The management of patients with pancreatic cysts is a significant issue in the health care community. We know from over 30 years of study that mucinous cysts account for a significant percentage of pancreatic cysts, and that mucinous cysts are precursors of pancreatic carcinoma. Based on autopsy studies, approximately 2.6% of the adult population and nearly 8% of elderly patients have a pancreatic cyst.1 How many of these cysts are mucinous and, of those that are mucinous, how many are malignant or sufficiently high-grade to prompt surgical management is a conundrum physicians are facing on an increasingly common basis.2, 3 The detection of mucinous cysts before the development of invasive carcinoma is the goal of management given that invasion is the single most important factor determining patient prognosis.4 Understanding the clinicopathological features of pancreatic cysts is critical in answering the important question of whether surgical management is warranted, and this requires a multidisciplinary approach.5, 6
In 2005, an international and multidisciplinary group of physicians met in Sendai, Japan and developed a consensus document to guide physicians in the diagnosis and management of patients with pancreatic mucinous cysts (also known as the “Sendai guidelines”).7 In brief, these guidelines called for the conservative management of asymptomatic patients with pancreatic mucinous cysts that were small (< 3 cm), not associated with a dilated main pancreatic duct (MPD) (≥ 6 mm) or mural nodule by imaging, and without “positive cytology.” Since their publication in 2006, studies have shown the Sendai guidelines to be highly sensitive (approximately 97%), but very nonspecific (approximately 29%) for the detection of a malignant mucinous cyst.8-10 Cyst size and nonspecific patient symptoms have proven to be poor predictors of malignancy.11, 12 Although the majority of patients with cancer were detected, many patients already had invasive cancers at the time of surgical resection. In addition, many elderly patients with significant comorbid conditions and mucinous cysts with low- to intermediate-grade (moderate) dysplasia were also undergoing surgical resection, which is riskier than conservative follow-up because the progression to carcinoma would likely take longer than the expected life expectancy of the patient. Most importantly, some asymptomatic patients with carcinoma or high-grade dysplasia but with no high-risk imaging features were being followed conservatively.12, 13
In 2010, a new international, multidisciplinary group of physicians (the 8 members of the original group plus 6 new physicians from various specialties) met in Fukuoka, Japan to revisit the Sendai guidelines and to discuss advances in the understanding of the biology, imaging, preoperative testing, and surgical management of patients with mucinous cysts, with the goal of providing a revised management algorithm.14 Because the levels of evidence for the issues addressed by the group were low, the guidelines are considered to be “consensus” rather than “evidence-based” guidelines. These revised guidelines14 address standardized definitions, define imaging criteria, and outline criteria for surgical resection as well as address definitions of histological parameters and surgical approaches in these patients. The preoperative issues of diagnosis that affect cytopathologists are summarized below.
The 2 neoplastic mucinous cysts of the pancreas include mucinous cystic neoplasm (MCN) and intraductal papillary mucinous neoplasm (IPMN). Although both of these mucinous cysts are precursors to invasive carcinoma, they differ with regard to their biological, radiological, pathological, and molecular characteristics.4, 15 MCNs typically occur in middle-aged females, and although the majority are low grade at the time of detection,16 surgical resection is recommended because lifelong surveillance is more expensive and anxiety-producing than a distal pancreatectomy, which is curative for patients with MCN without invasive carcinoma. IPMNs are classified into 3 types: main-duct (MD)–IPMN, branch-duct (BD)–IPMN, and combined type. The definitions of the MD and BD types have varied in the literature, with some investigators using imaging characteristics and others using histological characteristics to define them, with only an approximately 70% concordance between the 2 methods.17, 18 This lack of standardization has caused significant confusion as well as inaccuracy in the comparison of studies using different definitions. The 2012 guidelines recommend defining IPMN by imaging characteristics because such a classification is important for preoperative management decisions. MD-IPMN is defined as a mucinous neoplasm that either segmentally or diffusely involves the MPD being dilated > 5 mm in diameter. BD-IPMN is defined as a mucinous pancreatic cyst measuring > 5 mm that communicates with the MPD.14
The definition of malignancy is also addressed. The Sendai guidelines as well as many prior studies classified high-grade, preinvasive neoplasia as carcinoma in situ, in keeping with the 2000 World Health Organization19 and Armed Forces Institute of Pathology4 definitions. Because the World Health Organization 2010 definition20 confined malignancy to only those mucinous cysts with invasive carcinoma, defining preinvasive cysts as premalignant with either low-, intermediate-, or high-grade dysplasia, the 2012 guidelines have followed suit. Removing cysts with high-grade dysplasia from the malignant category is a very important factor to keep in mind when evaluating studies and determining the appropriateness of surgical management. “Malignancy” is no longer the preoperative detection goal because invasion significantly decreases patient prognosis. Ideally, only those patients with high-grade dysplasia will undergo surgical resection, but this is a tall order for preoperative testing.
Preoperative evaluation uses a combination of clinical history, gender, imaging characteristics, cytology, and cyst fluid analysis. Using imaging, an MPD measuring ≥ 10 mm is now considered a high-risk feature for MD-IPMN (in contrast to the Sendai guidelines, in which an MPD measuring ≥ 6 mm was considered high risk), thus warranting surgical resection without tissue confirmation; an MPD measuring 5 mm to 9 mm in size is now considered “worrisome” for MD-IPMN and should be evaluated further before surgery. Reliable distinguishing features considered diagnostic of BD-IPMN include multiplicity and visualization of a connection to the MPD. High-risk stigmata for malignancy that lead to surgery without further testing of BD-IPMN include obstructive jaundice associated with a cyst in the pancreatic head, an association with an MPD measuring ≥ 10 mm, or an enhancing solid component. Worrisome features on imaging that are suggestive of malignancy in a patient with BD-IPMN but that warrant further evaluation before surgery include cyst size ≥ 3 cm, thickened enhancing cyst walls, an MPD measuring 5 mm to 9 mm, nonenhancing mural nodules, an abrupt change in the caliber of the MPD with distal pancreatic atrophy, and lymphadenopathy.14
Further evaluation is generally conducted with the aid of endoscopic ultrasound–guided fine-needle aspiration. Endoscopic ultrasound provides higher resolution imaging that can evaluate the septa, mural nodules, and other features of the cyst while simultaneously aspirating the cyst contents for evaluation. Cyst fluid provides a wealth of information from the macroscopic appearance to the microscopic details, and from the biochemical to the molecular attributes. With this information, the nature of the cyst can generally be determined as mucinous or nonmucinous on the one hand, and as low risk or high risk based on cytology on the other.
The most accurate test for determining the cyst as mucinous is the carcinoembryonic antigen (CEA) assay.21 Fresh, undiluted fluid from the cyst can be analyzed using an automated immunodiagnostics analyzer and reported in nanograms per milliliter. Studies have shown that concentrations of around 110 to 200 ng/mL are approximately 80% accurate in determining a cyst to be mucinous.13, 21-25 Raising the cutoff level improves the specificity of the test but at the expense of sensitivity.24 Not all mucinous cysts will have such an elevated CEA concentration, and some mucinous cysts have very low concentrations of CEA. Additional support for a mucinous etiology comes from KRAS analysis. KRAS is an early oncogenic mutation in the adenoma-carcinoma sequence and can be detected in patients with low-grade pancreatic intraepithelial neoplasia.26-28 As such, the presence of a KRAS mutation in and of itself cannot distinguish a benign from malignant mucinous cyst. A recent study demonstrated that GNAS mutations detected in cyst fluids separate IPMN from MCN but, similar to KRAS mutations, do not determine malignancy.29 The absence of a GNAS mutation also does not correlate with a diagnosis of MCN because not all IPMNs will demonstrate a GNAS mutation. As such, how this finding will impact preoperative patient management is yet to be determined.
Cytology is the most accurate test for the detection of malignancy in patients with mucinous cysts.21 A cytological diagnosis of “positive” or “malignant” is generally 100% specific but very insensitive for detecting high-grade dysplasia.30, 31 Aspirating the contents of the cyst is really a screening test for the detection of dysplastic cells, akin to a Papanicolaou test. There is no directed biopsy of the cyst lining; the cells that are collected are those floating in the cyst cavity and adjoining ducts and abraded from the cyst wall with the aspiration needle. The cells may well underestimate the highest grade of neoplasia in the cyst given the typical heterogeneity of the lining of both MCNs and IPMNs, especially IPMNs.32 In addition to the typical impediments to cytological interpretation of scant cellularity and poor cellular preservation, there is a lack of standardized, reproducible, diagnostic criteria for the various cyst types and grades coupled with a general lack of experience in the interpretation of pancreatic cysts. Although it may be difficult to accurately correlate cytological grade with histological outcome, simplifying the cytological interpretation and lowering the threshold for defining high-risk from “positive” to “suspicious” has been shown to improve the sensitivity of detecting malignancy and high-grade dysplastic cysts.31 Defining “high-grade atypia” in pancreatic mucinous cysts as high-grade dysplasia or invasive carcinoma may indeed provide a better and more reproducible high-risk cytological threshold for resection. Likewise, the absence of high-grade atypia has been shown to provide a negative predictive value of > 90% for malignancy in the context of the absence of the high-risk imaging features of a dilated MPD and mural nodule, a so-called triple-negative test.30
Our ability to preoperatively diagnose and manage patients with pancreatic mucinous cysts continues to evolve as we gain experience in all disciplines. The 2012 guidelines offer an improved algorithm for patient management that attempts to balance the risk of malignancy with the risk of surgery (Fig. 1).