—  SPECIALTY CONFERENCE  —

Cytopathology

Case 2 - Metastatic Hepatocellular Carcinoma from an Occult Primary

Charles D. Sturgis
Evanston Northwestern Healthcare
Northwestern University Feinberg School of Medicine
Chicago, IL





Click on slide thumbnail images for an enlarged view.

If you have any difficulties viewing these slides, email the webmaster.



Clinical History:
The patient was a 77 year-old female with no prior history of malignancy. She presented with neck pain and new onset right arm paralysis. MRI of the cervical spine revealed an abnormal diminished signal intensity in the right pedicle and lamina of C6 with a noted 4 x 4 x 3.6 cm soft tissue mass. The mass showed epidural extension into the spinal canal and neural foramen. Digital photomicrographs were taken from microscope slides prepared following image-guided biopsy of the spinal mass.


Case 2 - Figure 1
Diff-Quik, 20X
Case 2 - Figure 2
Papanicolaou stain, 40X
Case 2 - Figure 3
Diff-Quik stain, 40X

Case 2 - Figure 4
Papanicolaou stain, 60X
Case 2 - Figure 5
Cell block, 60X


Cytologic Diagnosis:
Metastatic carcinoma, consistent with hepatocellular primary

Cytologic Diagnosis of Hepatic primary:
Hepatocellular carcinoma

Cytologic Findings:
Five digital images were provided for review. Image one (20X Diff Quik) demonstrates a hemodiluted background without necrosis or diathesis. A flat, cohesive group of lesional cells is noted in the center of the field. The cells are round to polygonal with moderately abundant granular cytoplasm and indistinct cytoplasmic borders at this magnification. In image two (40X Papanicolaou) a three dimensional aggregate of neoplastic cells is seen. The cells have abundant coarsely granular eosinophilic cytoplasm and somewhat better defined cell borders than in the low power Diff Quik image. Nucleoli are visible. Small flattened cells with tapered nuclei are noted at the peripheral edges of the cell cords. The fragment is at least six cells in thickness from side to side. The digital image numbered as three (40X Diff Quik) emphasizes the abundance of cytoplasm in some of the lesional cells with focal dark blue-black globular pigment material as well as focal coarse eosinophilic cytoplasmic globules. Binucleation is present. In the high power image number four (60X Papanicolaou), the polygonal shapes of the lesional cells becomes more obvious. Moderately abundant coarsely granular cytoplasm and nucleoli are key features of this small cohesive lesional cell group. The block from this case was paucicellular. Image five (60X Hematoxylin and Eosin) shows erythrocytes and a small aggregate of lesional cells, one with an intranuclear cytoplasmic invagination. Subsequent immunohistochemical studies performed on the cell block revealed the lesional cells to be immunoreactive (positive) for pan-cytokeratin, cytokeratin 7, and the hepatocyte antigen HepPar 1 / HSA. The lesional cell groups were nonreactive (negative) for cytokeratin 20, estrogen receptor, and thyroid transcription factor-1.

Cytologic and Clinical Follow-up:
Following the image-guided cervical spine needle aspiration, the patient underwent additional imaging and additional laboratory testing. Her serum alpha-fetoprotein (AFP) was measured at 17.1 ng/ml (upper limit of normal range 8 ng/ml). Other tumor markers including CEA were within normal limits. Computerized tomographic studies of the chest, abdomen and pelvis demonstrated a 1 cm partially calcified incidental right thyroid nodule, a 5 cm heterogeneously enhancing mass within the right hepatic lobe, multiple bilateral 2 to 4 mm lung nodules and colonic diverticulosis. The thyroid nodule was needle aspirated under ultrasound guidance resulting in a diagnosis of "benign thyroid nodule." The liver mass was also needle aspirated under ultrasound guidance yielding 13 direct smears and a cell block. The liver lesion was diagnosed as "hepatocellular carcinoma" based on morphology alone.

Discussion:
Cases of metastatic adenocarcinoma arising from clinically occult primary sites are seen with some frequency by practicing cytopathologists. Metastatic deposits from occult primary solid tumors are encountered in the context of needle aspiration biopsies of lymph nodes, lung, liver, adrenal, bone, soft tissue, skin, brain, and other body sites. As expected, epidemiologically "common" occult primary tumors are those that most typically lead to metastatic disease. Adenocarcinomas are the most usual malignancies to give rise to metastatic deposits from unknown origins, and the most frequent primary sites (when discoverable) are lung, large bowel and the pancreaticobiliary tract. [1, 2, 3] In postmenopausal females, breast and gynecologic tract primary tumors should be added to the list. In males, the prostate gland may be considered as a common primary site. Even with highly cellular samples and quality immunocytochemical analyses, there are instances in which a primary source simply cannot be accurately pinpointed. Inability to specifically label a metastatic focus as being derived from a certain site is most common in poorly differentiated / high grade neoplasms and in those cases in which the burden of metastatic disease is widespread, involving multiple anatomic sites.

Although epidemiologically common adenocarcinomas are the tumors most frequently encountered in the work up of metastases from unknown primaries, rarities make for good stories. A veritable treasure trove of unexpected and occult metastatic malignant neoplasms has been reported popping up in the bones of the cervical spine (as in the hepatocellular carcinoma in the current case). This list includes but is not limited to such infrequently encountered entities as pituitary carcinoma, glioblastoma multiforme, malignant meningioma, chordoma, renal cell carcinoma, sebaceous carcinoma, carcinoid tumor of the testis, and paraganglioma. [4, 5, 6, 7, 8, 9, 10, 11] Occult thyroid carcinomas presenting as cervical spine disease also exist. [12, 13] When a history of primary malignancy at another body site is not provided, the cytologist starts at the beginning with light microscopic evaluation. Time invested with a critical "reading of the slides" can pay the largest dividends.

Diagnosing a hepatocellular carcinoma presenting as a cervical spine mass is akin to standing behind a one-way glass and picking a Hobbit out of a lineup of felon cytopathologists. If one doesn't think to put Hobbits in the differential diagnosis, then the wrong "criminal" may go free. Critically interpreting the slides and keeping an open mind makes all of the difference in fingering the correct culprit. The cytomorphologic features of metastatic hepatocellular carcinoma are essentially the same as those of primary hepatoma. Classical architectural / low power features of hepatocellular carcinoma include complex branching and anastomosing thick trabeculae with sharp borders and endothelial appliques. Traversing capillary-sized blood vessels are often noted, and acinar formations are possible. Cellular / high power features include morphologic similarities to non-neoplastic hepatocytes such as polygonal-shaped cells with distinct cell borders and dense granular cytoplasm. Central round nuclei with prominent nucleoli are typical. Low grade lesions and their metastases are typically monotonous with increased nuclear /cytoplasmic ratios. Binucleation and intranuclear inclusions are common. Bile pigment may be seen, as can intracytoplasmic hyaline globules. [14, 15, 16]

Of note, not all primary and secondary hepatoid epithelial malignancies are truly derived from hepatocytes, and endothelial investiture surrounding trabeculae of malignant cells is not entirely specific for hepatocellular carcinoma. Hepatoid morphologies have been reported in (among others) cases of pancreatic adenocarcinoma, gastric adenocarcinoma, ovarian adenocarcinoma, adenocarcinoma of the lung, yolk sac tumor and immature teratoma. [17, 18, 19, 20, 21, 22, 23, 24, 25] The floating island pattern of neoplastic epithelial cells surrounded by appliques of intimately adherent flattened endothelial cells can be seen in adrenal cortical carcinomas and occasionally in other highly vascular tumors such as renal cell carcinoma and tumors derived from neuroendocrine tissues. In a sense, a critical review of the cytomorphology of a metastatic carcinoma may lead one to the diagnosis of a hepatoid malignancy, but ancillary testing will likely be pursued to confirm the cytologic impression whenever possible. In addition to correlating morphology with immunocytochemistry and imaging studies, measurement and tracking of certain serum tumor markers may be of value in discerning primary sites. Serum alpha-fetoprotein, an oncofetal antigen, is frequently used for screening patients suspected of harboring hepatocellular carcinoma in high risk populations. Although this serum marker is elevated in approximately 70% of patients with hepatoma, it is not entirely specific and can be elaborated by germ cell neoplasms, solid tumors of the lung and endometrium and by hepatoblastoma. [22, 26, 27] In addition, mildly to moderately elevated serum levels of alpha-fetoprotein have been reported in hepatic inflammatory conditions. [28]

Multiple cases of hepatocellular carcinoma either presenting with or rapidly developing metastatic disease to the spine, neck, head and upper thorax are known in the world's literature. Published cases of hepatoma presenting with vertebral, cranial, parotid gland and chest wall metastases highlight myriad possible anatomic starting points for the cytologic work up of hepatoma. [29, 30, 31, 32] A retrospective Taiwanese series of 395 individuals with pathologically verified hepatocellular carcinoma demonstrated that 20 (5%) of the patients had bone metastases at initial presentation and that in this subset of patients the most common initial presentation was with a spinal lesion with neurologic compression. [33] Acutely symptomatic epidural hematoma arising in association with metastatic disease has also been reported as the presenting symptom of an otherwise unknown hepatoma. [34]

As hepatocellular carcinoma can present with variable symptoms and as non-hepatocellular carcinomas may appear hepatoid by light microscopy, ancillary testing of tumor tissue is generally undertaken when cellularity allows. Ancillary testing of hepatocellular carcinomas can be of great importance not only in metastatic foci but also in the work up of liver masses proper, as the liver is a common site for metastatic tumors from neoplasms that are primary to other organs and as metastatic disease to the liver is far more common that primary liver carcinoma in North America. [35] In the future, quantitative polymerase chain reaction methodologies and serial analysis of gene expression based on classification maps may allow pathologists to identify malignancies of unknown origin without reliance upon the art of cytomorphologic interpretation. [36] Until that time, we in diagnostic anatomic pathology will likely continue to rely upon algorithms centered around glass slide based immunohistochemical and immunocytochemical testing. Published algorithms for the workup of metastatic adenocarcinoma of unknown primary are helpful; however, these treatises typically focus on "standard" or "frequently encountered" primaries such as a panel including CA125, CDX2, cytokeratins 7 and 20, estrogen receptor, gross cystic disease fluid protein 15, lysozyme, mesothelin, prostate specific antigen, and thyroid transcription factor 1 that is promoted for separating breast, colon, lung, ovary, pancreas, prostate, and stomach cancers. [37] OCT4 is a relatively new marker that is reported to be sensitive and specific for metastatic germ cell tumors and might be of value in ruling out one of these lesions with a hepatoid phenotype. [38] Pertinent positive markers for hepatocellular carcinoma include pancytokeratin and cytokeratin Cam5.2. Many hepatocellular carcinomas express cytokeratins 8 and 18, as is the case with benign hepatocytes. Approximately 90% of hepatomas are immunoreactive with HepPar1, while approximately 40% are immunoreactive for alpha-fetoprotein. Alpha-1-antitrypsin positivity is noted in 50% of hepatocellular carcinomas. The majority of hepatomas are nonreactive with antibodies to cytokeratins 7 and 20; however, the pattern of 7 and 20 expression is highly variable. Monoclonal CEA is typically negative in hepatocellular carcinomas; however, polyclonal CEA antibodies are known to decorate these lesions with a canalicular pattern of staining roughly 70% of the time. Markers such as CD34 may prove useful in highlighting wrapping endothelial cells. It is important to note that IHC results must be interpreted in combination with the overall morphology and with thought given to the larger context of the clinical setting. [39, 40]

For cytodiagnosticians, the challenge of finding an occult primary site based on studies of cells harvested from a metastatic deposit is akin to solving a complicated puzzle based on synthesis of data points taken from combined clinical, radiographic, cytomorphologic, cytochemical, and immunocytochemical correlations. It should be emphasized that behind each one of these challenging puzzles sits a patient. In most instances, the answer to the puzzle will determine the specific patient's treatment. Noticeably varied interventions result from the wording of cytopathology reports. The cytopathology report may trigger initiation of systemic anticancer therapies, administration of targeted radiation, surgery, radio frequency ablation or pain management and hospice referral depending upon the label given to the metastatic deposit and the likely primary source. Empirical chemotherapy with platinum-containing regimens has been shown to benefit some patients who present with metastatic disease from unknown primary sites. [3] If given choices, certainly all patients would prefer specific therapies targeting specific diseases over "blanket" approaches to bad situations. One retrospective study of patients with cervical spine metastases from tumors of unknown primary source showed that detection of the primary tumor was the only factor that significantly influenced patient survival. [41] Looking carefully at every brush stroke on the cytopathology slides allows us to envision the most complete pictures, and seeing the big picture may make all of the difference for a given patient.

References:
  1. Al-Brahim N, Ross C, Carter B, Chorneyko K. The value of postmortem examination in cases of metastasis of unknown origin, 20-year retrospective data from a tertiary center. Ann Diagn Pathol.2005;9(2):77-80.

  2. Blaszyk H, Hartmann A, Bjornsson J. Cancer of unknown primary: clinicopathologic correlations. APMIS.2003;111(12):1089-1094.

  3. Yakushiji S, Ando M, Yonemori K, Kohno T, Shimizu C, Katsumata N, Fujiwara Y. Cancer of unknown primary site: review of consecutive cases at the National Cancer Center Hospital of Japan. Int J Clin Oncol.2006;11(6):421-425.

  4. Ceyhan K, Yagmurlu B, Dogan BE, Erdogan N, Bulut S, Erekul S. Cytopathologic features of pituitary carcinoma with cervical vertebral bone metastasis: a case report. Acta Cytol.2006;50(2):225-230.

  5. Utsuki S, Tanaka S, Oka H, Iwamoto K, Sagiuchi T, Fujii K. Glioblastoma multiforme metastasis to the axis. case report. J Neruosurg.2005;102(3):540- 542.

  6. Pinsker MO, Buhl R, Hugo HH, Mehdorn HM. Metastatic meningioma WHO grade II of the cervical spine: case report and review of the literature. Zentralbl Neurochir.2005:66(1):35-38.

  7. Arnautovic KI, Al-Mefty O. Surgical seeding of chordomas. J Neurosurg. 2001;95(5):798-803.

  8. Schijns OE, Kurt E, Wessels P, Luijckx GJ, Beuls EA. Intramedullary spinal cord metastasis as a first manifestation of a renal cell carcinoma: report of a case and review of the literature. Clin Neurol Neurosurg. 200;102(4):249-254.

  9. Antuna SA, Mendez JG, Cincunegui JA, Lopez-Fanjul JC. Metastatic lesion of the cervical spine secondary to an extraocular sebaceous carcinoma. Acta Orthop Belg. 1996;62(4):229-232.

  10. Shimura S, Uchida T, Shitara T, Nishimura K, Murayama M, Honda N, Koshiba K. Primary carcinoid tumor of the testis with metastasis to the upper vertebrae report of a case. Nippon Hinyokika Gakkai Zasshi. 1991:82(7):1157-1160.

  11. North CA, Zinreich ES, Chistensen WN, North RB. Multiple spinal metastases from paraganglioma. Cancer. 1990;66(10):2224-2228.

  12. Habra MA, Vassilopoulou-Sellin R. Cervical spine metastasis mimicking thyroid bed uptake in follicular thyroid carcinoma. Thyroid. 2005:15(3):298-299.

  13. Kupferman ME, Patterson M, Mandel SJ, Livolsi V, Weber RS. Patterns of lateral neck metastasis in papillary thyroid carcinoma. Arch Otolaryngol Head Neck Surg. 2004;130(7):857-860.

  14. Pisharodi LR, Lavoie R, Bedrossian CW. Differential diagnostic dilemmas in malignant fine needle aspirates of liver: a practical approach to final diagnosis. Diagn Cytopathol. 1995;12(4):364-371.

  15. Demay R, Liver in The Art and Science of Cytopathology. ASCP Press. Chicago. 1996;1017-1052.

  16. Geisinger KR, Stanley MW, Raab SS, Silverman JF, Abati A. Liver in Modern Cytopathology. Churchill Livingstone. Philadelphia. 2004; 505-542.

  17. Matsueda K, Yamamoto H, Yoshida Y, Notohara K. Hepatoid carcinoma of the pancreas producing protein induced by vitamin K absence or antagonist II (PIVKA-II) and alpha-fetoprotein (AFP). J Gastroenterol. 2006;41(10):1011-9.

  18. Caruso RA, Basile F, Fedele F, Zuccala V, Crisafulli C, Fracassi MG, Quattrocchi E, Venuti A, Fabiano V. Gastric hepatoid adenocarcinoma with autophagy- related necrosis-like tumor cell death: report of a case. Ultrastruct Pathol. 2006;30(4):301-307

  19. Kwon JE, Kim SH, Cho NH. No ancillary finding is valid to distinguish a primary ovarian hepatoid carcinoma from metastatic hepatocellular carcinoma. Int J Gynecol Cancer. 2006;16(4):1691-1694.

  20. Terracciano LM, Glatz K, Mhawech P, Vasei M, Lehmann FS, Vecchione R, Tornillo L. Hepatoid adenocarcinoma with liver metastasis mimicking hepatocellular carcinoma: an immunohistochemical and molecular study of eight cases. Am J Surg Pathol. 2003;27(10):1302-1312.

  21. Hayashi Y, Takanashi Y, Ohsawa H, Ishii H, Nakatani Y. Hepatoid adenocarcinoma in the lung. Lung Cancer. 2002;38(2):211-214.

  22. Hiroshima K, Iyoda A, Toyozaki T, Haga Y, Baba M, Fujisawa T, Ishikura H, Ohwada H. Alpha-fetoprotein-producing lung carcinoma: report of three cases. Pathol Int. 2002;52(1):46-53.

  23. Gopaldas R, Kunasani R, Plymyer MR, Bloch RS. Hepatoid malignancy of unknown origin – a diagnostic conundrum: review of literature and case report of collision with adenocarcinoma. Surg Oncol. 2005;14(1):11-25.

  24. Matsuyama TA, Kushima M, Yamochi-Onizuka T, Ota H. Placental yolk sac tumor with divergent endodermal differentiation. Int J Gynecol Pathol. 2004;23(4):398-402.

  25. Taniyama K, Ohta S, Suzuki H, Matsumoto M, Nagashima Y, Tahara E. Alpha- fetoprotein-producing immature mediastinal teratoma showing rapid and massive recurrent growth in an adult. Acta Pathol Jpn. 1992;42(12):911-915.

  26. Ishak KG, Goodman ZD, Stocker JT. Hepatocellular carcinoma in atlas of tumor pathology tumors of the liver and intrahepatic bile ducts. Third Series Fascicle 31. AFIP Press. Washington D. C. 1999;199-230.

  27. Takeuchi K, Kitazawa S, Hamanishi S, Inagaki M, Murata K. A case of alpha- fetoprotein producing adenocarcinoma of the endometrium with a hepatoid component as a potential source for alpha-fetoprotein in a postmenopausal woman. Int J Gynecol Cancer. 2006;16(3):1442-1445.

  28. Amanjee S, Kew MC, Van Staden L. The pathogenesis of raised serum alpha- fetoprotein levels in amoebic and pyogenic hepatic abscesses. Trop Gastroenterol. 1990;11(2):99-201.

  29. Lauzon A, Lefebvre R. Primary epithelioma of the liver revealed by a vertebral metastasis and invasion of the adjacent tissues. Union Med Can. 1973;102(2):383-386.

  30. Lee JP. Hepatoma presenting as craniospinal metastasis: analysis of sixteen cases. J Neurol Neurosurg Psychiatry. 1992;55(11):1037-1039.

  31. Romanas MM, Cherian R, McGregor D, Wu Y, May CL, Baranda JC. Hepatocellular carcinoma diagnosed by fine-needle aspiration of the parotid gland. Diagn Cytopathology. 2004:30(6):401-405.

  32. Qureshi SS, Shrikhande SV, Borges AM, Shukla PJ. Chest Wall metastases from unknown primary hepatocellular carcinoma. J Postgrad Med. 2005;51(1):41-42

  33. Liaw CC, Ng KT, Chen TJ, Liaw YF. Hepatocellular carcinoma presenting as bone metastasis. Cancer. 1989;64(8):1753-1757.

  34. McIver JI, Scheithauer BW, Rydberg CH, Atkinson JL. Metastatic hepatocellular carcinoma presenting as epidural hematoma: case report. Neurosurgery. 2001;49(2):447-449.

  35. Centeno BA. Pathology of liver metastases. Cancer Control. 2006;13(1):13-26.

  36. Buckhaults P, Zhang Z, Chen YC, Wang TL, St Croix B, Saha S, Bardelli A, Morin PJ, Polyak K, Hruban RH, Velculescu VE, Shih I. Identifying tumor origin using a gene expression-based classification map. Cancer Res. 2003;63(14):4144-4149.

  37. Dennis JL, Hvidsten TR, Wit EC, Komorowski J, Bell AK, Downie I, Mooney J, Verbeke C, Bellamy C, Keith WN, Oien KA. Markers of adenocarcinoma characteristic of the site of origin: development of a diagnostic algorithm. Clin Cancer Res. 2005;11(10):3766-3772.

  38. Lau SK, Chang KL. OCT4: a sensitive and specific immunohistochemical marker for metastatic germ cell tumors. Adv Anat Pathol. 2006;13(2):76-79.

  39. Dabbs DJ. Immunohistochemistry of the gastrointestinal tract, pancreas, bile duct, gallbladder, and liver in diagnostic immunohistochemistry. Churchill Livingstone. Philadelphia. 2002;333-406.

  40. Wee A. Diagnostic utility of immunohistochemistry in hepatocellular carcinoma, its variants and their mimics. Appl Immunohistochem Mol Morphol. 2006;14(3):266-272.

  41. Rodriguez Gutierrez A, Calvo Boizas E, Alonso Rodriguez O, Soria Carreras P, Solbes Solbes R, Soler Ripoll JJ, Gomez Gonzalez JL, Fonseca Sanchez E. Results of treating cervical metastases of an unknown primary tumor a report of 29 cases. Acta Otorrinolaringol Esp. 1999;50(4):305-309.