Pancreas FNA


Guoping Cai, M.D.

Pancreatic lesions comprise a wide spectrum of entities, ranging from non-neoplastic to neoplastic lesions, and benign to malignant neoplasms [1]. Accurate preoperative diagnosis of pancreatic lesions is crucial for proper clinical management. Routine tissue diagnosis with laparotomy or core needle biopsy is seldom performed due to significant morbidity and potential mortality. In addition, core needle biopsy may have potential risk of needle tract tumor implantation. Fine needle aspiration (FNA) has been increasingly used in the diagnosis of pancreatic lesions because of relatively high diagnostic performance and low complication rate [2, 3].

FNA biopsy of pancreatic lesions can be performed via a percutaneous transabdominal or endoscopic approach [3, 4, 5]. Transabdominal biopsy is performed under the guidance of either ultrasound or computed tomography (CT). With the development of linear array echoendoscopy, endoscopic ultrasonography guided FNA (EUS-FNA) has become the choice of diagnostic approach in most institutions [4, 6]. Dependent on the location of the lesion of interest, the biopsy can be performed via a transgastric or transduodenal route. A transduodenal approach is commonly used for the lesions located in the head or uncinate process of the pancreas. If the lesion is located in the body or tail, a transgastric approach is generally used. Recognition of gastric or duodenal epithelial contaminants can be essential to avoid misinterpretation during cytological evaluation of the specimens [7].

Dependent on the nature of pancreatic lesions, different diagnostic approaches may be employed. For solid lesions, a definite diagnosis is highly desirable and can be achieved by cytomorphologic analysis with or without ancillary studies [1, 2, 4, 5, 6, 7]. To render a diagnosis of pancreatic neoplasms other than ductal adenocarcinoma, adjunct ancillary studies including immunocytochemistry may be required [2, 4]. Rapid on-site evaluation during the procedure helps ensure adequate sampling and appropriate specimen triage [8]. In our experience, if there are grossly visible particles in the needle rinse from an FNA pass that shows diagnostic cells on rapid smears, Cellient cell blocks are highly likely to contain diagnostic tissue fragments. If the lesion is cystic, EUS-FNA primarily serves as a tool for risk assessment, identifying the lesions with high grade dysplasia or malignant potential that may require surgical intervention [9, 10, 11]. To achieve this goal, an integrated approach that combines imaging studies, cytomorphologic evaluation, cyst fluid analysis and molecular tests is recommended [11, 12, 13]. For solid pancreatic neoplasms, terminology similar to that used in the histopathology is preferred. For cystic lesions, a descriptive diagnosis may be acceptable.

Ductal adenocarcinoma is the most common malignancy seen in the pancreas, accounting for 75% to 90% of all solid pancreatic masses. Most ductal adenocarcinomas can be diagnosed without difficulty based on cytomorphologic analysis alone. However, well-differentiated adenocarcinoma can impose a diagnostic challenge due to subtle cytologic atypia. There are some cytomorphologic features that may provide a diagnostic clue [1]. The epithelial groups seen in well-differentiated adenocarcinoma are often disorganized and have a drunken honeycomb pattern. In addition, these epithelial cells show variable degrees of anisonucleosis and nuclear membrane irregularity. Nevertheless, well-differentiated adenocarcinoma can still be difficult to diagnosis in some cases. Furthermore, superimposed chronic pancreatitis and pancreatic intraepithelial neoplasia may complicate the diagnosis [14, 15]. A few promising immunohistochemical markers to distinguish chronic pancreatitis from pancreatic adenocarcinoma have begun to emerge [16, 17], and cell blocks provide an ideal platform for immunohistochemistry. Cell block sections may display better tumor cell morphology and diagnostic growth patterns, and they can also demonstrate the relationship of tumor with surrounding tissue [18]. Cell block sections provide a common platform that can be shared with surgical pathologists to allow diagnosis of morphologic variants of pancreatic carcinomas.

Acinar cell carcinoma is a rare malignant tumor which is difficult to diagnose based on cytomorphologic features alone. It must be distinguished from pancreatic endocrine neoplasm as well as benign acinar tissue [1, 19]. Architectural features evident in cell block sections can be very important because the cytologic features may be very bland. Loss of normal acinar structures with the production of aberrant architectures such as solid and trabecular patterns can readily be appreciated on the cellblock sections. Zymogen granules in acinar cell carcinoma can be highlighted by diastase-digested PAS (D-PAS) stain, and trypsin and chymotrypsin immunostains.

Solid pseudo-papillary tumor and pancreatic endocrine neoplasm are two other entities that may present as a solid or partially cystic mass. They have characteristic cytomorphologic features which are well documented in the literature [1, 20, 21, 22, 23]. However, immunocytochemical studies are often performed to confirm the diagnosis. Immunochemically, solid pseudo-papillary tumor is positive for vimentin, beta-catenin, and CD10, and negative for cytokeratin [20, 23, 24]. Diagnosis of pancreatic endocrine neoplasms is supported by immunoreactivity with chromogranin and synaptophysin [2, 21, 22]. Thus, additional material should be saved for cell block preparation if the differential diagnosis includes the entities mentioned above.

Cystic pancreatic lesions include non-neoplastic, benign and malignant neoplasms. Evaluation of cytomorphologic features alone is often insufficient for accurate classification of these lesions. Limiting factors include scant cellularity, overlapping cytomorphologic features, and gastrointestinal contaminants. Incorporation of imaging findings, cyst fluid analysis, and molecular test results with cytological evaluation is highly recommended [11]. Imaging studies define the size, location, and complexity of the lesion and its relationship with the pancreatic duct. Cyst fluid analysis including CEA and amylase levels helps in separating non-neoplastic pancreatic pseudocysts from neoplastic non-mucinous and mucinous cysts [9, 12, 13]. Assessment of malignancy, or risk for malignancy, may benefit from molecular tests such as K-ras mutation analysis and loss of heterozygosity (LOH) of tumor suppressor genes [11, 12, 13].

Among the cystic lesions, intraductal papillary mucinous neoplasm (IPMN) has unique imaging findings and characteristic cytomorphologic features. The presence of papillary fragments, characteristic haphazard mixtures of goblet cells within mucinous epithelium, and a cystic mucinous background with histiocytes helps render the diagnosis of IPMN [25]. However, these cytomorphologic features, particularly true papillary fragments with fibrovascular cores, are uncommonly seen in the aspirates. Further, the distinction between normal gastric, normal small intestinal, and the sometimes minimally atypical epithelium of an IPMN can be difficult on monolayer preparations. Careful examination of cell block material may reveal these histologic diagnostic features. Overall, FNA has shown a relatively low diagnostic rate for IPMN [25, 26]. IPMN is cytologically almost indistinguishable from a mucinous cystic neoplasm [25]; imaging findings usually can make this distinction. Many cytologists lump IMPN and mucinous cystic neoplasms together by simply diagnosing "mucin-producing cystic neoplasm", with an attempt to characterize whether or not there is evidence of high grade atypia/dysplasia in the sample [27]. Since histology is still considered to be the gold standard for diagnosis of high grade dysplasia [28], cell blocks are particularly valuable for creating a common platform for developing consensus with surgical pathologists on this important diagnosis. Examples are shown. Cyst fluid analysis may have some value in predicting malignancy [29]. Serous cystadenoma can also be diagnosed by cytology and distinguished from mucinous cystic neoplasms [30,31,32].

It is not uncommon for a pancreatic mass to represent a metastasis. These patients may or may not have a documented history of malignancy. The most common carcinomas metastatic to the pancreas are renal cell carcinoma and lung carcinoma [33,34,35]. Malignant mesenchymal neoplasms may also involve the pancreas [36]. These metastatic tumors may display cytomorphologic features that overlap with primary pancreatic neoplasms. Cell block sections provide a better assessment of their morphologic features and can be used for immunocytochemical analysis, which is a crucial component for work-up of the origin and precise diagnosis.

Primary pancreatic lymphomas are rare [37,38,39]. More often, pancreatic lymphomas represent a systemic involvement by lymphoproliferative disorders. Diffuse large B-cell lymphoma and plasma cell neoplasm are probably the two most common lymphoproliferative disorders involving the pancreas. Although these lymphoproliferative disorders may have unique cytomorphologic features, ancillary immunophenotyping studies such as flow cytometry or/and immunocytochemistry are required for the diagnosis. It is crucial to recognize these lymphoid lesions during the on-site evaluation. Part of the aspirates should be saved for flow cytometry as well as processed for cell block preparation. Cell block preparation serves for both morphologic evaluation and potential immunocytochemical studies.

In this chapter, we will illustrate the benefits of combinational cytology and cell block preparations to improve FNA diagnosis of pancreatic lesions.