round or oval, or complex cystic with tubular cyst), wall thickness (thin, ≤1 mm; thick, >1 mm), internal surface (smooth or irregular), and other findings were analyzed with multiphasic CT with thin-section (2.5–3 mm) images. CT features between two groups were compared using univariate and multivariate stepwise logistic regression analyses. Results: On univariate analysis, the differences for the shape (p = 0.007), wall thickness (p = 0.011), and internal surface (p = 0.012) between benign and premalignant and malignant cysts were significant. A lobulated shape, a thin wall and a smooth internal surface were more frequent in benign cysts, whereas a round or oval shape or a complex cystic shape with tubular cyst, a thick wall and an irregular internal surface were more frequent in premalignant and malignant cysts. On multivariate analysis, the shape (p = 0.002) and wall thickness (p = 0.025) were significant CT features for differentiating benign from premalignant and malignant cysts. Conclusion: Shape and wall thickness are the main CT features for differentiating benign from premalignant and malignant macrocystic pancreatic lesions. © 2008 Elsevier Ireland Ltd. All rights reserved. Keywords: Pancreas; Pancreas neoplasms; Pancreas cysts; Pancreaticobiliary imaging; CT
1. Introduction Cystic lesions of the pancreas are being increasingly recognized because of frequent use of advanced imaging techniques. Of these, congenital cyst, serous cystadenoma (SCA) with macrocystic variant, mucinous cystic neoplasm (MCN), intraductal papillary mucinous neoplasm (IPMN) of the branch duct type and tumor with cystic change which includes solid pseudopapillary tumor, neuroendocrine tumor and pancreatic ductal adenocarcinoma all may have macrocystic morphologic features that overlap on imaging despite of different malignant potential [1–15]. Benign cysts such as congenital cyst and macrocystic SCA can clinically be managed in a non-surgical manner unless the lesions are symptomatic. In contrast, MCN, IPMN of the
∗
Corresponding author. Tel.: +82 2 3410 2548; fax: +82 2 3410 2559. E-mail address: [email protected] (S.H. Kim).
0720-048X/$ – see front matter © 2008 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.ejrad.2008.03.012
branch duct type and tumor with cystic change are premalignant or malignant, and so they warrant surgical resection [1–11]. CT is still the preferred imaging modality for both initial detection and characterization of these macrocystic pancreatic lesions. Thus, it is crucial to identify the useful CT features for making the differentiation of benign cysts from premalignant and malignant cysts so as to preclude unnecessary surgery because the morbidity after surgical management remains high and the outcomes of loosing pancreatic tissue may be serious. Although a few investigators [5,6] have reported some useful CT features for differentiating macrocystic SCA from mucinous cystadenoma and IPMN of the branch duct type, to the best of our knowledge, no reports are available concerning the usefulness of CT with thin-section images for differentiating benign from premalignant and malignant macrocystic lesions of the pancreas in a large population, along with the pathological proof. The purpose of this study was to assess the useful CT features for differentiating benign from premalignant and malignant macro-
S.H. Kim et al. / European Journal of Radiology 71 (2009) 122–128
cystic lesions of the pancreas by using multiphasic helical CT with thin-section images. 2. Materials and methods 2.1. Patients Institutional review board approval was obtained for this retrospective study and informed consent was waived. Between January 1999 and March 2007, we searched for all the pathologically proven macrocystic pancreatic lesions including pancreatic tumor with cystic or necrotic change more than 90%, which were equal to or larger than 2 cm in diameter, via a review of the archives of our departments of surgery, pathology, and radiology. We found a total of 253 consecutive patients who underwent pancreatic resection for pancreatic cystic lesions during this period. Of these, we excluded the patients (n = 13) with pseudocyst, which had clinical and laboratory evidence of pancreatitis and/or imaging evidence of pancreatic and peripancreatic inflammation, pancreatic ductal calcification or atrophy. The patients with typical microcystic SCA with multiple cysts smaller than 2 cm in diameter and IPMNs with the main duct type or the combined type were also excluded (n = 74). Of the remaining 166 patients, 92 patients were excluded because either CT was not performed or thick-section (thicker than 5 mm) CT images were performed. The remaining 74 patients were finally enrolled. The final diagnosis were macrocystic SCA in 12, simple cyst in two, lymphoepithelial cyst in three, MCN proven as adenoma in 22, borderline in two, and carcinoma in four, IPMN of the branch duct type proven as adenoma in 12, borderline in five, and carcinoma in three, and tumor with cystic change in nine, including solid pseudopapillary tumor in six, well differentiated neuroendocrine tumor in two, and ductal adenocarcinoma in one. We divided the patients into those with benign cysts including congenital cyst and macrocystic SCA (age range, 34–79 years; mean age, 52 years; 5 men, 12 women) and those with premalignant or malignant cysts (n = 57) including MCN, IPMN of the branch duct type and tumor with cystic change (age range, 19–73 years; mean age, 51 years; 19 men, 38 women). 2.2. CT examination All patients underwent contrast-enhanced CT of the upper abdomen. Different CT machines were used during this period; a single-detector helical CT scanner (HiSpeed Advantage; GE healthcare) (n = 15) and MDCT scanners with four (LightSpeed QX/I; GE healthcare) (n = 16), eight (LightSpeed Ultra 8; GE Healthcare) (n = 11), 16 (LightSpeed 16; GE Healthcare) (n = 16), and 40 (Brilliance 40; Philips Medical Systems, Eindhoven, Netherlands) (n = 16) detectors, respectively. The scanning parameters were 120 kVp and 175–189 mAs. After unenhanced CT scanning, a total of 120 mL of nonionic iodinated contrast material (Ultravist 300 [300 mg I/mL iopromide], Schering AG, Berlin, Germany) was administered intravenously with an automatic injector at a rate of 3–4 mL/s. The images were obtained at 35 s, 55 s and 70 s after the initiation of the contrast material injection, which represent the arterial, pancreatic and
123
portal venous phases, respectively. With the single-detector helical CT scanner, the arterial and pancreatic phase images were obtained in a craniocaudal direction with 3 mm slice thickness and a 3 mm reconstruction interval. With the MDCT scanner, the arterial and pancreatic phases images were obtained in a craniocaudal direction with 2.5 mm slice thickness and a 2.5 mm reconstruction interval. The unenhanced and portal venous phase images were obtained with 5 mm slice thickness and a 5 mm reconstruction interval. 2.3. Image analysis All CT images were interpreted on 1536 × 2048 (3 M pixel) monitor (CCL350i, TOTOKU, Japan) of a picture archiving and communication system (PACS) (GE Medical Systems Integrated Imaging Solutions, Mt Prospect, IL, USA) with stack-view capability. Unenhanced and contrast-enhanced images were evaluated together. Subjective visual criteria were used rather than density measurements. Two abdominal radiologists (7- and 27-year experience, respectively), who were blinded to the specific diagnosis and clinical information, reviewed the CT images by consensus to define the following morphological features of the lesions; maximum dimension (measured in centimeters), location (head, neck, body or tail), shape, wall thickness (measured in millimeters), content within the lesion (homogeneous or inhomogeneous), internal surface (smooth or irregular), presence or absence of calcification in the wall and/or septum, septa, mural nodule, dilatation of the main pancreatic duct, peripancreatic fat infiltration, peripancreatic enlarged lymph node or vascular involvement as based on the reports [3,5–23]. The presence of calcification was determined on the unenhanced images. For the size of the cysts, we used 3 cm in diameter as a cut-off value for being relatively small or large [16]. The shape of the lesion was categorized as lobulated, round or oval, and complex cystic with tubular cyst based on the reports [5,6]. The lobulated shape was defined as the presence of rounded contours that could not be described as a well-defined border of the same circle. The round or oval shape was defined as a round or oval configuration with a broad well-defined border. The complex cystic shape with tubular cyst was defined as any shaped cystic lesions with at least one tubular cyst. A tubular cyst was defined as one having a conic section, in which the plane of the tubular cyst was not parallel to the axis, base, or generatrix of the intersection cone as based on the report [6]. The wall of the cyst was measured with using exclusively thin-section (2.5–3 mm) images obtained during the arterial or pancreatic phase by one of the reviewers, with using both a compass and a digital caliper on the PACS. The wall was considered thin if the wall thickness was 1 mm or less, and it was considered thick if it was thicker than 1 mm in diameter for at least a quarter of the lesion circumference as based on the reports [22,23]. If the wall of the cyst was not discriminated, the wall was considered to be imperceptible and included in the group of thin wall [5]. Any irregularity of the internal surface was classified as irregular. Peripancreatic lymphadenopathy was diagnosed when one or more peripancreatic lymph nodes with a short-axis diameter of 10 mm or more were identified. The peripancreatic fat
124
S.H. Kim et al. / European Journal of Radiology 71 (2009) 122–128
infiltration was defined if an increased attenuation relative to that of subcutaneous fat was identified. The vascular involvement was defined when the occlusion of any nearby veins with engorged collaterals or irregular narrowing or apparent encasement of any nearby arteries was seen. 2.4. Statistical analysis The mean maximum dimensions of two groups were compared using Mann–Whitney U test. For univariate analysis, statistical differences of the CT features of two groups were analyzed with the Chi-square test and Fisher’s exact test. A multivariate stepwise logistic regression model was used to determine the most significant CT features differentiating two groups. Significant differences were defined as p values less than 0.05. All statistical analyses were performed with the SPSS software package (version 11.5; SPSS, Chicago III). 3. Results 3.1. Univariate analysis Table 1 summarizes the different CT features observed in two groups. The benign cysts ranged in maximum dimension from 2.0 to 9.5 cm (mean ± standard deviation, 4.3 ± 2.2 cm)
Fig. 1. A 57-year-old woman with a macrocystic serous cystadenoma in the head of the pancreas. CT image at pancreatic phase shows a lobulated cystic lesion (arrow) with the imperceptible wall, homogeneous content, and smooth internal surface in the head of the pancreas. Note a thin septum (arrowhead).
and this was similar to premalignant and malignant cysts (range, 2.0–12 cm; mean ± standard deviation, 4.7 ± 2.9 cm) (p = 0.969). The lobulated shape was more frequently seen in benign cysts (Figs. 1 and 2) than in premalignant and malignant cysts (Figs. 3 and 4). In contrast, the round or
Table 1 CT features of benign and premalignant and malignant cysts of the pancreas CT feature
Benign cyst (n = 17)
Premalignant and malignant cyst (n = 57)
Size ≤3 cm >3 cm
7 (41) 10 (59)
25 (44) 32 (56)
Location Head Body or tail
9 (53) 8 (47)
17 (30) 40 (70)
Shape* Lobulated Round or oval Complex cystic with tubular cyst
9 (52.9) 8 (47.1) 0
12 (21.1) 30 (52.6) 15 (26.3)
p 0.845
0.080
0.007
Wall thickness* ≤1 mm (thin) >1 mm (thick)
14 (82) 3 (18)
27 (47) 30 (53)
Content Homogeneous Inhomogeneous
15 (88) 2 (12)
39 (68) 18 (32)
Internal surface Smooth Irregular
15 (88) 2 (12)
31 (54) 26 (46)
Calcification Septa Mural nodule Dilatation of main pancreatic duct Peripancreatic fat infiltration Peripancreatic lymphadenopathy Vascular involvement
3 (18) 11 (65) 1 (6) 2 (12) 0 0 0
16 (28) 48 (84) 11 (19) 13 (23) 6 (11) 4 (7) 7 (12)
0.011
0.131
0.012
0.533 0.095 0.274 0.496 0.326 0.568 0.192
Note: Numbers in parentheses are percentages. * On multivariate stepwise logistic regression analysis, shape (p = 0.002; 95% confidence interval [1.460, 5.805]) and wall thickness (p = 0.025; 95% confidence interval [1.199, 15.826]) were statistically significant.
S.H. Kim et al. / European Journal of Radiology 71 (2009) 122–128
Fig. 2. A 53-year-old woman with a macrocystic serous cystadenoma in the tail of the pancreas. CT image at pancreatic phase shows a lobulated cystic lesion (long arrow) with the thin wall (short arrow) in the tail of the pancreas. Note septa and homogeneous content, smooth internal surface.
oval shape (Figs. 5–7) was more frequently seen and the complex cystic shape with tubular cyst (Fig. 8) was exclusively seen in premalignant and malignant cysts as compared with benign cysts (Figs. 9–11). The differences of the shape between two groups were significant (p = 0.007). The wall of the cysts were considered thick (>1 mm) in three benign cysts (Figs. 9 and 11) and in 30 premalignant and malignant cysts (Figs. 5–7), and this difference was significant (p = 0.011). The smooth internal surface was more frequently seen in benign cysts (Figs. 1, 2, 9–11) than in premalignant and malignant cysts (Figs. 3 and 4) (p = 0.012). No significant differences were found between two groups with respect to a size difference with 3 cm being used as cut-off value (p = 0.845), location (p = 0.080), content (p = 0.131), and the presence of calcification (p = 0.533), septa (p = 0.095), mural nodule (p = 0.274), and dilatation of the main pancreatic duct (p = 0.496). Peripancre-
Fig. 3. A 73-year-old woman with a mucinous cystadenoma of the pancreas. CT image at arterial phase shows a lobulated cystic lesion (arrow) with the imperceptible wall and smooth internal surface, which especially mimics such a benign cyst as a macrocystic serous cystadenoma, in the body of the pancreas.
125
Fig. 4. A 69-year-old woman with an intraductal papillary mucinous adenoma of the branch duct type in the pancreas. CT image at pancreatic phase shows a lobulated cystic lesion (arrow) with the imperceptible wall and smooth internal surface, which especially mimics such a benign cyst as a macrocystic serous cystadenoma, in the head of the pancreas.
atic abnormalities such as fat infiltration, lymphadenopathy and vascular involvement (Figs. 6 and 7) were only seen in six, four, and seven premalignant and malignant cysts, respectively, but no statistical significance was found compared with benign cysts. 3.2. Multivariate analysis Of three significant CT features on univariate analysis, multivariate stepwise logistic regression analysis showed that the shape (p = 0.002; 95% confidence interval [1.460, 5.805]) and wall thickness (p = 0.025; 95% confidence interval [1.199, 15.826]) were statistically significant CT features for differentiating benign from premalignant and malignant cysts.
Fig. 5. A 26-year-old woman with a mucinous cystadenoma of the pancreas. CT image at pancreatic phase shows a septated cystic lesion with round shape (arrow), surrounded by the thick wall in the head of the pancreas. The internal surface appears irregular (arrowheads).
126
S.H. Kim et al. / European Journal of Radiology 71 (2009) 122–128
Fig. 6. A 28-year-old woman with a mucinous cystadenocarcinoma of the pancreas. CT image at pancreatic phase shows a septated cystic lesion of the oval shape (long black arrow) with the thick wall and irregular internal surface (long white arrows) in the tail of the pancreas. The inhomogeneous content and calcifications (short arrows) in the septum and the wall are shown. Note perigastric engorged veins (arrowheads) due to the occlusion of the splenic vein (not shown).
4. Discussion The main clinical problem for the macrocystic pancreatic lesions is not detecting them, but rather, it is characterizing them, followed by deciding on the most appropriate therapeutic plan. However, neither imaging studies nor even cyst aspiration gives the physician the precise information to differentiate each benign, premalignant and malignant cystic lesion of the pancreas [4]. Therefore, it is clinically valuable to identify the useful CT features for making the differentiation of benign macrocystic lesions that can be closely followed unless the lesions are symptomatic, from premalignant and malignant macrocystic lesions that warrant surgical resection. Pseudocysts are the most common pancreatic cystic lesions [8]. Although pseudocysts and cystic neoplasms of the pancreas
Fig. 7. A 19-year-old woman with a solid pseudopapillary tumor of the pancreas. CT image at pancreatic phase shows a round cystic lesion (long arrow) with the thick wall and irregular internal surface (short arrows) in the tail of the pancreas. Note a solid component (asterisk), calcifications (black arrowheads) in septa, and perigastric engorged veins (white arrowheads).
Fig. 8. A 72-year-old man with an intraductal papillary mucinous adenoma of the branch duct type of the pancreas. (a) CT image at pancreatic phase shows a cystic lesion (arrow) with septa (arrowhead) and imperceptible wall, which is formed by cystically dilated branch duct in the head of the pancreas. Note mild dilatation of the main pancreatic duct. (b) The Coronal reformatted CT image shows a complex cystic shape with tubular cyst (arrowheads) with the irregular internal surface (short arrow) and mural nodule (long arrow).
Fig. 9. A 50-year-old woman with a macrocystic serous cystadenoma in the pancreas. CT image at pancreatic phase shows a round cystic lesion (arrow) with the thick wall, mimicking premalignant or malignant cyst, in the tail of the pancreas.
S.H. Kim et al. / European Journal of Radiology 71 (2009) 122–128
Fig. 10. A 36-year-old woman with a simple cyst of the pancreas. CT image at pancreatic phase shows a round cystic lesion (arrow) with the imperceptible wall and smooth internal surface in the body of the pancreas.
can mimic each other on imaging [24], a careful clinical history of pancreatitis with laboratory clues (amylase or lipase), and/or imaging evidence of pancreatic and peripancreatic inflammation, pancreatic ductal calcification or atrophy may allow diagnosis of a cystic lesion as a pseudocyst [3,8,9,12,18]. In our study, we excluded 13 patients with pseudocysts because the patients had clinical and laboratory evidence of pancreatitis and/or the imaging evidence of pancreatic and peripancreatic inflammation, pancreatic ductal calcification or atrophy. Instead, the patients with benign congenital cysts such as simple cyst and lymphoepithelial cyst, and premalignant or malignant IPMNs of the branch duct type were included in the study, and all of these may mimic other macrocystic lesions of the pancreas [1–3,6,9,10,13,19–21]. The present study showed that the shape and wall thickness were the main CT features for differentiating benign from premalignant and malignant macrocystic lesions of the pancreas.
127
Some investigators reported that the shape of macrocystic pancreatic neoplasms was useful for differentiating macrocystic SCAs, MCNs and IPMNs of the branch duct type from each other [5,6]. In our study, the lobulated shape was a common feature in benign cysts and the complex cystic shape with tubular cyst was only seen in premalignant and malignant cysts. We believe that these results are due to a large proportion of macrocystic SCA within benign cysts and due to premalignant or malignant IPMNs of the branch duct type which appear as multiple separate cysts with each cystic locule representing a dilated, tortuous pancreatic duct resulting in a tubular cyst [2,3,5,6,10–12]. In our study, a round or oval shape was slightly more frequent in premalignant and malignant cysts than in benign cysts. We think that this result may be influenced by MCNs and tumors with cystic change [5,6,13,19–21]. Some investigators have reported that benign cysts such as macrocystic SCAs had a thin wall that measured 0.1 cm in thickness [22,23] and no predominant enhancement of the wall of these lesions was seen [5,7]. In our study, we often found the wall of benign cysts appeared to be imperceptible, equal to or less than 1 mm in diameter, as compared with premalignant and malignant cysts being thick-walled on thin-section CT. MCNs, IPMNs of the branch duct type and tumors with cystic change may more frequently have an irregular internal surface compared to macrocystic SCAs and congenital cysts because MCNs and IPMNs of the branch duct type may have papillary excrescences into the cystic lumen, and tumor with cystic change may frequently be caused by necrosis or hemorrhage; moreover, all these lesions may have mural nodule or a solid component [3,6,9,10,12,13,15]. Our results showed that the presence or absence of irregular internal surface was a helpful CT feature for differentiating benign from premalignant and malignant cysts. CT features such as size, location, calcification, septa and mural nodule (or a solid component) may be helpful in differentiating benign and malignant cystic neoplasms [3,9,10,14,16,17]. However, in our study, these CT features were not statistically different between benign and premalignant and malignant cysts. The imaging overlap between cystic lesions or less frequent occurrence of these findings may be possible reasons on these results. In our study, peripancreatic abnormalities such as fat infiltration, lymphadenopathy and vascular involvement were only seen in premalignant and malignant cysts. We think that these CT features may be of additional value for differentiating benign from premalignant and malignant cysts. Our study has some limitations. First, we retrospectively investigated CT features that could facilitate differentiation between benign and premalignant and malignant macrocystic pancreatic lesions that were pathologically proven, which might be a bias in case selection. Second, the enrolled patients were collected over 8 years. Although only the CT images of optimal quality were included, the imaging parameters did vary. 5. Conclusion
Fig. 11. A 50-year-old man with a lymphoepithelial cyst of the pancreas. CT image at pancreatic phase shows an oval cystic lesion (arrow) with the thick wall and septum (arrowhead), mimicking a premalignant or malignant cyst, in the body of the pancreas.
Our results show that the shape and wall thickness are significant independent CT features for differentiating benign from premalignant and malignant macrocystic pancreatic lesions, in
128
S.H. Kim et al. / European Journal of Radiology 71 (2009) 122–128
which a lobulated shape and a thin wall (≤1 mm) are more frequently seen in benign macrocystic pancreatic lesions and a round or oval shape or a complex cystic shape with tubular cyst and a thick wall (>1 mm) are more frequently seen in premalignant and malignant macrocystic pancreatic lesions. References [1] Sakorafas GH, Sarr MG. Cystic neoplasms of the pancreas; what a clinician should know. Cancer Treat Rev 2005;31(7):507–35. [2] Klimstra DS. Cystic, mucin-producing neoplasms of the pancreas: the distinguishing features of mucinous cystic neoplasms and intraductal papillary mucinous neoplasms. Semin Diagn Pathol 2005;22(4):318–29. [3] Demos TC, Posniak HV, Harmath C, Olson MC, Aranha G. Cystic lesions of the pancreas. AJR 2002;179(6):1375–88. [4] Spinelli KS, Fromwiller TE, Daniel RA, et al. Cystic pancreatic neoplasms: observe or operate. Ann Surg 2004;239(5):651–9. [5] Cohen-Scali F, Vilgrain V, Brancatelli G, et al. Discrimination of unilocular macrocystic serous cystadenoma from pancreatic pseudocyst and mucinous cystadenoma with CT: initial observations. Radiology 2003;228(3):727–33. [6] Kim SY, Lee JM, Kim SH, et al. Macrocystic neoplasms of the pancreas: CT differentiation of serous oligocystic adenoma from mucinous cystadenoma and intraductal papillary mucinous tumor. AJR 2006;187(5):1192–8. [7] Khurana B, Mortele KJ, Glickman J, Silverman SG, Ros PR. Macrocystic serous adenoma of the pancreas: radiologic–pathologic correlation. AJR 2003;181(1):119–23. [8] Curry CA, Eng J, Horton KM, et al. CT of primary cystic pancreatic neoplasms: can CT be used for patients triage and treatment? AJR 2000;175(1):99–103. [9] Sahani DV, Kadavigere R, Saokar A, Fernandez-del Castillo C, Brugge WR, Hahn PF. Cystic pancreatic lesions: a simple imaging-based classification system for guiding management. Radiographics 2005;25(6):1471– 84. [10] Kawamoto S, Horton KM, Lawler LP, Hruban RH, Fishman EK. Intraductal papillary mucinous neoplasm of the pancreas: can benign lesions be differentiated from malignant lesions with mulidector CT? Radiographics 2005;25(6):1451–70.
[11] Chiu SS, Lim JH, Lee WJ, et al. Intraductal papillary mucinous tumour of the pancreas: differentiation of malignancy and benignancy by CT. Clin Radiol 2006;61(9):776–83. [12] Kim YH, Saini S, Sahani D, Hahn PF, Mueller PR, Auh YH. Imaging diagnosis of cystic pancreatic lesions: pseudocyst versus nonpseudocyst. Radiographics 2005;25(3):671–85. [13] Park M, Kim KW, Lim JS, et al. Unusual cystic neoplasms in the pancreas: radiologic–pathologic correlation. J Comput Assist Tomogr 2005;29(5):610–6. [14] Procacci C, Carbognin G, Accordini S, et al. CT features of malignant mucinous cystic tumors of the pancreas. Eur Radiol 2001;11(9):1626–30. [15] Buetow PC, Rao P, Thompson LD. Mucinous cystic neoplasms of the pancreas: radiologic–pathologic correlation. Radiographics 1998;18(2):433–49. [16] Sahani DV, Saokar A, Hahn PF, Brugge WR, Fernandez-del Castillo C. Pancreatic cysts 3 cm or smaller: how aggressive should treatment be? Radiology 2006;238(3):912–9. [17] Procacci C, Graziani R, Bicego E, et al. Serous cystadenoma of the pancreas: report of 30 cases with emphasis on the imaging findings. J Comput Assist Tomogr 1997;21(3):373–82. [18] Procacci C, Biasiutti C, Carbognin G, et al. Characterization of cystic tumors of the pancreas: CT accuracy. J Comput Assist Tomogr 1999;23(6):906–12. [19] Fiamingo P, Veroux M, Gringeri E, et al. True solitary pancreatic cyst in an adult: report of a case. Surg Today 2005;35(11):979–83. [20] Kim YH, Auh YH, Kim KW, Lee MG, Kim KS, Park SY. Lymphoepithelial cysts of the pancreas: CT and sonographic findings. Abdom Imaging 1998;23(2):185–7. [21] Shinmura R, Gabata T, Matsui O. Lymphoepithelial cyst of the pancreas: case report with special reference to imaging–pathologic correlation. Abdom Imaging 2006;31(1):106–9. [22] Huh JR, Chi JG, Jung KC, Choe KJ, Yoon YB. Macrocystic serous cystadenoma of pancreas: a case report. J Korean Med Sci 1994;9(1):78–85. [23] Fujiwara H, Ajiki T, Fukuoka K, Mitsutsuji M, Yamamoto M, Kuroda Y. Macrocystic serous cystadenoma of the pancreas. J Hepatobiliary Pancreat Surg 2000;7(1):92–6. [24] Scott J, Martin I, Redhead D, Hammond P, Garden OJ. Mucinous cystic neoplasms of the pancreas: imaging features and diagnostic difficulties. Clin Radiol 2000;55(3):187–92.