The identification of metastatic solid tumors in lymph nodes is one of the most important tasks in diagnostic surgical pathology. When regional lymph nodes are examined as part of a resection specimen, the primary neoplasm is known and the major task is to simply identify the presence of metastases. However, up to 5% of patients with solid tumors initially present with lymphadenopathy as a result of metastasis from an occult primary tumor (1,2)

In many biopsy specimens, routine histologic examination provides clues to the site of origin (e.g., melanin pigment in malignant melanoma). However, poorly differentiated metastatic solid tumors lack these histologic clues. Although these neoplasms can be loosely grouped into categories based on their histologic features, such as epithelioid, anaplastic, spindle cell, or small-cell, the differential diagnosis of each of these categories is broad. Immunophenotypic analysis can add much information, thus increasing the pathologist’s ability to predict the site of origin. Either flow cytometry or immunohistochemistry can be used for this purpose. However, immunohistochemistry is ideally suited to the workup of metastatic solid neoplasms because these metastases often involve lymph nodes partially or focally, thereby making visual recognition of the metastasis and its immunoreactivity required (Fig. 6.1).

Many immunohistochemical stains can be used for identification and classification of metastatic neoplasms, and this number is growing (Table 6.1). In general, the antibodies used can be grouped into initial and secondary panels influenced, in large part, by the histologic features of the neoplasm (e.g. epithelioid, spindle cell, etc.). In general, the commonly used first-tier diagnostic antibodies are highly specific but variably sensitive for detection of particular tumor types. This group includes reagents specific for keratin (usually a cocktail of antikeratin antibodies), epithelial membrane antigen (EMA, also known as MUC1) S100 protein, vascular markers (CD31 and CD34), LCA (CD45RB), germ cell tumor markers (PLAP, OCT4), spindle cell tumor markers (e.g., desmin, smooth muscle actin), and plasma cell–associated markers (CD38, CD138).

In metastatic carcinomas of unknown origin, a second group of antibodies can be used to suggest a possible primary site. Currently, the keratin expression pattern is probably the most widely used, particularly those antibodies specific for keratins 7 and 20 (3). For example, gastrointestinal carcinomas are typically negative for keratin 7 and positive for keratin 20. Expression or lack thereof of other types of keratin is also helpful. Hepatocellular carcinoma is typically negative for keratin 19, in contrast with cholangiocarcinoma. The quality of keratin immunoreactivity also can be helpful. Punctate or dot-like cytoplasmic staining, as is commonly observed in Merkel cell carcinoma and small-cell carcinoma, is suggestive, but not specific for these neoplasms.

Other antibodies are also available that, when reactive, are more or less tumor-specific. Perhaps the best examples of these reagents are prostatic acid phosphatase and prostate-specific antigen. Additional specific antibodies are being developed or validated; RCC for renal cell carcinoma and HepPar for hepatocellular carcinoma are two examples. Other antibodies, although not specific, are suggestive of a limited number of possible primary sites. An example is estrogen receptor positivity which is suggestive of a breast or gynecologic-tract primary site.

Immunophenotypic data are key components of the World Health Organization (WHO) lymphoma classification and necessary for the diagnosis of most specific lymphoma entities (4). Some lymphoma types have a characteristic immunophenotype, and recognition of this immunophenotype in the appropriate clinical and morphologic context strongly supports the diagnosis. Immunophenotype is also essential in the differential diagnosis of morphologically similar entities.

As stated earlier, two major methods are used for determining the immunophenotype of hematologic neoplasms: immunohistochemistry and flow cytometry (see Chapter 7). Flow cytometry immunophenotyping has many advantages, some of which include high sensitivity, simultaneous assessment of multiple antigens in a single cell, and the ability to analyze large numbers of cells rapidly and quantitatively. The major drawback of using flow cytometry is that one cannot directly visualize the architectural features of a biopsy specimen or the cytologic features of the cells being immunophenotyped. Although not an issue when a lymph node biopsy specimen is replaced by a homogeneous population of neoplastic cells, this drawback of flow cytometry can become an issue when the neoplasm represents only a small proportion of all cells in the biopsy specimen, especially when the cytologic features of the neoplastic cells must be correlated with the immunophenotype.

There are a number of specific circumstances in which using immunohistochemistry for immunophenotyping is more advantageous than flow cytometry. As stated earlier, any neoplasm in which direct visual examination of the tumor architecture or the neoplastic cells themselves is helpful is a potential indication
for the use of immunohistochemistry. Other circumstances also arise when flow cytometry analysis is compromised for one reason or another, and immunohistochemistry can be used in its place. The following is a list of circumstances in which immunohistochemistry has advantages over flow cytometry. (Flow cytometry and its advantages are discussed in Chapter 7.) This list is meant to be illustrative, rather than a complete catalog of all possible circumstances.

Many techniques have been developed for immunohistochemical analysis and, in recent years, many steps in the process of immunostaining have become automated. Here, we briefly describe the principles of immunohistochemical analysis. For additional details, the reader is referred to a number of textbooks on immunohistochemistry.

The first step in immunohistochemical analysis is the application of a primary antibody. Polyclonal and monoclonal antibodies are used. Most polyclonal antibodies are derived from rabbits, whereas monoclonal antibodies are derived from mice. However, rabbit monoclonal antibodies have been recently developed that appear to be more sensitive than mouse monoclonal antibodies. If a mouse monoclonal antibody is used, the second step involves an anti-mouse immunoglobulin complexed with a large molecule that will allow amplification of the signal. In the commonly used avidin–biotin complex (ABC) immunoperoxidase method, the anti-mouse immunoglobulin is complexed with biotin. The third step involves adding avidin–biotin–horseradish peroxidase–antiperoxidase complex. The peroxidase then reacts with a chromogen to produce a reaction product that is discretely localized according to the normal site(s) of the cellular antigen being assessed. The reaction product may be localized to the surface, cytoplasm, or nucleus depending on the normal distribution of the antigen in the cell.

Using the ABC system, two chromogens are commonly used: diaminobenzidine (DAB) or 5–amino-9–ethyl carbazole (AEC). We prefer DAB because it yields a brown reaction product that fades little over time. AEC yields a red-brown reaction product that is aesthetically attractive, in the eyes of some more so than DAB, but the reaction product can fade over time as diffusion occurs, particularly if cover-slipped incorrectly.

The alkaline phosphatase-antialkaline phosphatase (APAAP) detection system also is commonly used for immunohistochemistry. Use of APAAP is advantageous because many types of cells have endogenous peroxidase (e.g., eosinophils) that complicate analysis using the ABC approach. Methods for blocking endogenous peroxidase can be used, but these methods can decrease the sensitivity of detecting certain antigens. Some cells have endogenous alkaline phosphatase, but generally in fewer cells than those that carry endogenous peroxidase, and the endogenous alkaline phosphatase is relatively easy to block. The APAAP and ABC methods can be combined for double immunolabeling.

Both frozen and fixed, paraffin-embedded tissue sections can be used for immunohistochemical analysis. To obtain the best results with frozen tissue, it is best to snap-freeze the tissue in isopentane/liquid nitrogen (or dry ice) or acetone/dry ice baths. Tissue frozen in OCT compound in a cryostat also can be used, although ice crystals are more likely to form and freeze-thaw artifact can destroy antigens. The great advantage of using frozen sections is that virtually every antibody used for flow cytometry immunophenotyping can be used for immunohistochemical analysis of frozen sections. Another advantage is that surface immunoglobulin expression by B-cells is not destroyed by the process of freezing tissue, and therefore clonality can be assessed. The disadvantages are mainly two: (a) frozen tissue is often not available, and (b) the architectural and cytologic features within the biopsy specimen are not as clearly seen as can be appreciated in routinely stained tissue sections (Fig. 6.6).