Endoscopic Ultrasonography in the Evaluation of Pancreatic Cysts

Key Points

•

The differential diagnosis of pancreatic cystic lesions is wide: the majority of these lesions are benign, but detection of mucinous neoplasms (IPMN and MCN) is important because these cysts may be malignant or have malignant potential.
•

The diagnostic accuracy of EUS based on morphology alone is limited.
•

A combination of EUS features, fluid cytology, carcinoembryonic level, mucin staining, and molecular markers is used to differentiate pancreatic cysts.
•

FNA of cystic lesions under antibiotic cover is safe, with low rates of bleeding, infection, and pancreatitis.
•

Accurate diagnosis and management of pancreatic cystic lesions require careful evaluation of the clinical setting, other imaging modalities, and multidisciplinary collaboration.

Introduction

Pancreatic cysts, once thought to be rare, are now detected more frequently as a result of the increased use of high-resolution imaging. Between 2% and 13% of patients undergoing computed tomography (CT) or magnetic resonance imaging (MRI) with no symptoms or history of pancreatic disease are found to have pancreatic cysts. These lesions represent a broad spectrum of pathologic changes from benign cysts to premaligant and malignant cysts. Pancreatic cysts thus represent an important and increasing disease burden and pose a difficult diagnostic and management problem: that is, how to accurately predict which lesions contain malignancy and require resection versus those that can be followed safely by interval imaging or require no further follow-up.

Despite advances in CT and MRI, the ability of cross-sectional modalities to identify the exact nature of a cyst remains limited. Endoscopic ultrasonography (EUS) is ideally suited for the imaging of pancreatic lesions because of its high resolution and ability to sample cystic lesions. This chapter discusses the different types of pancreatic cysts, their endosonographic features, and the role of fine-needle aspiration (FNA) for cytologic and tumor marker analysis. A diagnostic approach to patients with pancreatic cysts is also described.

Types of Pancreatic Cysts

There are many different types of pancreatic cysts, including cysts with no or very low malignant potential, those that have the ability to develop high-grade dysplasia (HGD) or invasive carcinoma (IC), and cysts that harbor IC ( Table 15.1 ). Pseudocysts are common in general practice but account for less than 10% of resected pancreatic cysts. In modern surgical series, the most commonly resected pancreatic cysts are intraductal papillary mucinous neoplasms (IPMNs), mucinous cystic neoplasms (MCNs), and serous cystadenomas (SCAs), which account for 50%, 16%, and 12% of resected cysts, respectively. Solid neoplasms can also undergo cystic degeneration, with cystic pancreatic neuroendocrine tumors (PanNETs), solid pseudopapillary neoplasms (SPNs), and cystic pancreatic ductal adenocarcinomas accounting for between 1% and 9% of resected cysts. The clinical presentation, endoscopic features, and management of these cysts are discussed below.

TABLE 15.1

Classification of Pancreatic Cysts

No or Very Low Malignant Potential	Malignant Potential	Malignant
Pseudocyst	IPMN	Pancreatic ductal adenocarcinoma
Serous cystadenoma	MCN	Neuroendocrine tumor
Lymphoepithelial cyst		Solid pseudopapillary neoplasm
Retention cyst		Pancreatoblastoma
Congenital cyst		Acinar cell cystadenocarcinoma
Lymphangioma

IPMN, Intraductal papillary mucinous neoplasm; MCN, mucinous cystic neoplasm.

Diagnostic Approach

Clinical History and Imaging

Taking a good clinical history is important. Key questions are whether there is a history of pancreatitis, jaundice, or recent onset of diabetes; is there any pancreatic type of abdominal or back pain, anorexia, or weight loss? The presence of any of these features is worrisome for the presence of HGD/IC, and these patients should undergo a very careful EUS and be evaluated by a multidisciplinary team. Is there a personal or family history of related cancers to suggest multiple endocrine neoplasia (MEN) 1, von Hippel–Lindau (VHL) syndrome, or any history to suggest an increased risk of pancreatic adenocarcinoma, hereditary nonpolyposis colorectal cancer (HNPCC), Peutz-Jeghers syndrome, BRCA1 / BRCA2 mutation, or familial atypical multiple mole melanoma (FAMMM)?

Most patients have already undergone cross-sectional imaging before they are referred for EUS, but if not, an MRI or pancreatic protocol CT scan is helpful. From this scan and the clinical features, a diagnosis may be apparent. It may also be clear that the patient requires surgical resection. EUS is indicated when the diagnosis is unclear or where the patient has worrisome symptoms. A pragmatic algorithm for the differential diagnosis and management of pancreatic cysts is shown in Fig. 15.1 .

Endoscopic Ultrasound

The EUS approach to examining the pancreas is described in detail in Chapter 12 , and FNA techniques are described in Chapter 20 . The general EUS approach to pancreatic cysts is described in this section; the appearances of specific cysts are described later.

When a cystic lesion has been identified, it is important to note the number of cysts; their exact location and size; and whether the cyst is within, adjacent to, or outside of the pancreas. This information may influence management ( Table 15.2 ). If the lesion is clearly a pseudocyst, it is assessed for the need and suitability for EUS-guided drainage. The cyst itself should be examined to determine the wall thickness and the presence of a mural nodule or associated mass (see “ Examination Checklist ”). The size of the individual cysts (microcystic, macrocystic [>1 cm], or a mixture of both), the presence and thickness of any septations ( Fig. 15.2 ), and the presence of echo-dense mucus or debris within the cyst should be documented. The size of the main pancreatic duct (MPD) in the head, body, and tail, whether it communicates with the cyst, the presence of mucin or a mural nodule within the pancreatic duct, or any focal dilation should be noted.

TABLE 15.2

Features of Cystic Pancreatic Lesions

	SCA	MCN	IPMN	SPN	Pseudocyst	Cystic Neuroendocrine Tumor
Location	Any location	Body/tail	Arise from main PD or side branch	Any location	Any location	Any location
Location	Any location	Body/tail	Head > body/tail	Any location	Any location	Any location
Malignant potential	Extremely low	Moderate	Moderate. High when main PD involved	Yes	None	High when tumor >2 cm or if a high mitotic rate (Ki-67)
EUS features	Multiple small cysts; often microcystic “honeycomb”; central fibrosis or calcification (“central scar”)	Macrocystic, septations, nodularity or papillary projections. Peripheral calcifications seen in 10%–25%	Dilated main PD or dilated side branch	Mixed solid-cystic; hemorrhagic center	Unilocular, variable size and wall thickness; echogenic material; features of acute/chronic pancreatitis	Round, well-demarcated hypoechoic solid lesion with anechoic area of cystic degeneration
EUS features	Macrocystic and solid variants also possible		Can have mural nodule or mass	Mixed solid-cystic; hemorrhagic center
Communication with PD	No	Rare	Yes	No	Sometimes	No
Cytology	Bland glycogen-positive cuboidal cells	Columnar/cuboidal, mucin-positive cells; may show mild, moderate, marked atypia or invasive adenocarcinoma	Columnar/cuboidal mucin-positive cells; may show mild, moderate, marked atypia or invasive adenocarcinoma	Heterogeneous; eosinophilic, papillary cells, PAS-positive deposits, vimentin positivity	Macrophages, inflammatory cells, debris	Stains positive for the following immunohistochemical markers: chromogranin A, synaptophysin, CDX, CD56, and neuron-specific enolase
Cyst fluid	Low viscosity	High viscosity	High viscosity	Low viscosity, bloody and necrotic	Low viscosity, blood-stained or turbid	Low viscosity; may be blood-stained
Amylase	Variable	Variable	Variable	Variable	High	Variable
CEA	Low	Usually high	Usually high	Unknown	Low	Low
Molecular marker analysis	VHL mutation and LOH in Chr 3	KRAS mutation	KRAS and GNAS mutations	CTNNB1 mutation	No mutations	Not applicable

CEA, Carcinoembryonic antigen; Chr, chromosome; EUS, endoscopic ultrasonography; IPMN, intraductal papillary mucinous neoplasia; LOH, loss of heterozygosity; MCN, mucinous cystic neoplasm; PAS, periodic acid–Schiff; PD, pancreatic duct; SCA, serous cystadenoma; SPN, solid pseudopapillary neoplasm; VHL, von Hippel–Lindau.

There are several features on EUS that are worrisome for malignant transformation of the cyst. These include the presence of a thick wall or septum, an associated solid mass, or a mural nodule ( Fig. 15.3 ). The presence of focal dilation of the MPD, a pancreatic duct measuring ≥10 mm, or an MPD measuring between 5 and 9 mm with a mural nodule are concerning, and are associated with an increased risk of malignant transformation. One of the EUS features that is most difficult to differentiate in predicting malignant potential is a mural nodule versus a “mucin ball.” Several features that can be used and are quite helpful in differentiating these two lesions are listed in Fig. 15.4 and shown in Video 15.1 . Mucin typically has a round, hyperechoic, well-demarcated outer edge with a hypoechoic center. When these features are present on EUS it is associated with mucin in 90% of cases. In addition, mucin can be shown to move within the cyst either by moving the patient’s position or by targeting the lesion during endoscopic ultrasonography-guided fine-needle aspiration (EUS FNA), which is then shown to move with the needle (see Video 15.1 ). One of the key questions when a mural nodule is found is whether there is a difference between a 2-mm and a 20-mm mural nodule. One Japanese group tried to answer this question by classifying mural nodules into four types; type I has 1- to 2-mm fine papillary protrusions; type II has a larger polypoid nodule; type III has a larger, protruding component with a thickened wall; and type IV is a papillary nodule associated with an ill-defined hypoechoic area within the parenchyma. This group found that mural nodules increased in size with a mean diameter of 5, 6, 11, and 20 mm for types I to IV, respectively. Furthermore types III or IV were associated with IC in almost 90% of the patients. The take-home message is that large mural nodules or those associated with an ill-defined mass are highly concerning for the presence of HGD or IC. One technique that can be very helpful is contrast-enhanced EUS (CE EUS). Microbubbles are injected into a peripheral vein, which circulate through the pancreas 30 to 40 seconds later. The technique works on the hypothesis that a malignant tumor has a different vascular pattern than either normal pancreatic tissue or mucin. CE EUS may demonstrate vascularity in mural nodules and is useful for differentiating a mural nodule from a mucin ball. A prospective study found that CE EUS correctly identified 75% of mural nodules with HGD or IC. Tissue harmonicecho (THE) imaging is a further development in EUS. Preliminary studies appear promising and show superior image visualization of mural nodules compared with normal B-mode imaging. Further information about CE EUS can be found in Chapter 5 . Finally, it is important to inspect the entire parenchyma and not just the cyst, particularly in individuals with IPMNs, who can develop concomitant pancreatic adenocarcinoma in a region separate from the cyst.

Video 15.1

Video Demonstrating the Classic Appearance of Mucin Within a Cyst

Confocal Laser Endomicroscopy

Confocal laser endomicroscopy (CLE) is a novel imaging technique in which a low-power laser illuminates and scans a single plane of tissue and generates a real-time optical image or biopsy of the tissue ( Fig. 15.5 ; Video 15.2 ). A needle-based CLE (nCLE) has been developed that can be passed through a 19-gauge EUS FNA needle and enable imaging of the epithelium to several hundred microns. There have been two multicenter studies, one of which was retrospective (INSPECT) and one prospective (CONTACT), that have evaluated the role of nCLE in differentiating SCAs from other types of cysts. The result was that nCLE had excellent (100%) specificity and moderate (59% to 69%) specificity for differentiating SCAs from other types of cysts. A multicenter prospective study (DETECT) found that nCLE again had excellent specificity (100%) and good sensitivity (80%) for discriminating mucin-producing cysts (IPMNs and MCNs) from other types of cysts. The initial results are promising; however, there are significant limitations in the studies published to date. These include that consecutive patients were not always enrolled, many small cysts were excluded, and, most importantly, that the nCLE diagnosis was confirmed by surgical pathology in only 7% to 23% of the cases, with the diagnosis being made by an expert group using imaging and the analysis of cyst fluid. As shown in the next section, the analysis of cyst fluid has limited accuracy. Larger, prospective studies in consecutive patients that compare the results of nCLE with surgical pathology are required to fully evaluate the potential and role of this technology.

Video 15.2

Video of a Confocal Laser Endomicroscopy Examination of a Mucinous Cyst

Endoscopic Ultrasound-Guided Fine-Needle Aspiration

EUS morphology alone is imperfect at identifying the exact type of pancreatic cyst. A large prospective multicenter study found that the accuracy of EUS imaging features for diagnosing mucin-producing cysts was only 51%, although a more recent study from the Netherlands reported a sensitivity of 78%. Slightly worrying in the Dutch study, however, was the sensitivity for detecting malignancy in pancreatic cysts, which was only 25% for EUS. Other retrospective studies have shown a higher accuracy for the detection of malignant/potentially malignant lesions with a reported sensitivity of 91%. However, despite its high resolution, EUS alone clearly has limitations.

EUS FNA is often performed in patients with pancreatic cysts in the Western countries, as it provides additional information that can be helpful in confirming the type of cyst or detecting the presence of HGD or IC. There is no specific type or size of needle that should be used; the choice is determined by the size of the cyst, its location, and the presence of vessels around the cyst. If possible, a single pass is made into the cyst to minimize the risk of complication and the fluid in that cyst locule is fully aspirated ( Fig. 15.6 ). Aspirating the cyst locule to dryness may decrease the risk of infection, although the evidence for this is slim. Lesions that have a vascular component, such as cystic PanNETs, often yield bloody samples if large-bore needles are used; in such instances, a smaller-gauge needle (e.g., 22 or 25 gauge) may be helpful. Microcystic SCAs are particularly challenging to FNA. It is often difficult or impossible to obtain sufficient fluid for analysis from small cysts. In these cases, a 19-gauge needle can sometimes be used to obtain a core biopsy of the cyst. New core biopsy needles have been developed that can be helpful with lesions such as microcystic SCA (discussed in detail in Chapter 21 ). These needles have been shown to be helpful in solid pancreatic lesions; however, there are as yet no data for pancreatic cysts. One helpful technique is to target the wall of the cyst with a gentle to-and-fro motion to try to aspirate cells from the wall itself rather than simply aspirating fluid from the center of the cyst. This has been shown to increase the cytologic diagnostic yield for mucin-producing and malignant cysts.