Introduction
Metastatic tumours are the most common neuro-oncological lesions developing in the central nervous system (CNS). The metastatic tumours may herald the dissemination of the malignancy already being diagnosed or treated. In some cases (2-18% [3,22,32]), they may present as a cancer of unknown primary (CUP) and in these instances they require careful determination of the source of origin of the tumour. From the medical and economic points of view the histopathological determination of the phenotypic and genotypic profiles seems to be the most prudent. It enables an easy, relatively rapid and cheap method of tumour analysis [11]. Besides the detailed clinical evaluation of the patient, imaging (ultrasonography, endoscopy, computed tomography, positron emission tomography, magnetic resonance imaging, etc.) and laboratory procedures (serum tumour markers) are necessary to identify the primary tumour site. However, according to various reports, these analyses succeed in only approximately 20% of cases and their average cost is around $18,000 per patient. On the other hand, immunohistochemical profiling of the metastatic tumour requires an average of $2,000 per patient with a success rate of around 70% [11,33]. Therefore, cost-effectiveness analysis and enlarging specificity of immunohistochemical markers makes the pathological examination an important adjunct to the metastatic work-up.
The central nervous system is most frequently
a target of metastatic dissemination from lung (18-60%) and breast carcinomas (5-21%), melanoma (4-16%), genitourinary (3-10%) and gastrointestinal (5-12%) malignancies [2,3,6,14]. Those cancers arise in the CNS following haematogenous dissemination in most cases, as the lymphatic vasculature in the brain is non-existent [23]. Brain or spinal cord may also be involved by contiguous infiltration of the malignant tumour from the surrounding tissues. This may occur in the vertebral metastases of the prostatic carcinomas that may invade the structure of the spinal meninges.
Depending on the site of origin and the biological properties of the malignant process, the involvement of the CNS may manifest either as a solid, well-circumscribed tumour(s) in the brain or spinal cord parenchyma, dura-based lesions or a not well localized process in the arachnoid space. Proper identification of the secondary involvement of the CNS is vital for the optimal treatment modalities in a patient, as some of the metastatic lesions may mimic primary brain neoplasms. Precise phenotyping and, in some cases, genotyping of the tumour are necessary to reach an unequivocal diagnosis and these issues are the main focus of the current review. In some cases, the identification of the histological diagnosis of the metastasis may provide the patient with beneficial systemic treatment, as occurs in cases of pulmonary small cell carcinoma or lymphomas.
Epidemiology
The true incidence of metastases to the CNS is unknown but it is believed to be underestimated [32]. Involvement of the CNS may occur as a solitary lesion or multiple metastases [29]. The anatomical location of metastatic tumours is dependent on the vascularization of the CNS, biological propensities of the metastatic deposits and, to a certain degree, the primary site of origin. As the prevailing pathway of spread is via the bloodstream, most of the metastases are located in the brain hemispheres (80%), especially in the parietal lobe, followed by the frontal and occipital lobes [3]. The cerebellum and the brainstem are affected less frequently, whereas leptomeningeal and dura-based metastases occur quite uncommonly [13,28]. Leptomeningeal involvement of the CNS usually occurs in patients with pulmonary and mammary carcinomas [28]. Dural metastases are the consequence of the direct spread of the tumour from the bone (skull, vertebra), with lung, prostate and breast carcinomas predominating [13,20].
Generally, metastases from the lung and breast carcinomas and melanoma most commonly present in the CNS; other malignancies (gastric, female reproductive tract, pancreatic, colorectal) are identified less frequently in the CNS and usually late in the course of the disease.
Histopathological identification of the original site of metastasis
Determination of the metastatic nature of the brain tumour relies on the typical morphological features of the lesion. In recent years, determination of the tumour immunophenotype has added much information in regard to the potential source of origin. As epithelial tumours make most of the secondary CNS deposits, cytokeratin profile is indispensable in determining the site of origin. Cytokeratin 7 (CK7) and cytokeratin 20 (CK20) are basic parameters in this regard. Additional information may be acquired with specific markers, e.g. thyroglobulin (thyroid follicular epithelium), calcitonin (thyroid C cells), prostate specific antigen – PSA (prostate gland), renal cell carcinoma marker – RCCMa (kidney), etc.
Metastatic pulmonary carcinoma
Among non-small lung carcinomas, adenocarcinoma is most commonly identified in the CNS metastases. These lesions usually are CK7+/CK20- and may present expression of low molecular cytokeratins (CAM5.2) [3]. In addition, thyroid transcription factor (TTF-1) is frequently identified in the tumour nuclei of pulmonary adenocarcinomas (Fig. 1A) [34]. This marker is a member of the Nkx homeodomain-containing transcription factor family and is expressed during embryogenesis of the thyroid gland, lungs and forebrain [21]. Apart from adenocarcinomas, pulmonary large cell carcinomas, small cell carcinomas and neuroendocrine carcinomas may also show TTF-1 positivity [21,26]. On the other hand, squamous cell carcinomas of the lung usually do not show expression of that marker [26]. Small cell carcinomas (SCCs) of the lung, as belonging to the spectrum of neuroendocrine cell tumours, may show expression of neuroendocrine markers (chromogranin A, synaptophysin, neural cell adhesion molecule – NCAM / CD56, etc.) in addition to the CK7-/CK20- profile [3]. As small cell neuroendocrine carcinomas usually are morphologically indistinguishable, it is prudent to distinguish pulmonary SCC from others, and CK20 and TTF-1 expression may be a helpful panel in this regard as it discriminates cutaneous neuroendocrine carcinoma (Merkel cell carcinoma) from its pulmonary counterpart [9].
Metastatic mammary carcinoma
Breast carcinoma metastasizing to the CNS may be of any histological type but usually it is positive with CK7 and negative with CK20 [3]. Oestrogen (Fig. 1B) and progesterone receptors are often positive [3], although their expression is not restricted to that site of origin. There are also pulmonary carcinomas and other less commonly metastasizing carcinomas involving CNS that may present these two steroid receptors (endometrial, ovarian, etc.) [18]. Gross cystic disease fluid protein-15 (GCDFP-15) is also commonly expressed in mammary carcinomas but not unequivocally specific [19]. In contrast to pulmonary adenocarcinomas, TTF-1 is not typical for breast-derived metastases [3].
Metastatic colorectal carcinoma
Most metastatic tumours are adenocarcinomas and they usually retain morphological features of the primary lesion. In favour of colorectal derivation, the presence of ‘dirty’ necrosis and ‘garland-like’ growth pattern are frequently identified [18]. In addition, CK7-/CK20+ cytokeratin profile, expression of carcinoembryonic antigen (CEA) and an intestine-specific homeodomain transcription factor CDX2 (Fig. 1C) help in confirming this anatomical derivation of the tumour [3].
Metastatic ovarian carcinoma
Rarely, ovarian cancer may involve the CNS. The most common ovarian primary is serous carcinoma, which morphologically may mimic primary papillary tumours of the brain, e.g. choroid plexus carcinoma. While showing expression of CK7+/CK20- phenotype, serous carcinoma frequently strongly expresses WT-1, which enables differential diagnosis [18]. Additionally, BerEP4 expression may be a useful adjunct to the diagnosis as it is usually present in ovarian carcinomas [1], while only rarely being identified in choroid plexus tumours [12].
Metastatic gastric carcinoma
Metastases of gastric carcinoma may occur in the CNS. They present with CK7+/CK20- profile and in addition may show expression of CEA [3]. As this phenotype (CK7+/CK20-/CEA+) is also shared by some other carcinomas (pancreatic, colon, lung) close morphological and clinical correlations are inevitable.
Metastatic renal carcinoma
Most primary renal cancers present as the clear cell carcinoma subtype. Due to the specific morphology it may closely mimic some of the CNS primary tumours, e.g. haemangioblastoma. This is sometimes an important issue clinically, especially in patients with von Hippel-Lindau syndrome. This familial tumour syndrome dependent on germline (inborn) mutations of the VHL gene [15,27] may present both with haemangioblastoma (sometimes multicentric) and renal cell carcinoma [30]. Morphologically, haemangioblastoma is composed of stromal and vascular cells. The former are arranged in alveolar structures and have abundant clear cytoplasm, which closely mimic the structure of clear cell carcinoma of the kidney. Nuclear pleomorphism of the stromal cells present in some cases may additionally confer atypical features to these cells, making problematic an unequivocal differentiation between the two on morphological grounds alone. Haemangioblastoma is a tumour of uncertain histogenesis, but it quite constantly shows expression of NCAM (CD56) and S100 protein, and lack of EMA and cytokeratins. In contrast, renal cell carcinoma shows expression of RCCMa, EMA and CD10, while being CK7-negative [7].
Metastatic melanoma
Known for its enormous morphological plasticity, malignant melanoma always needs to be taken into account in the histopathological differentiation of tumours metastasizing to the CNS. It may show epithelioid, spindle and small cell anaplastic morphology, but phenotypically may be discriminated by expression of melanosome-bound proteins. Most commonly used are HMB45, Melan-A (MART-1) and, although less specific but more sensitive, especially in spindle cell melanomas, S100 protein. This in combination with lack of cytokeratin expression may allow proper diagnosis. Malignant melanoma of the CNS is usually a metastasis from extraneural sites (skin, anus, eye bulb, etc.) but it should be remembered that it may also develop as a primary CNS lesion. Most commonly such tumours arise as meningeal-based lesions [5].
Lymphomas
These tumours may affect the CNS as a single tumour or multifocal lesions usually in supratentorial location. Bilateral symmetrical subependymal lesions are highly suggestive of CNS lymphoma. Sometimes, these lesions may be discreet in imaging procedures and may mimic small infarctions (intravascular large B-cell lymphoma). The lymphoma infiltrates of the CNS are sensitive to steroids and may vanish within hours after such treatment [16].
The abundance of nosological entities among lymphoid neoplasms makes a brief review of their histological differentiation difficult. However, in the CNS there is a predominance of a certain set of lymphomas that preferentially affect this system. Most commonly, high-grade B-cell lymphomas involve the brain and the spinal cord, with diffuse large B-cell lymphoma being the most frequent one [16,24]. The common leukocyte antigen CD45 (LCA) is shared by most lymphomas. Pan-B (CD20, CD79a) or pan-T (CD3, CD4, CD8, CD2) markers are usually helpful in determination of the exact lineage of the proliferation, and in combination with the morphology of the tumour provides a strong basis for precise classification. Extremely differentiated plasma cells lack expression of pan-B markers and show presence of other specific antigens, i.e., CD138, VS38c (Fig. 2B). In some cases, use of steroids decreasing intracranial pressure in the period preceding diagnostic biopsy may alter the structure of the tumour [25]. Increase in apoptotic bodies, macrophages and necrosis are typical manifestations of such treatment.
Anaplastic tumours
Poorly differentiated tumours may involve the CNS as one of the potential targets in their dissemination throughout the body. They require differentiation from primary anaplastic tumours of the CNS, e.g. PNET, glioblastoma, germ cell tumours, etc. Thorough clinical metastatic work-up and immunohistochemical analysis (sometimes molecular genetics) can resolve this issue in most cases. Poorly differentiated sarcomas should be taken into account. Ewing sarcoma presents typical translocation t(11;22) that may be identified by molecular techniques (PCR, FISH). This tumour usually expresses CD99 (MIC-2), which is quite typical for this lesion, although it may also be shared by T-cell lymphoblastic lymphomas and poorly differentiated synovial sarcoma [10]. The latter may be differentiated by coexpression of epithelial markers (epithelial membrane antigen, EMA and cytokeratins) and non-random translocation t(X;18), which may also be identified by molecular techni-
ques. Another small round blue cell tumour that should be taken into account is rhabdomyosarcoma. Its anaplastic cells extensively present vimentin, and only differentiating rhabdomyoblasts may show expression of desmin and actin. Very specific for rhabdomyosarcomas are myoD1 and myogenin (myf-4) [8]; however, only nuclear reactivity should be regarded as a true marker of myogenic differentiation [31].
The CNS is frequently affected by metastases of choriocarcinoma; these are usually haemorrhagic tumours presenting with an apoplectic onset of disease. In the blood clots, one can identify typical morphological structure of syncytiotrophoblasts and cytotrophoblasts. The former expresses beta-chorionic gonadotrophin (ß-hCG) that may be identified in the tissue or, alternatively, in the blood. Rarely, other malignancies may spread into the CNS, e.g. sarcomas (malignant peripheral nerve sheath tumour, gastrointestinal stromal tumour [Fig. 2B]), carcinoid tumours, etc. [4;17].
Summary
Metastatic lesions in the CNS are very frequently identified in everyday neurosurgical and neuropathological practice. Therefore, acquaintance with and use of the immunohistochemical markers are essential for the correct histopathological diagnosis and the optimal treatment of patients with CNS malignancies.
References:
1. Attanoos RL, Webb R, Dojcinov SP, Gibbs AR. Value of mesothelial and epithelial antibodies in distinguishing diffuse peritoneal mesothelioma in females from serous papillary carcinoma of the ovary and peritoneum. Histopathology 2002; 40: 237-244.
2. Barnholtz-Sloan JS, Sloan AE, Davis FG, Vigneau FD, Lai P, Sawaya RE. Incidence proportions of brain metastases in patients diagnosed (1973 to 2001) in the Metropolitan Detroit Cancer Surveillance System. J Clin Oncol 2004; 22: 2865-2872.
3. Becher MW, Abel TW, Thompson RC, Weaver KD, Davis LE. Immunohistochemical analysis of metastatic neoplasms of the central nervous system. J Neuropathol Exp Neurol 2006; 65: 935-944.
4. Borowska-Lehman J, Rzepko R, Chrostowski L, Jende P, Gorska-Dubowik M. Metastatic carcinoid tumor in the brain. Folia Neuropathol 1994; 32: 265-268.
5. Brat DJ, Giannini C, Scheithauer BW, Burger PC. Primary melanocytic neoplasms of the central nervous system. Am J Surg Pathol 1999; 23: 745-754.
6. Burger PC, Scheithauer BW, Vogel FS. Surgical Pathology of the Nervous System and Its Coverings. Churchill Livingstone, New York; 2002.
7. Cameron RI, Ashe P, O'Rourke DM, Foster H, McCluggage WG.
A panel of immunohistochemical stains assists in the distinction between ovarian and renal clear cell carcinoma. Int J Gynecol Pathol 2003; 3: 272-276.
8. Cessna MH, Zhou H, Perkins SL, Tripp SR, Layfield L, Daines C, Coffin CM. Are myogenin and myoD1 expression specific for rhabdomyosarcoma? A study of 150 cases, with emphasis on spindle cell mimics. Am J Surg Pathol 2001; 25: 1150-1157.
9. Cheuk W, Kwan MY, Suster S, Chan JK. Immunostaining for thyroid transcription factor 1 and cytokeratin 20 aids the distinction of small cell carcinoma from Merkel cell carcinoma, but not pulmonary from extrapulmonary small cell carcinomas. Arch Pathol Lab Med 2001; 125: 228-231.
10. Dei Tos AP, Wadden C, Calonje E, Sciot R, Pauwels P, Knight JC, Dal Cin P, Fletcher CD. Immunohistochemical demonstration of glycoprotein p30/32mic2 (CD99) in synovial sarcoma. A potential cause of diagnostic confusion. Appl Immunohistochem 1995; 3: 168-173.
11. Elsheikh TM, Silverman JF. Differential diagnosis of metastatic tumors. In: Silverberg SG, DeLellis RA, Frable WJ, LiVolsi VA, Wick MR (eds.). Silverberg's Principles and Practice of Surgical Pathology and Cytopathology. Churchill Livingstone Elsevier, 2006; pp. 167-192.
12. Gottschalk J, Jautzke G, Paulus W, Goebel S, Cervos N. The use of immunomorphology to differentiate choroid plexus tumors from metastatic carcinomas. Cancer 1993; 72: 1343-1349.
13. Laigle-Donadey F, Taillibert S, Mokhtari K, Hildebrand J, Delattre JY. Dural metastases. J Neurooncol 2005; 75: 57-61.
14. Lassman AB, DeAngelis LM. Brain metastases. Neurol Clin 2003; 21: 1-23.
15. Latif F, Tory K, Gnarra J, Yao M, Duh FM, Orcutt ML, Stackhouse T,
Kuzmin I, Modi W, Geil L. Identification of the von Hippel-Lindau disease tumor suppressor gene . Science 1993; 260: 1317-1320.
16. Louis DN, Ohgaki H, Wiestler OD, Cavenee WK. WHO Classification of Tumours of the Central Nervous System. IARC, Lyon; 2007.
17. Matyja E, Naganska E, Gorski R, Zabek M. Multiple brain metastases from malignant peripheral nerve sheath tumour (MPNST). Folia Neuropathol 2004; 42: 43-48.
18. McCluggage WG, Wilkinson N. Metastatic neoplasms involving the ovary: a review with an emphasis on morphological and immunohistochemical features. Histopathology 2005; 47: 231-247.
19. Monteagudo C, Merino MJ, LaPorte N, Neumann RD. Value of gross cystic disease fluid protein-15 in distinguishing metastatic breast carcinomas among poorly differentiated neoplasms involving the ovary. Hum Pathol 1991; 22: 368-372.
20. Mut M, Schiff D, Shaffrey ME. Metastasis to nervous system: spinal epidural and intramedullary metastases. J Neurooncol 2005; 75: 43-56.
21. Oliveira AM, Tazelaar HD, Myers JL, Erickson LA, Lloyd RV. Thyroid transcription factor-1 distinguishes metastatic pulmonary from well-differentiated neuroendocrine tumors of other sites. Am J Surg Pathol 2001; 25: 815-819.
22. Pavlidis N, Briasoulis E, Hainsworth J, Greco FA. Diagnostic and therapeutic management of cancer of an unknown primary. Eur J Cancer 2003; 39: 1990-2005.
23. Piccirilli M, Brunetto GM, Rocchi G, Giangaspero F, Salvati M. Extra central nervous system metastases from cerebral glioblastoma multiforme in elderly patients. Clinico-pathological remarks on our series of seven cases and critical review of the literature. Tumori 2008; 94: 40-51.
24. Rudnik A, Larysz D, Blamek S, Larysz P, Bierzynska-Macyszyn G, Wlaszczuk P, Bazowski P. Primary pituitary lymphoma. Folia Neuropathol 2007; 45: 144-148.
25. Schwechheimer K, Braus DF, Schwarzkopf G, Feller AC, Volk B, Muller-Hermelink HK. Polymorphous high-grade B cell lymphoma is the predominant type of spontaneous primary cerebral malignant lymphomas. Histological and immunomorphological evluation of computed tomography-guided stereotactic brain biopsies. Am J Surg Pathol 1994; 18: 931-937.
26. Srodon M, Westra WH. Immunohistochemical staining for thyroid transcription factor-1: a helpful aid in discerning primary site of tumor origin in patients with brain metastases. Hum Pathol 2002; 33: 642-645.
27. Stolle C, Glenn G, Zbar B, Humphrey JS, Choyke P, Walther M, Pack S, Hurley K, Andrey C, Klausner R, Linehan WM. Improved detection of germline mutations in the von Hippel-Lindau disease tumor suppressor gene. Hum Mutat 1998; 12: 417-423.
28. Taillibert S, Laigle-Donadey F, Chodkiewicz C, Sanson M, Hoang-Xuan K, Delattre JY. Leptomeningeal metastases from solid malignancy: a review. J Neurooncol 2005; 75: 85-99.
29. Varadhachary GR, Talantov D, Raber MN, Meng C, Hess KR, Jatkoe T, Lenzi R, Spigel DR, Wang Y, Greco FA, Abbruzzese JL, Hainsworth JD. Molecular profiling of carcinoma of unknown primary and correlation with clinical evaluation. J Clin Oncol 2008; 26: 4442-4448.
30. Vogelstein B, Kinzler KW The Genetic Basis of Human Cancer. McGraw-Hill, New York; 2002.
31. Wang MP, Marx J, McNutt MA, Rutledge JC, Gown MA. Expression of myogenic regulatory proteins (myogenin and MyoD1) in small blue round cell tumors of childhood. Am J Pathol 1995; 147: 1799-1810.
32. Wesseling P, von Deimling A, Aldape KD. Metastatic tumours of the CNS. In: Louis DN, Ohgaki H, Wiestler OD, Cavenee WK (eds.). WHO Classification of Tumours of the Central Nervous System. IARC, Lyon; 2007; pp. 248-251.
33. Wick MR, Ritter JH, Swanson PE. The impact of diagnostic immunohistochemistry on patient outcomes. Clin Lab Med 1999; 19: 797-814.
34. Yatabe Y, Mitsudomi T, Takahashi T. TTF-1 expression in pulmonary adenocarcinomas. Am J Surg Pathol 2002; 26: 767-773.
