Challenges and Pitfalls in Pancreatobiliary Cytopathology Open Access

One of the most challenging areas of cytopathology is the evaluation of the pancreas and the biliary system since these organs can be difficult to access and frequently yield minimal and fragmented specimens [1]. Moreover, there is significant morphologic overlap between numerous neoplastic and nonneoplastic mimickers [1]. Here we describe a few challenging examples with potential pitfalls when examining pancreatobiliary cytopathology. All cases represent real-life routine experience encountered during daily cytopathology practice in a busy academic hospital (tertiary) center.

A 66-year-old male was found on an abdominal MRI to have a 2.4-cm heterogeneously enhancing pancreatic tail mass and an enlarged peripancreatic lymph node suspicious for metastasis. Patient underwent an endoscopic ultrasound-guided fine needle aspiration (FNA) of the mass that yielded a paucicellular specimen composed of clusters and single cells with abundant granular cytoplasm and monomorphic nuclei with indistinct nucleoli (Fig. 1). Cell block was extremely limited and contained scant atypical cells in a background of benign pancreatic parenchyma. The case was diagnosed as suspicious for acinar cell carcinoma (ACC). The lesion was resected, and final pathology report revealed a neuroendocrine tumor with involvement of regional lymph nodes.

ACC is a malignant pancreatic neoplasm showing acinar cell differentiation [2, 3] that accounts for about 1–2% of pancreatic neoplasms in adults and 15% in children [4]. The overall survival for ACC is usually between ductal adenocarcinoma of the pancreas and pancreatic endocrine neoplasms [5, 6]. The clinical manifestation usually varies according to the location and stage and can include symptoms of lipase hypersecretion like subcutaneous fat necrosis and polyarthralgia, usually in patients with extensive metastasis [2, 7]. Morphologically, they are characterized by highly cellular specimens containing structures resembling normal acini with cells with moderate amounts of granular eosinophilic cytoplasm containing zymogen granules and uniform nuclei with single prominent nucleoli [2, 3].

In view of the resemblance with benign pancreatic acini, the cytological diagnosis of ACC can be challenging at times specially in paucicellular specimens since hypercellularity is usually a key feature used to identify ACC [2, 3]. Moreover, the broad differential diagnoses of solid pancreatic neoplasms which include poorly cohesive ductal adenocarcinoma, ACC, pancreatoblastoma, solid pseudo-papillary and neuroendocrine neoplasms should be considered since often, and especially in hypocellular specimens, the distinction between those entities can be challenging and sometimes not possible in the biopsy/aspirate specimen and being reserved for the final resected specimen [2, 3]. It is always recommended to collect material for cell block when suspecting ACC or neuroendocrine tumors because immunohistochemistry and special stains can facilitate the diagnosis. In conclusion, caution rendering a definite diagnosis is advised when dealing with paucicellular specimens with atypical morphological features, and the cytopathologist should keep a broad differential when dealing with such specimens.

An 84-year-old man with a past medical history of chronic pancreatitis (CP) was found on an abdominal CT scan to have a 2.3 cm ill-defined hypoenhancing mass in the head/neck area of the pancreas with circumferential encasement of the portal vein and upstream dilation of the main pancreatic duct concerning for malignancy. There were associated cystic lesions in the pancreas which was also showing background atrophy. An FNA was obtained from the mass and showed rare, atypical cells in a background of CP. An FNA of the nearby separate cystic lesion revealed at least high-grade epithelial atypia, from which an invasive carcinoma could not be ruled out (Fig. 2a–c). A subsequent core biopsy from the mass was performed 2 months later and revealed detached atypical epithelial cells (Fig. 2d).

CP is a chronic progressive disease with an annual incidence of 5–8 and prevalence of 42–73 cases per 100,000 adults in the USA [8‒10]. The diagnosis is usually obtained through a combination of clinical, laboratory, and radiological findings [11]. Advanced, late-stage CP may present with mass-forming fibrosis and can be indistinguishable radiologically from a neoplastic process, specially in cases with exuberant fibrosis like autoimmune pancreatitis [12]. To complicate matters, CP is a risk factor for development of pancreatic cancer [11]. Therefore, cytological specimens from patients with CP with the goal to exclude malignancy are a common scenario among daily practice of pathology.

Morphologically, CP is characterized by the cardinal triad of fibrosis, loss of acinar tissue (atrophy), and duct changes [12]. The duct changes may include distorted and/or dilated ducts with luminal obstruction by proteinaceous material and loss of epithelium [12]. Reactive epithelial atypia including chromatin changes, irregular nuclear contours, prominent nucleoli, and increased nuclear-to-cytoplasm ratio may be present and can be extensive which might mimic malignancy, specially in small, limited specimens. Moreover, CP specimens may harbor focal pancreatic intraductal neoplasia which coupled with the radiologic appearance of fibrosis mimicking invasive cancer can lead to a wrong diagnosis of invasive carcinoma. In conclusion, despite the increasing demand for pathologists to make a diagnosis of invasive carcinoma on fine needle and small core biopsies of the pancreas, one should be cautioned doing so in a setting of CP even if the imaging is suspicious for malignancy.

A 35-year-old man with past medical history of inflammatory bowel disease and primary sclerosing cholangitis (PSC) was found to have a severe common bile duct stricture measuring 1.2 cm in length during a magnetic resonance cholangiopancreatography. The stricture was deemed “indeterminate” regarding neoplastic risk, and tissue sampling was recommended. Patient subsequently underwent an endoscopic retrograde cholangiopancreatography with a biliary brushing being obtained. Morphologically, the cytological specimen was poorly preserved, paucicellular and composed predominantly of small clusters and single cells revealing mild nuclear atypia (high nucleus-to-cytoplasm ratio, irregular nuclear contours, and chromatin distribution) (Fig. 3a, b). Despite the atypia, the scarcity and poor preservation of the material precluded a more definitive diagnosis, and the specimen was diagnosed as atypical, cannot exclude carcinoma. The specimen was submitted for FISH analysis using probes for CEP3, CEP7, and CEP17. The results were negative for polysomy with a note that there were limited number of cells available for analysis. The patient underwent a needle biopsy of the stricture that revealed an EBV-positive lymphoepithelioma-like carcinoma with positive in situ hybridization for EBER (Fig. 3c).

PSC is a rare, chronic cholestatic liver disease characterized by both intrahepatic and extrahepatic bile duct injury, leading ultimately to fibrosis with stricturing, cholestasis, biliary cirrhosis, and progressive liver dysfunction and failure [13]. The prevalence of PSC is up to 16.2 per 100,000 population, being higher in northern Europe and lower in Asia [14‒17]. Patients diagnosed with PSC are at an increased risk of development of cholangiocarcinoma with 10–20% of patients ultimately developing this malignancy [18].

The diagnosis of PSC is usually made with the combination of clinical, laboratory, and imaging data rarely requiring tissue diagnosis [13]. Nonetheless, biopsies are frequently performed, especially in scenarios where a differential diagnosis is considered and/or there is a need to exclude a superimposed neoplastic process [13, 19]. One example is the presence of biliary strictures in imaging studies that is simultaneously a key feature of PSC but may also indicate presence of malignancy [20]. In those situations, endoscopic-guided biliary sampling (brushing or needle biopsy) can be helpful but also comes with some pitfalls. Like pancreatitis, cholangitis can also be associated with reactive ductal epithelium atypia which should be distinguished from carcinoma. A high threshold for malignancy is needed for brushings from patients with PSC, primary biliary cirrhosis, biliary stones, or an indwelling stent since there is an overlap of morphological findings (presence of mitoses, prominent nucleoli, and nuclear enlargement) found in both benign and malignant processes [21, 22]. Extreme cases like false-positive diagnosis of carcinoma in patients with reactive changes due to PSC have been reported [23]. Certain morphological features including irregular chromatin distribution, parachromatin clearing, irregular nuclear membranes, loss of polarity, nuclear crowding, anisonucleosis, and variation in nuclear diameter more than three or four-fold within a group of cells tend to be associated with malignancy, but none of them are pathognomonic [24, 25].

In challenging cases, ancillary tests including immunohistochemical analysis for SMAD4/DPC4, p16 and p53 and FISH analysis of polysomy can be helpful in distinguishing between reactive and malignant bile duct processes, but they all require an adequate minimum number of cells to be correctly interpreted to avoid false negatives [26‒28]. In our patient, the FISH analysis was negative for polysomy, but a note highlighting limited number of cells was added. In conclusion, the threshold to diagnose malignancy in specimens from patients with PSC, primary biliary cirrhosis, stones, and catheter should be high, especially if dealing with limited material and ancillary tests results should be interpreted with caution in paucicellular specimens.

A 64-year-old woman was found to have multiple hypervascular solid masses in the pancreas head, neck, and body. The patient had a history of resected clear cell renal cell carcinoma (RCC) 11 years earlier. An endoscopic FNA of one of the pancreatic masses was performed and yielded a specimen containing clusters and single cells with moderately abundant eosinophilic to clear cytoplasm with small round nuclei with inconspicuous nucleoli (Fig. 4). A cell block was prepared and immunohistochemical studies revealed the suspicious cells to be positive for cytokeratin AE1-3, PAX-8, and carbonic anhydrase IX (CA-IX). A diagnosis of metastatic clear cell RCC was rendered.

A 56-year-old man was admitted to the Emergency Department with nausea, vomiting, headaches, and loss of coordination. A head MRI revealed multiple intraparenchymal metastatic lesions. Patient had no significant past medical history including oncologic history; therefore, no primary site for the metastases was known. The patient then underwent a full body PET-CT that revealed a poorly visualized abnormal FDG-avid area involving stomach and pancreas which was further characterized in the magnetic resonance cholangiopancreatography as few pancreatic lesions and peritoneal implants. An endoscopic FNA was performed and yielded a cellular specimen composed predominantly by small loosely cohesive clusters and single cells with variable N/C ratio, pleomorphic nuclei, and inconspicuous nucleoli (Fig. 5a). A cell block was prepared, and immunohistochemical studies were performed. The suspicious cells stained positive for S100p, SOX-10, and HMB45 and were negative for cytokeratin, INSM1, synaptophysin, chromogranin, and CD163 (Fig. 5b–d). A diagnosis of metastatic melanoma was rendered.

Pancreas is an uncommon site for metastases since the majority of tumors identified in the pancreas are primary [29, 30]. The overall incidence of pancreatic metastases varies from 2% to 5% of all pancreatic malignancies [31]. Most patients with secondary pancreatic tumors have widespread disease and are not candidates for resection [32]. Cases with metastases confined to the pancreas at the time of diagnosis are rare but do exist, for which surgical resection may offer survival benefits [33‒35]. Most pancreatic metastases are from a primary RCC [31, 36, 37]; however, metastases from other sites including lung, breast, colon, skin (melanoma), and sarcoma have been reported [29, 38‒40].

Metastatic neoplasms should be considered in the differential diagnosis of solid pancreatic masses when the imaging and/or morphological findings are atypical for primary disease even in the absence of documented clinical history of previous malignancy. Often, RCC mimics a primary pancreatic neoplasm like foamy-gland variant of the pancreatic ductal adenocarcinoma and clear cell (lipid-rich) variant of pancreatic neuroendocrine tumor of the pancreas [37, 38, 41, 42] and even ectopic adrenal cortical tissue [36]. Likewise, melanoma has been reported to give metastases to the pancreas and can frequently mimic primary pancreatic neoplasms, especially ACC and neuroendocrine tumors [43‒45]. In challenging situations, collecting material for cell block is paramount as immunohistochemical studies can tremendously help identifying the primary site of origin.

The increased utilization of endoscopic minimally invasive procedures that often yield small samples has contributed to an increase in evaluation of pancreaticobiliary lesions. In view of the limited nature of these samples, this is now one of the most challenging areas in the daily practice of cytopathology. When dealing with challenging cases during the rapid on-site evaluation, the cytopathologist must keep in mind the broad differential diagnoses and use the often scant material in an efficient way to maximize the diagnostic yield. In that regard, separating material for cell block is often the best practice since it can expand the possibilities of ancillary tests that can be performed.

This paper did not require prior approval of the Institutional Committee of Ethics in accordance with local/national guidelines. Written informed consent was obtained from the patient for use of the details of their medical case and any accompanying images.

The authors declare no conflict of interest.

This research did not require funding.

J.V.S.-C. and M.T.S. contributed to the study design and were responsible for the final review of the manuscript. J.V.S.-C. was primarily responsible for selecting the cases and writing the manuscript.

All data generated or analyzed during this study are included in this article. Further inquiries can be directed to the corresponding author.