Pseudocysts are sequelae of acute interstitial pancreatitis and require at least 4 weeks to form. A thin capsule of nonepithelialized granulation or fibrotic tissue forms a wall around amylase-rich fluid. Symptoms, when present, typically consist of abdominal pain and early satiety. Gastric outlet and/or biliary obstruction may occur as well. Usually pseudocysts are readily diagnosed by the patient’s history of acute or chronic pancreatitis . Without a clear history of acute or chronic pancreatitis , differentiating pseudocysts from MCN and even SCA and BD-IPMN may be difficult by imaging alone. On abdominal CT, pseudocysts typically appear round with a thin or thick wall. Calcifications and communication with the pancreatic duct may be present.

Serous cystadenomas are benign pancreatic cystic neoplasms, which very rarely become malignant with a couple of larger studies describing about a 1 % rate of malignancy [6, 7]. SCAs account for over 30 % of pancreatic cystic neoplasms, typically occur in women over the age of 60, and are defined on pathology by glycogen containing cuboidal epithelial cells (Fig. 31.1) [8]. They arise anywhere throughout the pancreas. When symptomatic, SCAs usually present with nonspecific symptoms, mainly abdominal pain, due to compression of adjacent organs by the cyst. Symptoms occur more commonly in larger cysts > 4 cm (77 %) compared to cysts < 4 cm (22 %).

Radiologically, SCAs may be classified into several subtypes. Over half SCAs are microcystic defined as each cyst compartment less than 1 cm while the rest are macrocystic (composed of cysts larger than 1 cm), mixed, or rarely solid (Fig. 31.2) [6]. A Japanese study of 172 SCA diagnosed by resection and typical imaging findings noted highest diagnostic accuracy for microcystic SCA (85 %) with significantly lower diagnostic rates (17–50 %) for macrocystic, mixed, and solid types. CT is only approximately 23 % accurate for SCA while diffusion-weighted MRI has demonstrated 100 % sensitivity and 97 % specificity for differentiating mucinous cysts from SCA [9, 10]. The pathognomonic central scar or “sunburst calcification” is present in only about 30 % of these cysts (Fig. 31.3) [11].

The natural history of SCAs is not well described; however, they appear to grow over time. Although an older study reported more rapid growth rate in cysts > 4 cm, a recent multicenter study failed to confirm these results [12, 13]. A rate of growth of 6.2 % per year or a doubling time of 12 years was calculated for the nonresected SCAs while resected SCAs grew faster (17 % per year for a doubling time of 4.5 years) [12].

Mucinous cystic neoplasms are premalignant parenchymal lesions defined by ovarian-like stroma on pathology (Fig. 31.4) and thus almost exclusively occur in women. They arise in the body and tail of the pancreas in approximately 95 % of patients. Unlike SCAs, presence of symptoms in mucinous cystic lesions is associated with malignancy . Other features concerning for malignancy in MCN include older age, large size especially > 6 cm, and presence of thick cyst wall, mural nodules, or peripheral eggshell calcification [14, 15]. The true incidence of malignancy in MCNs is unknown although recent studies suggest lower rates of invasive cancer (6–27 %) with only 5.5 % carcinoma in situ [15, 16]. Unlike SCA, MCNs usually appear smooth, well-defined, and unilocular or with a few septations (Fig. 31.5).

IPMNs are also mucinous cysts and arise from the pancreatic ductal epithelium of the main duct, side branches, or both. They occur more commonly in men between ages 50 and 60. While IPMNs usually arise in the head of the pancreas, they can occur anywhere in the pancreas as well as in multiple locations. There are three radiologic subtypes of IPMN: main duct (diffuse or segmental dilation of the main duct > 5 mm), branch duct (dilation of one or more side branches), and mixed-type (both main duct and side branch involvement; Figs. 31.6, 31.7, and 31.8). Typically one relies on MRI/MRCP to distinguish BD-IPMN from mixed-type and MD-IPMN. Rarely can MD-IPMN be diagnosed endoscopically by visualizing the “fish-mouth” papilla, which represents mucus emerging from a widely patulous papilla (Fig. 31.9). Unfortunately, radiology misclassifies 29 % of MD-IPMN and 21 % of BD-IPMN [17]. Up to 29 % of mixed-type IPMN are misdiagnosed as BD-IPMN [18]. By pathology, IPMN may also be classified as gastric, intestinal, oncocytic, or pancreaticobiliary type. There is increasing appreciation of the histologic subtypes and their relationship with malignancy. In 169 resected IPMNs, 73 % were gastric, 25 % intestinal, 2 % pancreaticobiliary, and 0.6 % oncocytic [18]. HGD and cancer occurred more frequently in intestinal (32/42) than gastric type (48/123, p< 0.0001). A recent retrospective study of surgically resected IPMNs did note that gastric IPMNs were more likely to be smaller without nodules or masses and higher CEA level [19]. However, the clinical utility of this histological grading remains uncertain as definitive grading is currently available only following surgical resection.

Correctly identifying the radiologic subtypes of IPMN is important because of differences in malignant potential and, therefore, management. MD-IPMN has the greatest malignant potential ranging from 40 to 70 % with nearly comparable risk in mixed-type IPMN. Symptoms do not appear to increase risk of malignancy in MD-IPMN although the passage of time does (1 year following first imaging or symptom onset, 20–42 % malignant; 2 years, 40–54 %; 5 years, 40–66 %) [20, 21]. Not all MD-IPMN and mixed-type IPMN behave the same as certain patients appear to have a more indolent course. MD-IPMN with normal serum CA 19-9 and cyst fluid cytology harbor lower rates of malignancy compared to other MD-IPMN (29 versus 60 %,p< 0.0001) [22]. Similarly mixed-type IPMN with noncircumferential involvement of the main duct in one or a few histologic sections and no gross abnormalities other than a dilated pancreatic duct had lower malignancy than mixed-type IPMN not meeting these criteria (17 versus 70%, p< 0.0001) [23].

Predictors of malignancy in BD-IPMN include those defined in the 2012 International Association of Pancreatology (IAP) guidelines (presence of a mass, mural nodules, dilated main pancreatic duct ≥ 1 cm, obstructive jaundice with cyst in the head of the pancreas, cytology suspicious or positive for malignancy), cyst size greater than 3 cm, and symptoms [22–28]. Symptoms attributable to IPMN include steatorrhea and diabetes with 15–30 % of IPMNs presenting with acute pancreatitis, which is believed due to obstruction from mucus plugging the ducts. In a small study of BD-IPMN, no asymptomatic patients had cancer while symptomatic patients developed more cancers over time (1 year, 15 %; 2 years, 30 %; 5 years, 37 %) [21].

Regarding the IAP guidelines, the absence of malignant features predicts benign cysts well with nearly 100 % negative predictive value while presence of them less accurately diagnoses malignancy [24]. The 2012 IAP guidelines were revised from the 2006 guidelines, and the former appears to predict malignancy better [29]. A major difference between the 2006 and 2012 guidelines is the removal of cyst size as a definite criterion for resection. Several studies contradict this de-emphasis of size. A metaanalysis of pathologically confirmed IPMNs and predictors of malignancy from the 2006 and 2012 guidelines identified cyst size > 3 cm as the strongest predictor of malignancy with odds ratio (OR) 62 followed by nodule (OR 9.3), pancreatic duct > 6 mm (OR 7.3), MD-IPMN (OR 4.7), and symptoms (OR 1.6) [30, 31].

Less common pancreatic cystic neoplasms include solid pseudopapillary neoplasm (SPEN), which occurs almost exclusively in young women. SPENs account for 1–2 % of pancreatic cystic neoplasms. They were first described in 1959 as Frantz or Hamoudi tumors and then renamed SPEN by the World Health Organization in 1996. SPENs are premalignant with reported 2–15 % local invasion or metastatic disease [32]. About 10–15 % of SPENs are malignant, and to date, no predictors of aggressive behavior have been identified [33]. These patients usually present with nonspecific abdominal pain and occasionally with an abdominal mass palpable on examination. SPENs may occur anywhere throughout the entire pancreas. Pathology reveals characteristic pseudopapillae with cystic spaces containing hemorrhage and cholesterol clefts in myxoid stroma alternating with solid tissue.

Neuroendocrine or acinar cell tumors can occasionally undergo cystic degeneration. Cystic neuroendocrine tumors account for only 8–17 % of pancreatic neuroendocrine tumors and are usually nonfunctional [34]. Acinar cystadenocarcinoma is extremely rare with fewer than 10 cases reported in literature and typically presents with abdominal pain and a multilocular cystic lesion [35]. Unlike most other pancreatic cystic lesions, SPEN and cystic neuroendocrine tumors usually have characteristic findings on imaging. SPEN typically presents as a large well-defined encapsulated mass with peripheral solid component and cystic degeneration in the center with areas of hemorrhage (Fig. 31.10) [36]. Peripheral calcification occurs rarely. Cystic neuroendocrine tumors are highly vascularized with early enhancement of the rim during early arterial imaging with MRI (Fig. 31.11) [34].

Pancreatic lymphangiomas are endothelium-lined cysts arising from the lymphatic system due to blocked lymphatics from inflammation or congenital anomaly [37]. During embryogenesis, ectopic lymphatic tissue lands in the pancreas and these cysts form from progressive dilation of insufficiently draining lymphatic vessels. Most pancreatic lymphangiomas occur incidentally in women in the body and tail of the pancreas. Complications include abdominal pain, hemorrhage , infection, and hydronephrosis. Lymphangiomas are difficult to distinguish from pancreatic cystic neoplasms on imaging and typically appear multiseptated, well-defined.

Lymphoepithelial cysts are another group of rare nonneoplastic pancreatic cysts accounting for 0.5 % of pancreatic cysts [38]. These typically occur in middle-aged men in the body or tail of the pancreas. The cysts are lined by stratified squamous epithelium with subepithelial lymphoid tissue and follicles. The equally rare simple or true cyst is lined by cuboidal epithelial cells and does not communicate with the pancreatic duct. They occur in about 10 % of patients with autosomal-dominant polycystic kidney disease. In addition, they are the most common pancreatic lesion seen in von Hippel-Lindau disease (up to 72 % of patients) [39].

Retention cysts are actually cystically dilated segments of pancreatic duct resulting from obstruction [38]. The obstruction may result from stones or stricture from chronic pancreatitis or cancer. Viscous mucus in cystic fibrosis may also clog the pancreatic duct.

Recent consensus by radiologists recommended MRI as the preferred imaging modality to characterize pancreatic cysts with its enhanced ability to detect septa, nodules, ductal communication, main duct involvement, and small branch duct cysts compared to CT [40–42]. MRI images should be obtained at 1.5 or 3 T with T1, T2, and 3-D, fat-saturated, gradient-echo T1 gadolinium-enhanced sequences in pancreatic, portal, and equilibrium phases with magnetic resonance cholangiopancreatography (MRCP). MRI also helps distinguish pseudocysts from cystic neoplasms by identifying internal debris within pseudocysts [43]. MRI is comparable to multidetector (MD) CT for diagnosing the specific type of cyst (40–70 % accuracy) and may be superior in identifying mucinous cysts (79–82 % accuracy) [44–46]. Both CT and MRI predict the presence of malignancy in pancreatic cysts with high accuracy (73−79 %). These accuracy rates are comparable to EUS imaging [47]. MRI and EUS have modest sensitivity (58−67 %) for detecting mural nodules, the presence of which are concerning for malignancy.

If MRI cannot be performed due to contraindications or patient intolerance, patients should undergo a “pancreatic protocol” abdominal CT scan. The MDCT should be dual-phase contrast-enhanced with images acquired during the pancreatic and portal venous phases which can be analyzed in 3-D.

Whether EUS and EUS-FNA adds useful information beyond radiology was examined recently. In 154 patients with resected cysts, all underwent EUS, 90 % had CT, and 34 % had MRI [48]. This study focused on the ability to differentiate neoplastic (MCN, IPMN, SPEN, cystic ductal adenocarcinoma , and cystic neuroendocrine tumor) from nonneoplastic (pseudocyst, simple cyst, benign epithelial cyst, duplication cyst , and SCA) cysts. Sensitivity of EUS with or without cytology , CEA, and amylase was superior to CT and MRI (76 versus 48 and 34 %, respectively, p< 0.0001). Although this study supports the value of EUS in identifying neoplastic cysts, the low diagnostic accuracy of CT and MRI contradicts other studies. In addition, this study only applies to surgically resected cysts; therefore, it may bias in favor of EUS or EUS-FNA compared to establishing the value of EUS or EUS-FNA among all cysts including the many patients who do not undergo surgery.

Despite this study, not everyone with an incidental pancreatic cyst needs EUS or EUS-FNA. The 2012 IAP guidelines for suspected MCN and IPMN suggest EUS for patients with pancreatitis or the following “worrisome features” on imaging: cyst size ≥ 3 cm, thick enhancing cyst wall, nonenhancing nodule, main pancreatic duct (MPD) 5–9 mm, abrupt change in MPD caliber with distal pancreatic atrophy, or lymphadenopathy [22]. Similarly, the American Gastroenterological Association (AGA) guidelines recommend EUS-FNA only for higher risk cysts with at least 2 high risk features (size ≥3 cm, dilated main pancreatic duct, solid component) or significant change during surveillance (e.g., increase in pancreatic duct diameter, development of solid component) [87]. Additionally, cysts between 1 and 3 cm could undergo EUS-FNA even without worrisome features to help classify them as mucinous or nonmucinous.

EUS imaging alone is insufficient for diagnosing mucinous cysts with 56 % sensitivity, 45 % specificity, and 51 % accuracy reported in Brugge’s seminal paper [49]. In addition, interobserver agreement among expert endosonographers for distinguishing mucinous and nonmucinous cysts by EUS imaging is only fair from an older study while a more recent study found moderate agreement for the actual diagnosis of the cyst [50, 51]. In this Dutch study, agreement among expert endosonographers (performed > 1000 pancreas EUS) was better than semi-experts (performed 50–200 pancreas EUS) with experts demonstrating good agreement for nodules, moderate agreement for solid component and communication of the cyst with pancreatic duct, and fair agreement for suspicion of malignancy [50].

During EUS, features to evaluate include cyst size, presence of septations, lobular versus smooth contour, thick cyst wall, solid component to the cyst, nodule within the cyst, evidence of communication between the cyst and pancreatic duct, and size of main pancreatic duct. The following features are predictive of mucus: lesion hypoechoic relative to adjacent tissue, smooth-edge with hyperechoic rim. Nodules are iso- or hyperechoic compared to adjacent tissue without a hyperechoic rim or smooth edge [52]. Rotating the patient during EUS and trying to move the lesion with the FNA needle can help differentiate mucus from a nodule. Diagnostic accuracy of EUS for a nodule is modest (57 %) in a pathology-based study of MCN and BD-IPMN. However, after training endosonographers in the above EUS criteria for differentiating a nodule from mucus, accuracy improved to 79 %. Sensitivity and specificity of EUS (75 and 83 %) were superior to CT (24 and 100 %) for nodules [52].