Diagnosis and Management of Cystic Pancreatic Lesions

Transcription

1 Gastrointestinal Imaging Best Practices/Review Sahani et al. Cystic Pancreatic Lesions Gastrointestinal Imaging Best Practices/Review Dushyant V. Sahani 1 Avinash Kambadakone 1 Michael Macari 2 Noaki Takahashi 3 Suresh Chari 4 Carlos Fernandez-del Castillo 5 Sahani DV, Kambadakone A, Macari M, Takahashi N, Chari S, Fernandez del Castillo C Keywords: cystic pancreatic lesion, MDCT, MRI DOI: /AJR Received February 25, 2012; accepted after revision July 18, Department of Radiology, Division of Abdominal Imaging and Intervention, Massachusetts General Hospital, Harvard Medical School, 55 Fruit St, White 270, Boston, MA Address correspondence to D. V. Sahani (dsahani@partners.org). 2 Department of Radiology, Division of Abdominal Imaging, New York University Langone Medical Center, New York, NY. 3 Department of Radiology, Division of Abdominal Imaging, Mayo Clinic, Rochester, MN. 4 Department of Medicine, Division of Gastroenterology, Mayo Clinic, Rochester, MN. 5 Department of Surgery, Division of Pancreatic Surgery, Massachusetts General Hospital, Harvard Medical School, Boston, MA. AJR 2013; 200: X/13/ American Roentgen Ray Society Diagnosis and Management of Cystic Pancreatic Lesions OBJECTIVE. The purpose of this review is to outline the management guidelines for the care of patients with cystic pancreatic lesions. CONCLUSION. The guidelines are as follows: Annual imaging surveillance is generally sufficient for benign serous cystadenomas smaller than 4 cm and for asymptomatic lesions. Asymptomatic thin-walled unilocular cystic lesions smaller than 3 cm or side-branch intraductal papillary mucinous neoplasms should be followed up with CT or MRI at 6 and 12 months interval after detection. Cystic lesions with more complex features or with growth rates greater than 1 cm/year should be followed more closely or recommended for resection if the patient s condition allows surgery. Symptomatic cystic lesions, neoplasms with high malignant potential, and lesions larger than 3 cm should be referred for surgical evaluation. Endoscopic ultrasound with fine-needle aspiration (FNA) biopsy can be used preoperatively to assess the risk of malignancy. Clinical Vignettes and Images The increasing use and improved spatial and contrast resolution of advanced crosssectional imaging techniques such as MDCT and MRI have resulted in a marked increase in the incidental detection of cystic pancreatic lesions. They are encountered in as many as 2.6% of abdominal MDCT examinations and 20% of MRI studies [1 3]. Larger cystic pancreatic lesions are typically symptomatic, and incidentally detected cystic pancreatic lesions are often small. Increased identification of cystic pancreatic lesions at MDCT and MRI presents a clinical conundrum for appropriate further management [4 7]. Accurate characterization of these cystic lesions is essential for further management, either surgical or conservative. In a select group of patients, endoscopic ultrasound and cyst aspiration can be performed for further characterization. Clinical vignettes are presented in Figures 1 4. The Imaging Question How do we characterize, diagnose, and appropriately manage incidentally found cystic pancreatic lesions? Background and Importance Cystic pancreatic lesions encompass a varied group of pancreatic abnormalities, includ- ing inflammatory (pseudocysts), benign (serous cystadenomas), precancerous (intraductal papillary mucinous neoplasms [IPMNs] and mucinous cystic neoplasms [MCNs]), and frankly malignant (cystadenocarcinomas) [8 12]. Cystic pancreatic lesions not only have diverse histologic and imaging appearances but also differ in clinical presentation, biologic behavior, growth pattern, and risk of malignancy (Table 1). Accurate risk stratification and decisions on treatment and follow-up strategy necessitate precise lesion characterization and diagnosis [2, 13 16]. The current management of common cystic pancreatic lesions is summarized in Table 2. Synopsis and Synthesis of Evidence The most common nonneoplastic cystic pancreatic lesions are pseudocysts, which usually arise as a sequela of pancreatitis or trauma. The most common cystic pancreatic neoplasms are IPMNs, MCNs, and serous cystaden omas (SCAs) [12, 17 20]. Although SCAs and pseudocysts are considered benign, IPMNs and MCNs have malignant potential [19, 21 23]. Other cystic pancreatic lesions account for fewer than 10% of cases and include uncommon pathologic findings such as solid pseudopapillary neoplasms, cystic pancreatic neuroendocrine neoplasms, cystic degeneration in other solid pancreatic neoplasms, lymphoepithelial AJR:200, February

2 Sahani et al. cysts, and cystic adenocarcinoma of the pancreas [5, 24, 25]. Diagnosis Imaging plays a crucial role in the management of cystic lesions of the pancreas, including lesion detection and characterization. Technologic innovations in MDCT and MRI have led to improvement in analysis and morphologic differentiation of cystic pancreatic lesions and are widely considered the primary imaging modalities in the care of patients with cystic lesions of the pancreas. In addition, advances in postprocessing have enabled enhanced definition of the extent of a lesion and its relation to adjacent structures. These techniques are particularly valuable in delineating the relation between the cystic lesion and the pancreatic duct, a key feature in differentiating sidebranch IPMNs from other cystic lesions [24, 26, 27]. The following imaging modalities can be used either independently or in combination to help in the diagnosis and management of cystic pancreatic lesions. A C MDCT MDCT is the primary modality for imaging of cystic pancreatic lesions, including IPMNs, owing to its high spatial and temporal resolution, speed of acquisition, wide availability, and ease of interpretation [2, 15, 16, 28]. The superior quality of 2D and 3D image displays generated from isotropic MDCT datasets has facilitated excellent depiction of the detailed regional pancreatic anatomy and precise definition of the morphologic characteristics of cysts [26, 29, 30]. Image postprocessing in a desired plane also allows determination of the communication between the cystic lesion and the main pancreatic duct, a key feature in the diagnosis of side-branch IPMNs [26]. However, the presence of a small duct or a collapsed duct can impede visualization of ductal communications [26]. MDCT has a reported accuracy of 56 85% for characterization of cystic pancreatic lesions, which is comparable to that of MRI [2, 15, 16, 28, 31]. Visser et al. [30] found that MDCT had an accuracy of 76 82% in establishing the diagnosis of malignancy in B D 58 histopathologically proven cystic pancreatic masses. In a study of 100 cystic pancreatic lesions, Chaudhari and colleagues [32, 33] reported an accuracy of 71 79% for MDCT for discriminating premalignant or malignant lesions from benign lesions. Lee et al. [2] reported a comparable accuracy ( %) of MDCT for differentiating benign from malignant cystic pancreatic lesions. However, subclassification of cystic lesions into histopathologic types is often difficult because of overlapping imaging features. Increasingly, the morphologic pattern depicted on MDCT images is being used to categorize cystic pancreatic lesions broadly into mucinous and nonmucinous types and then subdivide them on the basis of complex features into aggressive and nonaggressive lesions [2, 15, 22]. In a study of 114 patients with 130 cystic pancreatic lesions, Sahani and colleagues [15] stratified the lesions into mucinous and nonmucinous subtypes with an accuracy of 82 85% and reported an accuracy of 85 86% for recognizing aggressive biologic features. Similar accuracy has Fig. 1 MDCT images in four patients with incidental finding of pancreatic cystic lesion. A, 62-year-old-man with 1.2-cm cystic lesion (arrow) in pancreatic body without main pancreatic ductal dilatation or enhancing solid components. Coronal MDCT image shows main pancreatic duct intimately associated with cystic lesion, and communication between cystic lesion and main duct is evident. Lesion is typical of nonaggressive side branch intraductal papillary mucinous neoplasm that does not warrant surgical resection. B, 69-year-old woman with cystic lesion in pancreatic tail. Axial MDCT image shows cystic lesion in pancreatic tail (white arrows) with enhancing solid components (black arrow) and peripheral rim calcification. Findings are indicative of aggressive mucinous cystic lesion. Patient underwent distal pancreatectomy, and histopathologic examination revealed mucinous cystadenocarcinoma. C, 65-year-old woman with recurrent dull, aching epigastric pain. Axial MDCT image shows 3.5-cm microcystic lesion (arrow) in pancreatic body and tail with multiple small loculations. Patient underwent distal pancreatectomy because of considerable symptoms. Histopathologic finding was serous cystadenoma. D, 67-year-old man with pancreatic ductal dilatation. Curved multiplanar reformatted MDCT image shows main pancreatic ductal dilatation (arrows) in region of head and proximal body of pancreas with prominence of pancreatic tail indicative of main duct intraductal papillary mucinous neoplasm. Patient underwent Whipple procedure. Histopathologic finding was main duct intraductal papillary mucinous neoplasm with ductal carcinoma in situ. 344 AJR:200, February 2013

3 Cystic Pancreatic Lesions Fig year-old man with incidentally detected lesion in pancreatic tail. A, Axial MDCT image shows 8-mm low-attenuation lesion (arrow) in pancreatic tail. MRI was considered for evaluation of internal morphologic features of lesion and for assessing any communication with main pancreatic duct. B, Axial T2-weighted MR image shows T2 hyperintense lesion in tail of pancreas with mild nodular wall thickening (arrow) along lateral wall. C, Coronal T2-weighted MR image shows septate cyst (arrow). D, Axial T2-weighted gadolinium-enhanced MR image shows enhancing nodular thickening (arrow) of lateral wall. Distal pancreatectomy was performed because of complex morphologic features of cyst, especially enhancing solid component, and young age. Histopathologic finding was side-branch intraductal papillary mucinous neoplasm with lowgrade dysplasia. been reported for small cystic pancreatic lesions ( 3 cm). Sainani et al. [16] found that MDCT had % accuracy for differentiating mucinous and nonmucinous subtypes of small cystic pancreatic lesions. The presence of solid nodules, thick septations, and cyst wall thickening on MDCT images favors the diagnosis of an aggressive cystic lesion [2, 15, 16, 22, 30, 31]. Sahani and colleagues [15] reported that pancreatic protocol MDCT had sensitivities of 93.6%, 71.4%, and 86.4% for detecting morphologic features such as septa, mural nodules, and main pancreatic duct communication. Kim et al. [34] found that shape and wall thickness (> 1 mm) were two independent predictors of malignancy of a macrocystic pancreatic lesion. Tomimaru et al. [35] reported that the presence or absence of mural nodules on CT images had a sensitivity, specificity, and accuracy of 93%, 80%, and 86% in the diagnosis of malignant IPMNs. Sainani and colleagues [16] found that in the detection of small cystic pancreatic lesions ( 3 cm), MDCT had 73.9% sensitivity for the assessment of septa and 86% sensitivity for depiction of ductal communication. MDCT has the additional advantage of depicting calcifications, which can be difficult to recognize on MR images. Despite its improved performance in the assessment of the biologic characteristics of pancreatic cysts, false-negative results can occur because the dysplastic changes in the cystic lesions do not have distinct MDCT features [15]. In addition, MDCT has limited utility for differentiating minimally invasive carcinoma from carcinoma in situ. Similarly, the recognition of internal details and pancreatic duct communication in a small cystic lesion can be challenging with MDCT. Sainani et al. [16] reported that MRI had higher sensitivity than MDCT in showing ductal communication of small cystic pancreatic lesions (100% vs 85.7%). In addition, inflammatory changes from concurrent pancreatitis can obscure the morphologic details of cystic pancreatic lesions. A C MRI and MRCP MRI of the pancreas with MRCP has emerged as a reliable tool for detecting and characterizing cystic pancreatic lesions. The superior soft-tissue and contrast resolution makes MRI a sensitive study for assessing the morphologic features of cystic lesions, including their communication with the main pancreatic duct [3, 16, 30, 36, 37]. Visser and colleagues [30] found that MRI had an accuracy of 85 91% in establishing the diagnosis of malignancy in cystic pancreatic lesions. Lee et al. [2] found that MRI had an accuracy of % in determining the malignancy of cystic pancreatic lesions. In particular, in small cystic lesions ( 3 cm), MRI facilitates confident assessment of the morphologic features of the cyst and reliably displays small cystic lesions not obvious on MDCT images [16]. Sainani et al. [16] reported that MRI had an accuracy of % for differentiating mucinous and nonmucinous subtypes of small pancreatic cystic lesions ( 3 cm). They also reported that MRI had a sensitivities of 91% and 100% in the assessment of septa and main pancreatic duct communication in small cystic pancreatic lesions. The transition from 2D software to higherquality 3D acquisition has resulted in more effective detection of connections with the main pancreatic duct compared with the 2D singleslab technique [37, 38]. Yoon et al. [38] found that compared with 2D MRCP, 3D MRCP facilitated superior evaluation of the pancreatic duct and the morphologic details of IPMNs. The 2D MRCP sequence is usually performed as a breath-hold coronal single-shot fast spinecho sequence or HASTE sequence [37, 38]. The 3D imaging technique is a high-spatialresolution MRCP sequence that entails either a breath-hold turbo spin-echo sequence or a respiratory-triggered fast spin-echo approach B D AJR:200, February

4 Sahani et al. [37, 38]. An additional advantage of MRI with or without MRCP is in the follow-up of young patients with cystic pancreatic lesions because MRI eliminates exposure to ionizing radiation. A C B D Although MRCP is more sensitive in displaying the details of cystic lesions, in most cases appropriately performed thin-section MDCT in combination with image processing (multiplanar reconstructions and curved reformations) can provide sufficient detail on cystic lesions to allow decision making [15, 16]. Variants in anatomy of the pancreatic ductal system can be confidently defined with both MDCT and MRCP. This capability is important in cases of ductal anatomic variants such as pancreas divisum. The presence of such an anomaly can influence the surgical approach. A main duct IPMN affecting the dorsal duct can be treated with newer surgical techniques involving dorsal pancreatectomy and sparing the ventral pancreas, thus avoiding biliary and pancreatic anastomoses [39]. Secretin-enhanced MRCP is a modified MRCP technique in which MRI is performed after stimulation of pancreatic exocrine function by IV injection of secretin [36, 37, 40]. Through stimulation of pancreatic secretion, secretin administration can improve the utility Fig year-old woman with incidentally detected cyst in pancreatic head. A, Axial T2-weighted MR image shows 17-mm hyperintense lesion (arrow) in pancreatic head. B, Gadolinium-enhanced T1-weighted fat-saturated MR image shows suspicious nodular enhancement (arrow) along cyst wall. Close imaging follow-up was performed. C and D, Follow-up MR images 1 year after B show nodule (arrow) within lesion on T2-weighted image (C) that was enhancing on gadolinium-enhanced T1-weighted fat-saturated image (D). Middle pancreatectomy was performed because of enlarging enhancing component in cyst. Histopathologic finding was side-branch intraductal papillary mucinous neoplasm with low- to moderate-grade dysplasia. Operative decision was based on development of suspicious features on follow-up images. Case falls into category of cyst with solid component. A B C Fig year-old woman undergoing follow-up of side-branch intraductal papillary mucinous neoplasm in pancreas. A, Axial T2-weighted MR image shows 12-mm T2 hyperintense lesion (arrow) in pancreatic neck that had communication with pancreatic duct on MRCP images (not shown). No enhancing solid components or main ductal dilatation was seen. Yearly surveillance with MRI was prescribed. B, Axial T2-weighted MR image 1 year after A shows stability of lesion size (arrow) and no development of suspicious features. C, Axial T2-weighted MR image 2 years after A shows stability of side-branch intraductal papillary mucinous neoplasm (arrow). T2 hyperintense simple cyst in left kidney is incidental finding. 346 AJR:200, February 2013

5 Cystic Pancreatic Lesions TABLE 1: General Features and Imaging Appearances of Commonly Encountered Pancreatic Cystic Lesions Cystic Pancreatic Neuroendocrine Neoplasm Intraductal Papillary Mucinous Neoplasm Main Duct Intraductal Papillary Mucinous Neoplasm Side Branch Characteristic Pseudocyst Sex M > F F > M (75%/25%) F a M > F (60%/40%) M > F (60%/40%) F b F = M Age (decade) 3rd 7th 6th 7th 4th 5th 6th 7th 6th 7th 2nd 4th 5th 6th Location Head, body, tail Head, body, tail Body, tail Head, body, tail Head, body, tail Body, tail Head, body, tail Oval Oval Solid Pseudopapillary Neoplasm c Mucinous Cystic Neoplasm Serous Cystadenoma Shape Variable Lobulated Oval Bunch of grapes Diffuse pancreatic duct dilatation Present (thin) Present (usually thick) Present Absent Present Present (thick) Wall Present (usually thin, thick if infected) Macrocystic NA Unilocular NA Macrocystic (< 6, each > 2 cm) Microcystic (> 6, each < 2 cm) Loculation Unilocular/ multilocular Central scar Absent Present (20 30%) Absent Absent Absent Absent Absent Calcifications Uncommon (rim) Present (central) Present (peripheral) Uncommon (septal) Uncommon Uncommon Present NA Absent Absent Uncommon Absent Absent Usually present as a channel Main pancreatic duct communication Present Absent Absent Absent Absent Present in mixed intraductal papillary mucinous neoplasms and large lesions Involved in chronic pancreatitis Main pancreatic duct involvement No specific imaging features Large size, local invasion or enlarged nodes (all have very low malignant potential) Main pancreatic duct > 10 mm; nodules present Size > 3 cm; main duct dilatation (> 6 mm); thick irregular wall, septa; mural nodules None Typically benign Solid areas and irregular wall; peripheral calcification Imaging predictors of malignancy Note Data from [14, 18 20, 24, 25, 27, 60, 69, 70]. NA = not applicable. a Mucinous cystic neoplasms occur almost exclusively in women [19]. b More than 90% of patients with solid pseudopapillary neoplasm are women [25]. c All solid pseudopapillary neoplasms have very low malignant potential. of MRCP in evaluating ductal anatomy and the communication of small cystic pancreatic lesions [36, 37, 40]. However, the clinical benefit of secretin MRCP in the care of patients with IPMN and other cystic lesions is currently unknown. PET and PET/CT As in a number of other malignancies, PET has a potential advantage in detection of metastatic spread of invasive pancreatic neoplasms [41]. Hybrid PET/CT with 18 F-FDG allows assessment of tumor extent and microinvasion in minimally invasive disease, which can be missed with other imaging techniques [42, 43]. The sensitivity of PET/CT for detecting carcinoma in situ and borderline lesions remains unsatisfactory [44, 45]. A few studies have shown that PET/CT performs marginally better than MDCT alone in the detection and characterization of malignant cystic pancreatic neoplasms [35, 46]. In a study of 72 patients, Tomimaru et al. [35] found that with a cutoff maximum standardized uptake value of 2.5, FDG PET had sensitivity, specificity, and accuracy of 93%, 100%, and 96% in the diagnosis of malignant IPMN. Likewise, Sperti and colleagues [42] found that FDG PET was useful in the differentiation of benign and malignant IPMNs with a specificity, sensitivity, and accuracy of 92%, 97%, and 95%. Mansour et al. [45] found that PET had sensitivity and specificity of 57% and 85% in determination of the malignancy of pancreatic cystic tumors. The major limitations of PET/CT in the evaluation of pancreatic cystic lesions include higher cost, false-negative results for borderline and in situ tumors, and false-positive uptake in areas of lesion-associated pancreatitis and postbiopsy changes [35, 46]. Therefore, there are not enough data to justify a role of PET/CT in the characterization of cystic pancreatic lesions [47]. Endoscopic Ultrasound Endoscopic ultrasound is an excellent imaging technique for detecting signs predictive of malignancy or aggressiveness in cystic pancreatic lesions. Such signs include internal septations, mural nodules, solid masses, vascular invasion, and lymphatic metastasis [48, 49]. An additional benefit is its capability for sampling fluid and solid components and depicting debris and wall thickness [50]. Endoscopic ultrasound also shows the details of the pancreatic parenchyma and the pancreatic duct. In a study involving 50 patients, Kim et al. [51] found that endoscopic ultrasound had sensitivity, specificity, and accuracy of AJR:200, February

6 Sahani et al. TABLE 2: Management of Commonly Encountered Cystic Lesions of the Pancreas Lesion Malignant Potential Recommendation Pseudocyst None Referral to gastroenterologist or pancreatic surgeon if lesion is symptomatic Serous cystadenoma Very low (malignant lesion is termed serous cystadenocarcinoma) Serial imaging annually for 3 y; referral to surgeon if lesions is symptomatic or larger than 4 cm; for patients at poor surgical risk, endoscopic ultrasound (fine-needle aspiration to confirm diagnosis and rule out malignancy) Mucinous cystic neoplasm 6 36% prevalence of invasive carcinoma [19] (malignant lesion is termed mucinous cystadenocarcinoma) Resection if patient s condition allows surgery Side-branch intraductal papillary mucinous neoplasm Main-duct intraductal papillary mucinous neoplasm 90.5%, 86.2%, and 88% for differentiating cystic from solid pancreatic lesions. Kim and colleagues also found that the sensitivity of endoscopic ultrasound for characterization of septa (77.8%), mural nodules (58.3%), main pancreatic duct dilatation (85.7%), and main pancreatic duct communication (88.9%) was comparable to that of MRI. However, endoscopic ultrasound is invasive and operator dependent, and these limitations have led to considerable variability in determining accuracy in differentiating benign and malignant lesions [48, 52 54]. Ahmad et al. [55] found only fair agreement (κ = 0.24) between experienced endosonographers in the diagnosis of neoplastic versus nonneoplastic cystic pancreatic lesions. In addition, several investigators have noted difficulty in sampling lesions smaller than 3 cm [48, 52 54]. Currently, endoscopic ultrasound with or without aspiration is used in the following instances: indeterminate MDCT or MRCP findings; care of patients at high surgical risk owing to comorbid conditions or advanced age, which precludes them from undergoing extensive surgery; and confirmation of the malignant status of a cystic lesion before it is resected [48, 52 54]. Endoscopic ultrasound guided cyst fluid aspiration is often performed in conjunction 6 46% risk of development of high-grade dysplasia or malignancy [18] 57 92% risk of development of high-grade dysplasia or malignancy within 5 y; follow-up typically not conducted because the prevalence of carcinoma and carcinoma in situ at diagnosis is high [18] Resection if lesion is symptomatic, larger than 30 mm, or mural nodules or main duct dilatation larger than 6 mm is present; if lesion is not resectable, imaging follow-up a is recommended; yearly follow-up imaging if lesion is smaller than 10 mm, 6- to 12-mo follow-up imaging if mm, 6-mo follow-up if > 20 mm Resection if patient s condition allows surgery Solid pseudopapillary neoplasm Low malignant potential [25] Resection if patient s condition allows surgery Cystic pancreatic neuroendocrine neoplasm Variable malignant potential [25] Resection if patient s condition allows surgery a Follow-up guidelines are based on Sendai criteria [22]. with endoscopic ultrasound for definitive diagnosis (Table 3). In a multicenter trial that included 341 patients with cystic pancreatic lesions [56], endoscopic ultrasound had low sensitivity (56%) and specificity (45%) for differentiation of mucinous and nonmucinous cystic lesions on the basis of endoscopic ultrasound morphologic features. However, based on results of cytologic (fine needle aspiration [FNA]) evaluation, sensitivity, specificity, and accuracy were 34.5%, 83%, and 51%. Biochemical analysis of the cyst fluid aspirate for estimation of carcinoembryonic antigen (CEA), mucin, and amylase concentrations can facilitate reliable differentiation of mucinous and nonmucinous cystic neoplasms [18 20, 25, 56]. A cutoff CEA concentration of 192 ng/ml has been found to have 84% specificity in differentiation of mucinous from nonmucinous lesions [54, 56 59]. Cyst fluid amylase concentration also is helpful in differentiating pseudocysts from lesions that are not pseudocysts [56, 58, 59]. Although amylase concentrations less than 250 U/L are helpful for excluding pseudocysts, concentrations greater than 250 U/L are nonspecific because they occur not only in pseudocysts but also in benign IPMNs and MCNs [58]. FNA cytologic evaluation of the cyst fluid is also performed for cyst characterization, but the yield of cytologic evaluation is often limited by low cellularity of the fluid aspirate [18 20]. Evidence-Based Management Guidelines Imaging Appearance Optimal management of cystic pancreatic lesions begins with morphologic classification into one of four types: unilocular, microcystic, macrocystic, and cysts with solid components [27]. Unilocular cysts are thinwalled simple cystic lesions without internal septa, solid components, or calcifications [24, 27]. Pseudocysts are the most common lesion in this category, and usually, features of pancreatitis, such as inflammation, atrophy, and pancreatic parenchymal calcifications, are also seen [24, 27]. In rare instances, IPMNs, SCAs (< 10%), MCNs, and lymphoepithelial cysts present as unilocular cysts [27, 60]. Microcystic lesions typically present with multiple tiny cysts (more than six, each measuring < 2 cm) with lobulated outlines and thick or fleshy stroma [20, 27, 60, 61]. The microcystic appearance is typically seen in SCAs, and the pathognomic fibrous central scar is present in only 30% of cases [20, 27, 60, 61]. Microcystic lesions can have avid enhancement on arterial phase images after 348 AJR:200, February 2013

7 Cystic Pancreatic Lesions TABLE 3: Fluid Characteristics of Pancreatic Cystic Lesions Characteristic Pseudocyst Serous Cystadenoma Mucinous Cystic Neoplasm Intraductal Papillary Mucinous Neoplasm Nature Turbid, hemorrhagic Thin and clear, possibly bloody Thick and viscous Thick and viscous Possibly bloody Viscosity Low Low High High NA Mucin content Low Low High High NA < 5 ng/ml < 5 ng/ml High (> 192 ng/ml) High (> 192 ng/ml) NA Carcinoembryonic antigen concentration Amylase concentration High (> 250 U/L) Low (< 250 U/L) Variable a Variable a Low Glycogen content None Abundant None None None Note Data from [20, 24, 58]. NA = not applicable. a High concentration of amylase (> 250 U/L) can be seen in benign intraductal papillary mucinous neoplasms and benign mucinous cystic neoplasms. IV contrast injection owing to the presence of a vascular epithelial lining. This effect is especially pronounced in lesions with a very small cyst size, causing them to masquerade as solid pancreatic neoplasms such as neuroendocrine tumors and metastatic lesions from a primary cancer such as renal cell carcinoma or melanoma [60]. Delayed phase contrast-enhanced images can show the microcysts and the enhancing stroma. Similarly, T2-weighted MR images can confirm the presence of high-signal-intensity microcysts [20, 60]. Most SCAs have a microcystic appearance on images. Oligocystic and macrocystic patterns of SCA have been described in fewer than 10% of patients, and they can be difficult to differentiate from mucinous neoplasms on imaging [20, 61, 62]. Macrocystic lesions are composed of fewer cysts than are microcystic lesions, and the cysts are often larger than 2 cm in diameter [18, 19, 27]. MCNs and side-branch IPMNs are included in this category. Patient demographics (age, sex) and presence or absence of cyst communication can be used to differentiate MCNs and side-branch IPMNs [18, 19, 27]. MCNs are common among middle-aged women, are usually well defined, and are often located in the pancreatic tail [18, 19, 24, 27]. Side-branch IPMNs are commonly detected in older men and are more frequently located in the proximal pancreas (head and uncinate process) [18, 19, 24, 27]. An important differentiating feature between MCN and IPMN is visualization of pancreatic ductal communication. If a clear channel of communication with the pancreatic duct is visualized, the diagnosis of side-branch IPMN is almost certain because SCAs and MCNs do not communicate with the pancreatic ductal system [16, 26]. Cysts with solid components include true cystic tumors (MCNs, IPMNs) and solid pancreatic neoplasms associated with a cystic component, which includes tumors such as pancreatic neuroendocrine neoplasm, solid pseudopapillary neoplasm, adenocarcinoma of the pancreas, and metastatic lesions [27]. Both MDCT and MRI can depict the presence of enhancing solid components in a cystic lesion, which is diagnostic for this category of lesions. The lesions encountered in this category are either frankly malignant or have high malignant potential. Therefore, surgical resection is the preferred management [27, 63]. Management Guidelines With MDCT and MRI, a selective management approach can be considered for each patient after factors such as clinical presentation, age, sex, and surgical risk are accounted for [12, 22, 63, 64] (Figs. 5 and 6). The eventual management paradigm should weigh the risk of aggressiveness and the benefit of pancreatic resection. It should also include risk of development of advanced dysplastic or invasive changes in presumed mucinous lesions [12, 22, 63]. Surgery is often recommended for symptomatic cystic lesions, cystic lesions having complex morphologic features (e.g., solid components), and cystic lesions detected in patients younger than 50 years [12, 22, 63]. Asymptomatic SCAs larger than 4 cm often are resected because of a high likelihood of rapid growth and a propensity to development of symptoms [65, 66]. Because mucinous lesions have a higher propensity toward aggressive biologic behavior at detection and toward later transformation, knowledge of the mucinous nature of cystic lesions influences management [22, 31, 67, 68]. In selected patients, endoscopic ultrasound guided cyst aspiration and FNA can be considered if the imaging findings are indeterminate or the risk of surgery outweighs the benefits. Because MCNs are encountered in young patients and are premalignant or malignant, Solid Pseudopapillary Neoplasm they are usually surgically treated at diagnosis [22]. The imaging predictors of malignancy in MCNs include large size (> 4 cm) and the presence of mural nodules and eggshell calcification [22]. Because the natural history of MCNs can follow a stepwise progression to malignancy, these lesions typically require a more aggressive approach, even when obvious imaging evidence of malignant behavior is lacking at the initial presentation [15, 22]. The prevalence of biologic aggressiveness of cystic lesions varies from 44.6% to 60% [6, 15, 30]. Biologically aggressive cystic lesions include those with overtly malignant features and lesions with higher likelihood of becoming malignant (histopathologic finding of moderate- to high-grade dysplasia) [10, 68, 69]. The prevalence of potential malignancy is higher in mucinous than in nonmucinous lesions [15, 31, 67, 68]. Mucinous cystic lesions with low-grade dysplastic changes (adenomas) are generally considered benign, and the risk of malignant transformation is unknown. Therefore, aggressive monitoring after surgical resection is not necessary [70 73]. On the basis of involvement of the pancreatic duct, IPMNs are classified as either main duct IPMN, side-branch IPMN, or mixed variant IPMN involving both the main pancreatic duct and side branches [14, 18, 22, 27]. IPMNs have distinct histologic subtypes: gastric, intestinal, pancreatobiliary, and oncocytic [74]. Main duct IPMNs often have intestinal-type epithelium, and side-branch IPMNs usually have gastric-type epithelium [74]. Although all morphologic variants of IPMN can progress to cancer, invasive adenocarcinoma originating in gastric-type IPMNs is associated with a significantly worse survival rate than that originating from other types of IPMNs [74]. However, the imaging features are not specific for differentiating the various histologic AJR:200, February

8 Sahani et al. Unilocular Symptomatic Follow-up imaging Pancreatic cystic mass detected with MDCT or MRI 4 cm Consider surgical resection depending on comorbidities and risk Microcystic < 4 cm Possible MCN Benign Macrocystic or cyst with septa (indeterminate) See Figure 6 Possible IPMN < 3 cm Indeterminate 3 cm Cyst with solid components or with suspicious features Indeterminate EUS with or without aspiration Suspicious features Consider surgical resection depending on comorbidities and risk Fig. 5 Flowchart shows management guidelines for pancreatic cystic lesions seen on imaging. Suspicious features include presence of mural nodules, main duct dilatation, solid component, symptoms, and thick wall or septations. Differentiation of possible intraductal papillary mucinous neoplasm (IPMN) and mucinous cystic neoplasm (MCN) is based on classic imaging features. EUS = endoscopic ultrasound. variants of IPMNs. IPMNs can be managed either surgically or conservatively, depending on their characteristics, the clinical presentation, and the patient s age. Because of the higher risk of invasive cancer associated with main duct lesions, resection is recommended for all main duct and mixed variant IPMNs [14, 22, 75]. In patients whose condition is considered acceptable for surgery, the type of surgical resection is influenced by lesion location and the extent of tumor foci in the duct epithelium. For side-branch IPMNs, a surgical approach is undertaken if the patient has symptoms (pain, nausea, diarrhea, weight loss, jaundice), if there is main duct involvement or dilatation (> 6-mm), if the cyst is larger than 3 cm, or complex features such as a thick irregular wall, thick septa, and solid nodules are identified on imaging [22]. Because side-branch IPMNs without complex morphologic features usually have low malignant potential, surgical management is not always warranted [4, 22]. Although the incidence of potential malignancy is lower for smaller lesions (< 3 cm), the presence of suspicious features on images, even in a small cystic lesion, should be approached more aggressively [8, 16]. Therefore, despite the low incidence of aggressiveness of mucinous cystic lesions 3 cm and smaller, the incidence is not low enough to dismiss the lesions entirely, and careful review of the imaging features is mandated. In addition, patients whose condition is found not suitable for surgical management often need frequent assessments for growth and change in imaging features [14, 15, 22]. A panel of experts have proposed the Sendai criteria as guidelines for the management of side-branch IPMNs [22]. The follow-up guideline varies in accordance with the size of side-branch IPMN [22]. Lesions smaller than 1 cm are evaluated annually; those measuring 1 2 cm are evaluated every 6 12 months; and those measuring 2 3 cm are imaged at intervals of 3 6 months [22]. However, authors of more recent studies have recommended considering a longer surveillance interval of 2 years for cystic lesions smaller than 3 cm after baseline detection in the absence of mural nodules [63, 76]. Accordingly, it would be prudent to perform followup evaluations every 2 years for side-branch IPMNs smaller than 2 cm and to perform annual evaluations for IPMNs measuring 2 3 cm [63, 76]. The choice of imaging modality for monitoring IPMNs depends on institutional preference and the patient s age. Although MDCT and MRI are both accepted methods for follow-up of these lesions, for adults younger than 50 years, MRI can be considered owing to concerns about radiation exposure from MDCT. Regardless of the type of imaging modality used, contrast-enhanced examinations are crucial for improving detection of enhancing solid components, the cyst wall, and septa. Contrast injection is desirable, but for patients with compromised renal function and those with lower cancer risk (small lesion, advanced age), follow-up CT or MRI can be performed without contrast injection. However, if suspicious features are observed during follow-up examinations, IV contrast medium should be used [77]. A vexing issue in the follow-up of pancreatic cystic lesions is the total duration of follow-up. It would be reasonable to increase the follow-up intervals to 2 years for lesions 2 cm and larger that are stable for 2 years. For lesions smaller than 2 cm, imaging follow-up can stop after stability has been found for 2 years. However, due consideration needs to be given to the patient s age, symptoms, and capability of undergoing surgical resection. At follow-up imaging, lesions that are indeterminate or have a growth spurt of more than 1 cm/year, endoscopic ultrasound can be performed to confirm the malignant nature (Fig. 6). In patients in whom endoscopic ultrasound findings indicate a cystic lesion is benign, follow-up is performed in accordance with the size of the cystic lesion. Surgical Management A variety of open and laparoscopic surgical options are available for patients whose condition allows surgery. For lesions in the head of the pancreas, such as an IPMN with one or more of the aforementioned suspicious features, either a standard Whipple procedure or pylorus-sparing pancreaticoduodenectomy can be performed [5, 39, 78]. Medial segmental pancreatectomy is performed for a lesion in the neck or the body of the pancreas. For lesions involving the tail of the pancreas, the spleen is assessed for involvement because it does not necessarily have to be removed with the distal pancreas [5, 39, 78]. Postsurgical Follow-Up The postsurgical follow-up of patients who have undergone resection of cystic pancreatic neoplasms depends on the histologic features. Benign MCNs do not recur and therefore require no postoperative follow-up [22]. Because the risk of local recurrence and distant metastasis is higher for malignant MCNs, postsurgical follow-up evaluations are needed every 6 months [22]. The postsurgical surviv- 350 AJR:200, February 2013

9 Cystic Pancreatic Lesions al rate for invasive IPMN varies between 35% and 60%, mortality being associated with cancer recurrence, most commonly local or extrapancreatic metastasis [79 81]. The risk of recurrence ranges from 3% to 11% [75, 79]. In cases of local recurrence, completion pancreatectomy may be necessary. It is important to emphasize that invasive IPMN has a better survival rate than pancreatic ductal adenocarcinoma [74]. Follow-up of a side-branch IPMN should be pursued carefully, and the time frame of follow-up should be based on the patient s risk and the lesion size. Guidelines laid down by the International Association of Pancreatology call for yearly follow-up evaluations of benign IPMNs and for imaging follow-up in conjunction with measurement of serum markers (CEA and CA19-9) after resection of invasive IPMN [7, 22]. The followup imaging protocol used in these scenarios should consider the aggressiveness of the resected lesion and the surgical margins. Most recurrences take place within 3 years. There have been reports [69, 82], however, of recurrence after 5 years. If the surgical margins are negative, benign lesions can be evaluated at intervals of 1 year. Patients with borderline lesions or carcinoma in situ, a positive margin, or indeterminate cystic lesions in the pancreatic remnant need more frequent evaluations for the first 2 3 years [83]. An added objective of follow-up imaging is to detect invasive cancer, Main duct IPMN Combined IPMN Symptomatic side-branch IPMN Risk and benefits of surgery Low risk Resection IPMN detected with MDCT or MRI High risk > 3 cm Positive cytologic result because patients with IPMN are at increased risk of development of synchronous or metasynchronous invasive ductal adenocarcinoma [14] (Fig. 3). The Gastroenterologist s Perspective The detection of a potentially malignant lesion in the pancreas causes considerable anxiety to both patient and physician. The prevalence of incidentally identified cystic lesions of the pancreas is high, but it is increasingly becoming apparent that only a small minority of such lesions progress to cancer. Imaging, especially MRI, MDCT, and endoscopic ultrasound, plays an important role in risk stratification, avoiding unnecessary surgery, and safe follow-up of lesions that are not resected. Most incidentally identified cystic lesions can be safely followed up. Current international guidelines help in this regard. They are highly sensitive to identification of high-risk lesions (in situ and invasive cancer) but need further refinement to improve their positive predictive value for high-grade pathologic findings. The Pancreatic Surgeon s Perspective Cystic neoplasms of the pancreas are one of the most common indications for pancreatic surgery. Because most of the cystic lesions that we remove do not contain invasive cancer and are asymptomatic, we can state EUS 2 3 cm Negative cytologic result Size < 1 cm follow-up yearly Asymptomatic side-branch IPMN < 2 cm Follow-up with imaging Size 1 2 cm follow-up every 6 12 months Size 2 3 cm follow-up every 3 6 months Growth > 1 cm/y, increase in size to > 3 cm, suspicious features Fig. 6 Flowchart shows management guidelines for intraductal papillary mucinous neoplasm (IPMN). Highrisk factors for surgery are old age and presence of comorbid conditions. Low-risk factors for surgery are young age and no comorbid conditions. Suspicious features are mural nodules, main duct dilatation, solid component, symptoms, and thick wall or septations. Follow-up guidelines are based on Sendai criteria [22]. EUS = endoscopic ultrasound. that many of these operations are preventive. Most of these lesions are IPMNs that contain only low-, moderate-, or high-grade dysplasia (what we used to refer to as in situ carcinoma), and we remove them either because we cannot reliably exclude invasive cancer or because we believe that progression will inevitably occur and the lesion will become invasive, akin to the process that occurs in a colonic polyp. The decision to operate, however, is not straightforward. Although pancreatic surgery has become safer and the risk of dying after a Whipple procedure or distal pancreatectomy is less than 2% at most major medical centers, the frequency of complications is still high (> 40%), and the consequences of endocrine and exocrine insufficiency with loss of pancreatic tissue are not trivial. These risks have to be carefully weighed against the potential benefit. Striking the right balance can be difficult because most of these lesions occur in elderly persons, and our knowledge of the natural history of IPMNs is incomplete. Practice Recommendations Annual imaging surveillance is generally sufficient for benign serous cystadenomas smaller than 4 cm and for asymptomatic lesions. Asymptomatic thin-walled unilocular cystic lesions smaller than 3 cm or side-branch IPMNs should be followed up with CT or MRI at 6 and 12 months interval after detection and then annually for 3 years. Cystic lesions with more complex features or with growth rates greater than 1 cm/year should be followed more closely or recommended for resection if the patient s condition allows surgery. Symptomatic cystic lesions, neoplasms with high malignant potential, and lesions larger than 3 cm should be referred for surgical evaluation. Endoscopic ultrasound with FNA biopsy can be used preoperatively to assess the risk of malignancy. Recommendations for Further Research Despite great strides in noninvasive imaging and endoscopic ultrasound in the characterization of pancreatic cystic lesions, current imaging techniques are not accurate in the differentiation of cystic lesions associated with carcinoma in situ or high-grade dysplasia from benign lesions. Though MDCT and MRI can reliably depict cystic lesions with obvious aggressive biologic features, their value for prediction of the biologic behavior of all the cysts is limited. Advanced techniques such as PET/MRI with targeted radioisotopes have AJR:200, February

10 Sahani et al. potential for characterization of changes at the molecular level. Endoscopic ultrasound guided FNA and fluid analysis already have a specific role in surgical decision making [84]. However, though the commonly used cyst fluid CEA levels are useful in differentiating mucinous from nonmucinous cystic lesions of the pancreas, they have no value for assessing the degree of dysplasia in mucinous lesions [59, 84]. Pancreatic cyst fluid DNA analysis that includes identification of KRAS2 gene mutations is being explored for predicting the aggressiveness of pancreatic cysts [84]. Few initial studies have shown promise with this technique over FNA and other types of cyst fluid analysis in predicting the biologic features of cysts [84]. Further studies are necessary to validate reliable molecular and imaging markers for accurate prediction of the biologic behavior of cysts and to guide the management of pancreatic cystic lesions. References 1. Laffan TA, Horton KM, Klein AP, et al. Prevalence of unsuspected pancreatic cysts on MDCT. AJR 2008; 191: Lee HJ, Kim MJ, Choi JY, Hong HS, Kim KA. Relative accuracy of CT and MRI in the differentiation of benign from malignant pancreatic cystic lesions. Clin Radiol 2011; 66: Zhang XM, Mitchell DG, Dohke M, Holland GA, Parker L. Pancreatic cysts: depiction on singleshot fast spin-echo MR images. Radiology 2002; 223: Correa-Gallego C, Ferrone CR, Thayer SP, Wargo JA, Warshaw AL, Fernandez-Del Castillo C. Incidental pancreatic cysts: do we really know what we are watching? Pancreatology 2010; 10: Ferrone CR, Correa-Gallego C, Warshaw AL, et al. Current trends in pancreatic cystic neoplasms. Arch Surg 2009; 144: Goh BK, Tan YM, Thng CH, et al. How useful are clinical, biochemical, and cross-sectional imaging features in predicting potentially malignant or malignant cystic lesions of the pancreas? Results from a single institution experience with 220 surgically treated patients. J Am Coll Surg 2008; 206: Lévy P, Jouannaud V, O Toole D, et al. Natural history of intraductal papillary mucinous tumors of the pancreas: actuarial risk of malignancy. Clin Gastroenterol Hepatol 2006; 4: Handrich SJ, Hough DM, Fletcher JG, Sarr MG. The natural history of the incidentally discovered small simple pancreatic cyst: long-term follow-up and clinical implications. AJR 2005; 184: Kimura W, Nagai H, Kuroda A, Muto T, Esaki Y. Analysis of small cystic lesions of the pancreas. Int J Pancreatol 1995; 18: Kirkpatrick ID, Desser TS, Nino-Murcia M, Jeffrey RB. Small cystic lesions of the pancreas: clinical significance and findings at follow-up. Abdom Imaging 2007; 32: Lahav M, Maor Y, Avidan B, Novis B, Bar-Meir S. Nonsurgical management of asymptomatic incidental pancreatic cysts. Clin Gastroenterol Hepatol 2007; 5: Spinelli KS, Fromwiller TE, Daniel RA, et al. Cystic pancreatic neoplasms: observe or operate. Ann Surg 2004; 239: ; discussion, Ohtsuka T, Kono H, Nagayoshi Y, et al. An increase in the number of predictive factors augments the likelihood of malignancy in branch duct intraductal papillary mucinous neoplasm of the pancreas. Surgery 2012; 151: Sahani DV, Lin DJ, Venkatesan AM, et al. Multidisciplinary approach to diagnosis and management of intraductal papillary mucinous neoplasms of the pancreas. Clin Gastroenterol Hepatol 2009; 7: Sahani DV, Sainani NI, Blake MA, Crippa S, Mino-Kenudson M, del-castillo CF. Prospective evaluation of reader performance on MDCT in characterization of cystic pancreatic lesions and prediction of cyst biologic aggressiveness. AJR 2011; 197:72; [web]w53 W Sainani NI, Saokar A, Deshpande V, Fernandezdel Castillo C, Hahn P, Sahani DV. Comparative performance of MDCT and MRI with MR cholangiopancreatography in characterizing small pancreatic cysts. AJR 2009; 193: Curry CA, Eng J, Horton KM, et al. CT of primary cystic pancreatic neoplasms: can CT be used for patient triage and treatment? AJR 2000; 175: Sakorafas GH, Smyrniotis V, Reid-Lombardo KM, Sarr MG. Primary pancreatic cystic neoplasms revisited. Part III. Intraductal papillary mucinous neoplasms. Surg Oncol 2011; 20:e109 e Sakorafas GH, Smyrniotis V, Reid-Lombardo KM, Sarr MG. Primary pancreatic cystic neoplasms revisited. Part II. Mucinous cystic neoplasms. Surg Oncol 2011; 20:e93 e Sakorafas GH, Smyrniotis V, Reid-Lombardo KM, Sarr MG. Primary pancreatic cystic neoplasms revisited. Part I. Serous cystic neoplasms. Surg Oncol 2011; 20:e84 e Sadakari Y, Ienaga J, Kobayashi K, et al. Cyst size indicates malignant transformation in branch duct intraductal papillary mucinous neoplasm of the pancreas without mural nodules. Pancreas 2010; 39: Tanaka M, Chari S, Adsay V, et al. International consensus guidelines for management of intraductal papillary mucinous neoplasms and mucinous cystic neoplasms of the pancreas. Pancreatology 2006; 6: Yamaguchi K, Kanemitsu S, Hatori T, et al. Pancreatic ductal adenocarcinoma derived from IPMN and pancreatic ductal adenocarcinoma concomitant with IPMN. Pancreas 2011; 40: Brugge WR, Lauwers GY, Sahani D, Fernandezdel Castillo C, Warshaw AL. Cystic neoplasms of the pancreas. N Engl J Med 2004; 351: Sakorafas GH, Smyrniotis V, Reid-Lombardo KM, Sarr MG. Primary pancreatic cystic neoplasms of the pancreas revisited. Part IV. Rare cystic neoplasms. Surg Oncol 2012; 21: Sahani DV, Kadavigere R, Blake M, Fernandez-Del Castillo C, Lauwers GY, Hahn PF. Intraductal papillary mucinous neoplasm of pancreas: multi-detector row CT with 2D curved reformations correlation with MRCP. Radiology 2006; 238: 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: Kawamoto S, Lawler LP, Horton KM, Eng J, Hruban RH, Fishman EK. MDCT of intraductal papillary mucinous neoplasm of the pancreas: evaluation of features predictive of invasive carcinoma. AJR 2006; 186: Taouli B, Vilgrain V, Vullierme MP, et al. Intraductal papillary mucinous tumors of the pancreas: helical CT with histopathologic correlation. Radiology 2000; 217: Visser BC, Yeh BM, Qayyum A, Way LW, Mc- Culloch CE, Coakley FV. Characterization of cystic pancreatic masses: relative accuracy of CT and MRI. AJR 2007; 189: Procacci C, Carbognin G, Accordini S, et al. CT features of malignant mucinous cystic tumors of the pancreas. Eur Radiol 2001; 11: Chaudhari VV, Raman SS, Vuong NL, et al. Pancreatic cystic lesions: discrimination accuracy based on clinical data and high resolution CT features. J Comput Assist Tomogr 2007; 31: Chaudhari VV, Raman SS, Vuong NL, et al. Pancreatic cystic lesions: discrimination accuracy based on clinical data and high-resolution computed tomographic features. J Comput Assist Tomogr 2008; 32: Kim SH, Lim JH, Lee WJ, Lim HK. Macrocystic pancreatic lesions: differentiation of benign from premalignant and malignant cysts by CT. Eur J Radiol 2009; 71: Tomimaru Y, Takeda Y, Tatsumi M, et al. Utility of 2-[ 18 F] fluoro-2-deoxy-d-glucose positron emission tomography in differential diagnosis of benign and malignant intraductal papillary-mucinous neoplasm of the pancreas. Oncol Rep 2010; 24: Fukukura Y, Fujiyoshi F, Sasaki M, Nakajo M. Pancreatic duct: morphologic evaluation with MR cholangiopancreatography after secretin stimula- 352 AJR:200, February 2013