—  SPECIALTY CONFERENCE  —

Surgical Pathology

Case 3 - Metastatic Pancreatic Neuroendocrine Tumor (PanNET 2)

Wendy L. Frankel, The Ohio State University, Columbus, OH





Virtual Slides as well as Still Images are displayed below.
For the fastest viewing of virtual slides, click:



under each thumbnail image below. You must have Aperio ImageScope installed on your PC.
If you do not already have Aperio ImageScope, Windows users with administrator privileges may download and install a free version in order to view USCAP Virtual Slides. Click the icon on the right to get your free copy:  
Or, click on slide thumbnail images to view each slide
in a Web-based slide viewer, which is somewhat slower.

If you have any difficulties viewing these slides, email or call George Clay at +1.724.449.1137.



Clinical History
75 year old woman presented in May 2008 with painless, jaundice, pruritus, post-prandial epigastric pain and weight loss over the past month. CT scan showed a 4.9 cm mass in the head of the pancreas that did not extend into the adjacent vessels. She was thought to have a resectable pancreatic adenocarcinoma and was brought to the operating room for a Whipple procedure. After opening the abdomen, she was found to have a single, small subcapsular liver lesion. A frozen section was requested to determine whether the surgeons should proceed with the Whipple procedure or place a stent and abort the procedure.


Case 3 - Slide 1
Click to view with ImageScope
Click to view with a Web-Based Viewer


Case 3 - Figure 1
Scanning power view of the frozen section of the liver lesion.

Case 3 - Figure 2
Medium power image of the previous showing a proliferation of cells arranged in nests and glands.

Case 3 - Figure 3
High power image of the previous showing a proliferation of glands containing pigment (right upper) and nests/glands (center and left).

Case 3 - Figure 4
High power image of the frozen section showing mild nuclear atypia without mitotic figures.

Case 3 - Figure 5
High power image of the frozen section showing focal sheet/nests of cells with mild cytologic atypia without mitotic figures.

Case 3 - Figure 6
Low power image of the permanent control of the frozen section of the liver.

Case 3 - Figure 7
High power image of the previous showing a fairly regular proliferation of cells arranged in glands and nests.

Case 3 - Figure 8
Cytokeratin AE1/3 immunostain showing weak staining in the tumor cells and strong staining in the proliferating bile ductules.

Case 3 - Figure 9
Synaptophysin immunostain showing staining in the tumor cells but not bile ductules.

Case 3 - Figure 10
The immunohistochemical stain for Ki67 showing a proliferative index of approximately 5 to 10%.

Case 3 - Figure 11
Medium power image of the pancreas resection specimen showing a proliferation of fairly regular and round tumor cells with adjacent benign pancreatic acini (right).

Case 3 - Figure 12
High power image of the previous showing a proliferation of fairly regular and round tumor cells arranged in nests and pseudoglands.

Introduction:
One of the most common liver lesions sampled for frozen section is a small subcapsular mass found during laparotomy for another lesion. In the case of a planned pancreas resection for pancreatic adenocarcinoma, the main concern is metastatic adenocarcinoma. The diagnosis has important clinical implications as it may alter the planned operation. The resection will continue with a benign diagnosis but will be aborted with adenocarcinoma.

Pathological/Microscopic Findings and any Immunohistochemical or Other Studies:

Operative/Gross findings:
A single tan subcapsular liver lesion was found that measured 0.5 cm. The remaining liver appeared unremarkable. The pancreatic tumor did not extend into the major blood vessels or other organs and no other possible metastases were noted. The pancreatic tumor was considered resectable.

Frozen section microscopic findings:
The lesion was entirely submitted for microscopic evaluation in one block. It is fairly well circumscribed and is composed of a proliferation of glands and nests (Fig 1,2). The cells are round and regular with a moderate amount of cytoplasm and focal nuclear atypia and hyperchromasia without mitotic figures or necrosis (Fig 3,4,5). No definite features of malignancy could be identified. Some of the glands at the periphery appear to contain pigment (Fig 3).

Permanent control microscopic findings:
The proliferation is composed of glands and nests and is regular without atypia or mitotic figures (Fig 6,7). An iron stain shows iron in the peripheral glands that had pigment identified on H&E staining.

Immunohistochemical stains were positive for cytokeratin AE1/3 (weak in lesion, strong in proliferating bile ductules at the periphery) (Fig 8), synaptophysin (Fig 9) and chromogranin. Ki67 shows a proliferation index of 5 to 10% (Fig 10) and p53 shows focal weak staining in lesional cells. The pancreas resection specimen contains a tumor composed of round and regular tumor cells arranged in nests and pseudoglands without cytologic atypia or mitotic figures (Fig 11,12).

Differential Diagnoses:
Gross finding- Small subcapsular liver lesion:
  • Bile duct adenoma (peribiliary hamartoma)

  • von Meyenburg complex (biliary microhamartoma)

  • Metastatic pancreatic adenocarcinoma

  • Metastatic neuroendocrine tumor

  • Other metastasis

  • Reactive bile duct proliferation

  • Hemangioma

  • Granuloma

  • Other primary liver tumors or tumor-like lesions (focal nodular hyperplasia, hepatic adenoma, hepatocellular carcinoma)
Microscopic finding-Proliferation of glands/nests:
  • Bile duct adenoma (peribiliary hamartoma)

  • von Meyenburg complex (biliary microhamartoma)

  • Reactive bile duct proliferation

  • Metastatic pancreatic adenocarcinoma

  • Metastatic neuroendocrine tumor

  • Other metastasis

Final Diagnosis:
Metastatic pancreatic neuroendocrine tumor (PanNET 2)

Case Discussion:

Pancreas cancer diagnosis and frozen section:
The diagnosis of pancreatic adenocarcinoma is typically based upon clinical and radiographic findings. Patients may undergo resection without tissue diagnosis if they are operative candidates and the tumor appears resectable. Many surgeons feel that in patients suspicious for pancreatic adenocarcinoma, they would proceed with the resection regardless of the diagnosis, since chronic pancreatitis frequently surrounds adenocarcinoma. Preoperative tissue/cytologic diagnosis is performed in most who appear unresectable (planned bypass and medical treatment) and some who are resectable. Perioperative biopsy is performed most commonly for frozen section in those who have unusual findings at laparotomy, possible metastatic sites or to assess margins [4].

Pancreatic adenocarcinoma is considered unresectable when tumor extends to the celiac axis or is metastatic to distant sites such as the liver. The diagnosis of metastatic adenocarcinoma would stop a planned resection. In a patient with an otherwise resectable pancreatic tumor, benign liver lesions, would not alter the planned pancreatic resection. The diagnosis of a metastatic neuroendocrine tumor would not stop the planned pancreatic resection [4].

Frozen section of subcapsular liver lesions:
The microscopic differential of a proliferation of glands includes bile duct adenoma (peribiliary hamartoma), von Meyenburg complex (biliary hamartoma), reactive bile ductular proliferation, metastatic adenocarcinoma, metastatic neuroendocrine carcinoma, and other metastasis.

Bile duct adenomas have been shown to have the phenotype of normal peribiliary glands rather than bile ducts [2, 8]. Most bile duct adenomas are solitary but they can be multiple and are usually less than 1 cm in diameter. They are characterized by a well-circumscribed proliferation of small acini and tubules lined by cuboidal to columnar cells in fibrous stroma with inflammation. There is no hyperchromasia, atypia or mitosis [1, 5, 6, 10]. The tubules do not contain bile.

The von Meyenburg complex is thought to be a developmental malformation that may be associated with solitary bile duct cysts and polycystic liver disease. They may be multiple and are usually less 0.5 cm and are characterized by irregularly shaped duct-like structures within fibrous stroma. The duct-like structures are dilated or U-shaped and may contain proteinaceous or bile-stained secretions. They are lined by low cuboidal epithelium without atypia [5, 6, 10].

Reactive benign bile ductular proliferation can be due to a reaction to an adjacent lesion with obstruction or a more diffuse finding due to a medical liver condition. They do not typically present as a single lesion in the liver. The ducts can be angulated and appear infiltrative and there may be cytologic atypia. The presence of surrounding normal portal structures, edema and acute inflammation may be helpful in identifying this lesion [12].

Metastatic adenocarcinoma often presents as multiple liver lesions but can be solitary and small. There is usually cytologic and architectural atypia. The tumor cells have irregular nuclear membranes and mitotic figures. The growth pattern tends to be infiltrative with desmoplasia [3, 4, 9, 10, 12]. Of course, well-differentiated tumors can sometimes be mistaken for a benign lesion, most commonly a bile duct adenoma.

Metastatic neuroendocrine tumor in the liver can be multiple or solitary. The cells can be arranged in nests, acini, pseudoglands, trabeculae and sheets causing confusion with multiple other lesions in the liver [4]. Less common features including cystic architecture and clear cell and oncocytic change have been seen. The degree of atypia and mitotic activity can vary from none to extensive. Neuroendocrine tumors need to be distinguished from endocrine clusters that can occur in some bile duct adenomas [11].

Our case showed some features suggesting bile duct adenoma including a single small subcapsular lesion composed of tubules/glands with focal mild cytologic atypia and no mitotic figures. However, the focal nests of cells and cytologic atypia raised concern for metastatic adenocarcinoma. The proliferating bile ductules at the periphery of the lesion led to confusion by suggesting a mainly glandular lesion rather than a nested and pseudoglandular arrangement. The pigment (iron) was present in these ductules and adjacent liver but not in the tumor. The architecture together with some cytologic atypia raised the possibility of a neuroendocrine tumor.

Clinical course and pathologic workup:
The surgeon's working diagnosis based on preoperative clinical and radiologic findings was pancreatic adenocarcinoma. The intraoperative frozen section was called atypical, and the pathologist's differential was bile duct adenoma vs. metastatic adenocarcinoma. Since adenocarcinoma could not be excluded, the surgeon aborted the planned Whipple procedure and did a bypass procedure.

The distinction between benign and malignant glandular lesions in the liver is usually straight-forward on large, well-fixed specimens. Unfortunately, frozen sections are neither large nor well-fixed. The most helpful features include some of those that have been found helpful in frozen section diagnosis of pancreatic adenocarcinoma. Criteria that are frequently seen in adenocarcinoma include nuclear size variation greater than 4 to 1, partial gland lumina/cribriforming/single cells, and infiltrative architecture [3, 4, 9, 12]. Mitotic figures or necrosis should not be seen in most benign lesions. Hyperchromasia is not a useful finding at frozen section since many benign cells appear hyperchromatic with frozen artifact.

On H&E permanent control sections, the working diagnosis was initially similar to that on frozen section. However, there appeared to be too much atypia for a bile duct adenoma but not sufficient for pancreatic adenocarcinoma. The initial immunohistochemical stains performed included p53 and Ki67 to help distinguish bile duct adenoma from adenocarcinoma. The low p53 and Ki67 did not seem to be consistent with the favored diagnosis of adenocarcinoma.

The distinction of benign from malignant glandular lesions has been assessed using immunohistochemical stains in several studies and may be helpful for the final diagnosis [7, 13, 14]. Of course, this will not be useful at the time of frozen section. Hornick et al. studied bile duct adenomas, von Meyenburg complex, metastatic pancreatic and other adenocarcinomas. They showed that the expression of p53, TAG-72, mCEA, mesothelin and loss of Dpc4 could help distinguish metastatic pancreas adenocarcinoma from the benign lesions while CK7, CK8/18, CK19 or pCEA did not [7]. Tretiakova et al. studied biopsies, resections and cell blocks from FNAs and showed that the absence of CD10 in atypical biliary epithelial cells is a useful finding to distinguish malignant from benign bile duct lesions [14]. Others have found Ki67, bcl2, MUC4, S100P, von Hippel-Lindau gene product and IMP3, among other markers, useful [13].

After review of the case by other pathologists, the possibility of a metastatic neuroendocrine tumor was considered and synaptophysin, chromogranin and cytokeratin AE1/3 were performed, confirming the diagnosis. The iron stain showed that the iron was not in tumor cells but in the surrounding liver and proliferating bile ductules. The surgeon was surprised by the diagnosis since he expected that the pancreatic tumor was adenocarcinoma, not a neuroendocrine tumor.

Since PanNET metastatic to the liver is still considered best treated by resection (in contrast to adenocarcinoma), the patient was subsequently brought back to the operating room and a Whipple procedure was performed. A pancreatic neuroendocrine tumor (PanNET 2) was diagnosed with metastasis in 5/21 lymph nodes and an additional liver metastasis (not easily resectable) was treated postoperatively with transarterial chemoembolization (TACE). The pancreatic tumor was primarily arranged in nests with focal pseudoglands and no cytologic atypia.

Conclusion(s):
The pitfall that this case emphasizes is that if you do not consider a diagnosis, you will miss it. Do not ignore possibilities that are not considered by the surgeon (or the frozen section pathologist). A clue that could have been identified at frozen section was the focal nests of cells. Nests of cells are not usually seen in a bile duct adenoma, and the atypia was not sufficient for a typical adenocarcinoma. The nests could have prompted consideration of other less likely entities in the differential diagnosis. Additionally, requesting a portion of the primary pancreatic tumor may have been helpful.

The primary differential for a subcapsular liver lesion composed of nests/glands includes bile duct adenoma, von Meyenburg complex, reactive bile duct proliferation, metastatic adenocarcinoma or other metastatic tumors. Although the most important differential at frozen section is usually whether a lesion is benign or malignant, sometimes the tumor type may be important and could alter the surgical management. In this case, a differential that included neuroendocrine tumor may not have altered the operation and possibly atypia was the best diagnosis at the time. However, recognizing the possibility of neuroendocrine tumor was vital in the workup of the permanent sections and ultimately led to the proper treatment. It may not be possible to classify a tumor at frozen section, but giving a differential including a low grade tumor rather than typical adenocarcinoma or a diffuse lesion such as lymphoma may be helpful.

References:
  1. Allaire GS, Rabin L, Ishak KG, Sesterhenn IA. Bile duct adenoma. A study of 152 cases. Am J Surg Pathol. 12 (9)708-715, 1988.

  2. Bathal PS, Hughes NR, Goodman ZD. The so-called bile duct adenoma is a peribiliary gland hamartoma. Am J Surg Pathol. 20(7):858-864, 1996.

  3. Cioc A, Ellison EC, Proca DM, Lucas J, Frankel WL. Frozen section diagnosis of pancreatic lesions. Arch Pathol lab Med, 126(10):1169-1173, 2002.

  4. Frankel WL, Proca DM. Frozen Section Library: Pancreas. New York, NY: Springer; 2011: 73-80.

  5. Goodman ZD, Terracciano LM. Tumours and tumour-like lesions of the liver. In: Burt A, Portmann B, Ferrell L, eds. MacSween's Pathology of the Liver. Philadelphia, PA: Elsevier; 2011: 798-800.

  6. Govindarajan S, Peters RL. The bile duct adenoma. A lesion distinct from Meyenburg complex. Arch Pathol Lab Med. 108(11):922-924, 1984.

  7. Hornick JL, Lauwers GY, Odze RD. Immunohistochemistry can help distinguish metastatic pancreatic adenocarcinomas from bile duct adenomas and hamartomas of the liver. Am J Surg Pathol. 29(3):3813-89, 2005.

  8. Hughes NR, Goodman ZD, Bhathal PS. An immunohistochemical profile of the so-called bile duct adenoma clues to pathogenesis. Am J Surg Pathol 34(9):1312- 1318, 2010.

  9. Hyland, C, Kheir SM, Kashlan MB. Frozen section diagnosis of pancreatic carcinoma. A prospective, study of 64 biopsies. Am J Surg Pathol, 5:179-191, 1981.

  10. Lamps L, Misdraji J, Mino-Kenudson M, Yerian L, Kakar S, Deshpande V, Wang H, Kim GE, Yeh MM, Caradine KD. Liver: Tumors of the liver. In: Diagnostic pathology: Hepatobiliary and Pancreatic. Manitoba, Canada: Amirsys Publishing, Inc; 2011: 30-33.

  11. O'Hara BJ, McCue P, Markku M. Bile duct adenomas with endocrine component: Immunohistochemical study and comparison with conventional bile duct adenomas. Am J Surg Pathol. 16(1):21-25, 1992.

  12. Pai R, Wilcox R, Noffsinger A, Hart J. Liver, extrahepatic biliary tree, gallbladder, and pancreas. In: Taxy J, Husain A, Montag A, eds. Biopsy Interpretation Series Biopsy Interpretation: The Frozen Section. Philadelphia, PA: Lippincott Williams & Wilkins; 2010:228- 237.

  13. Tan G, Yilmaz A, DeYoung BR, Behling C, Lehman A, Frankel WL. Immunohistochemical analysis of biliary tract lesions. Appl Immunohistochem Mol Morphol 12(3):193-197, 2004.

  14. Tretiakova M, Antic T, Westerhoff M, Mueller J, Himmelfarb EA, Wang HL, Xiao S-Y.. Diagnostic Utility f CD10 in Benign and Malignant Extrahepatic Bile Duct Lesions. Am J Surg Pathol. 36(1):101-118, 2012.