Only a little more than 30 years ago a popular, widely used classification system for non-Hodgkin lymphomas was initiated based solely on hematoxylin and eosin (H&E)–stained slide interpretation. The “Working Formulation for Clinical Usage” was used for 12 years before the ever-mounting evidence for immunophenotyping and genetic characterization of lymphoid neoplasms was recognized as necessary for a more precise and scientific system. With the benefit of highly specific antibodies against human lymphoid biomarkers, both flow cytometry and immunohistochemistry (IHC) can be used to identify neoplastic lymphoid cells in relation to their normal cellular counterparts in the immune system. Some of the most useful antibodies for diagnosis actually represent “trickle down” products that derive from markers originally identified by genetic methods, such as BCL2 and ALK1. Today such methods are requisite for lymphoma diagnosis and of increasingly predictive value in guiding directed therapies, for example, tumor cell expression of CD20 for the use of rituximab.
To confidently interpret IHC in diagnosing lymphoma, one must first be familiar with the distribution of test antigens in normal and hyperplastic benign lymphoid tissues ( Fig. 6.1 ). Only then can deviations from the norm be gauged in relation to a diagnosis of malignancy. So-called built-in positive controls are for this reason particularly appropriate in lymphoid samples, since most, but not all, markers are expressed in reactive elements usually included in tumor biopsies.
It is critical to remember that the same good practice procedures required for high-quality conventional sections are all the more necessary for the production of good IHC. Tissue allocation must avoid the introduction of surface drying artifact, and blocks should be adequately thin to allow satisfactory fixation and processing. Selection of the proper fixative, sufficient fixation time, fluid processing, sectioning, and deparaffinization all require constant attention. When dealing with small tissue samples, a laboratory protocol calling for a number (6 to 10) of unstained sections mounted on coated slides is an important insurance policy for avoiding the misery of being left inadequate tissue for diagnostically essential IHC.
Antigens for Evaluation of Hematologic Disorders
B-Cell Associated Antigens
CD20 is a 35-kDa nonglycosylated membranous phosphoprotein. CD20 is acquired by late pre-B cells as they mature and is typically lost as the B cells become plasma cells. Although the exact function of the CD20 antigen is unknown, it is thought to be involved in B-cell regulation, differentiation, and calcium flux. L26 is the most commonly used commercial antibody. CD20 staining is membranous.
Currently, CD20 is the first-line B-cell lineage defining antibody. Therapeutic strategies involving the monoclonal antibodies directed against CD20 (i.e., rituximab containing chemotherapeutic regimens) have necessitated the use of alternative B-cell lineage defining markers in patients who have received such therapy and are suspected of relapse. Although rarely identified, CD20 expression has been noted on a small subset of nonneoplastic T cells. CD20 expression has also been identified in occasional cases of Hodgkin lymphoma, precursor B acute lymphoblastic leukemia, plasma cell neoplasms, and rarely in T-cell lymphomas.
PAX5 (see later), CD79a (see later), CD19, and CD22 are pan-B-cell antigens also available for staining in paraffin sections and are useful adjuncts to determine B-cell lineage in patients who have received treatment with rituximab or other anti-CD20 agents.
PAX5 (B-cell-specific activator protein) is a nuclear protein belonging to the paired-box containing (PAX) family of transcription factors. PAX5 is thought to commit B cell progenitors to the B-cell lineage by suppressing non–B-cell associated genes and activating B-cell specific genes. A broader regulatory role has been described, including regulation of cell adhesion and migration, inducing V-DJ immunoglobulin heavy chain recombination and facilitating early B-cell receptor signaling to promote development to the mature B-cell stage. The staining pattern in B cells is strong and nuclear; Reed-Sternberg cells show a characteristic weak/variable pattern of nuclear expression.
PAX5 is expressed by early B-cell precursors as well as mature B cells, and it is lost as B cells mature to plasma cells. It is expressed on normal B cells as well as their malignant counterparts and has become a valuable adjunct marker of B-cell lineage. Rare reports of PAX5 expressing non–B-cell lineage neoplasms have been described in anaplastic large cell lymphoma (ALCL) and lymphomatoid papulosis. Expression occurs in nonhematolymphoid malignancies, as PAX5 expression has been identified in atypical carcinoids, small cell lung carcinomas, and large cell neuroendocrine carcinomas.
CD79a is associated with the immunoglobulin receptor complex in the B-cell membrane. It is the B cell marker with the broadest sensitivity as it is expressed in early B-cell precursors (even before immunoglobulin heavy chain gene rearrangement), throughout B-cell maturity, and is eventually lost only in the latest plasma cell stage. The staining pattern is cytoplasmic.
Although a B-cell lineage antigen, CD79a expression has also been reported in cases of T-cell lymphoblastic lymphoma and acute myeloid leukemia (often in acute promyelocytic leukemia).
The BCL6 gene encodes a 79-kDa zinc finger binding protein thought to play a role in B-cell differentiation within the germinal center. BCL6 is expressed on B cells of germinal center origin with a nuclear staining pattern.
Expression of BCL6 protein, demonstrated through IHC, does not necessarily correlate with the presence of a BCL6 gene rearrangement. BCL6 rearrangements are one of the more common genetic abnormalities and are found in diverse lymphomas, not restricted to those arising from a B-cell germinal center origin. In contrast, BCL6 protein is normally expressed nonneoplastic germinal center B cells (GCBs). Lymphomas of germinal center origin, such as follicular lymphoma (FL) or Burkitt lymphoma (BL), express BCL6 protein demonstrated by IHC; however, this does not imply the presence of a BCL6 translocation. Although more often seen in cells of the B-cell lineage, BCL6 expression is reported in occasional ALCLs and peripheral T-cell lymphomas (PTCLs).
Other makers that identify germinal center origin are GCET1, HGAL/GCET2, and LMO2. These markers are used in various classifying systems for determining GCB versus nongerminal center B (NGC) origin of diffuse large B-cell lymphomas (DLBCLs).
Multiple Myeloma Oncogene-1/Interferon Regulatory Family 4
The multiple myeloma oncogene-1 (MUM1) belongs to a larger group termed the interferon regulatory family (IRF4) and encodes a transcription factor responsible for development in B, T, plasma, dendritic, and myeloid cells. There are two commercially available antibodies against human MUM1/IRF4 protein, Mum-1p (monoclonal) and ICSTAT (polyclonal). Staining is both nuclear and cytoplasmic.
MUM1 was originally recognized by its upregulation in multiple myeloma with t(6;14)(p25;q32); however, it was soon found not to be specific for plasmacytic differentiation. MUM1 protein expression has been described in numerous neoplasms including: lymphoplasmacytic lymphoma (LPL), chronic lymphocytic leukemia (CLL), FL, marginal zone lymphoma (MZL), DLBCL, primary mediastinal large B-cell lymphoma (PMLBCL), primary effusion lymphoma (PEL), BL, Hodgkin lymphoma, ALCL, PTCL-not otherwise specified (NOS), adult T-cell lymphoma/leukemia, and melanoma. MUM1 expression is also seen in nonneoplastic “activated” T cells, a subset of GCBs, and normal melanocytes. In general, MUM1 immunostaining is thought to parallel that of CD30 staining. Translocation of IRF4 is seen in a subset of cutaneous ALCL and a small subset of other T-cell lymphomas. The expression of MUM1/IRF4 can be seen in many T-cell lymphomas and be independent of the presence of the translocation. A rare pediatric lymphoma with features of DLBCL or FL, occurring mostly in the Waldeyer ring, has been identified. In contrast to most FLs, these express MUM1, and a substantial portion of these cases have IRF4/MUM1 translocations.
OCT2 and BOB1
OCT2 and its coactivator BOB1 are transcription factors of the POU homeodomain family that bind to the conserved octamer sites in the promoters of the immunoglobulin genes involved in B-cell differentiation and regulation. Staining is nuclear.
BOB1 is a novel 256-amino acid proline-rich protein that is seen predominantly in the B-cell lineage expressed in precursor and mature B cells. BOB1 interacts with either OCT1 or OCT2 to coactivate gene transcription by binding to a number of octamer sites throughout the genome. BOB1 is also called OBF1, OCT Binding Factor 146, OCA-B47, or Bob-148.
Given the postulated mechanisms of action and origins of classic Hodgkin lymphoma (CHL), OCT2 and BOB1 expression were thought to be absent in Reed-Sternberg cells, a fact which was largely used to differentiate the LP (“popcorn cells”) of nodular lymphocyte predominant Hodgkin lymphoma (NLPHL) from Reed-Sternberg cells of CHL. However, weak expression of OCT2 and BOB1 in the Reed-Sternberg cells of CHL has been reported in a subset of cases. Strong consistent expression of OCT2 and BOB1 is reported in NLPHL and in non-Hodgkin lymphomas. OCT2 and BOB1 are also expressed in a subset of acute myeloid leukemias, where their expression may have a prognostic relevance.
Other Important Markers
CD138 (syndecan-1) is a 200-kDa member of the transmembrane family of heparin sulfate proteoglycan proteins. Located on chromosome 2p23-24, CD138 is thought to be a receptor for matrix proteins and a cofactor for growth factors. The anti-syndecan 1 monoclonal antibody Mi15 is frequently used. The staining pattern is membranous.
CD138 is most commonly associated with plasmacytic differentiation and is seen on both benign plasma cells and their malignant counterparts (plasma cell myeloma [PCM]). Among hematolymphoid neoplasms, expression of CD138 is interpreted as evidence of plasmacytic differentiation; however, caution should be used as non-Hodgkin lymphomas and many nonhematolymphoid neoplasms may express this marker. CD138 expression is seen on nonneoplastic epithelial surfaces and correspondingly has been reported among various types of carcinomas. Nonneoplastic early precursor B cells and posttransplant lymphoproliferative disorders express CD138. Plasmablastic lymphoma, LPL and a small subset of cases of CLL express CD138. Expression of CD138 on mesenchymal neoplasms and tumors of melanocytic origin have also been reported.
CD30 is a membrane bound phosphorylated glycoprotein weighing 120 kDa. It is a member of the tumor necrosis factor (TNF) receptor superfamily 8 (TNFRSF8). Monoclonal antibodies used in paraffin include Ber-H2, HeFi, HRS-1, HRS-2, HRS-3, M44, M67, and C10. CD30 is expressed on normal activated T and B cells and virally transformed B and T cells (Epstein-Barr virus [EBV], human T-cell lymphotropic virus-1 or -2 and human immunodeficiency virus [HIV]). Monocytes/macrophages and granulocytes also constitutively express CD30. In lymph node and tonsil sections, a small subset of lymphocytes in the parafollicular areas express CD30. CD30 is thought to transduce a cell survival signal and be involved in the T-cell–dependent portion of the immune response. CD30 staining may be membranous or concentrated in the Golgi zone outside the nucleus (paranuclear).
CD30 is consistently overexpressed on Hodgkin/Reed-Sternberg (HRS) cells. However, CD30 expression is seen in several settings including lymphomatoid papulosis, ALCL, adult T-cell lymphomas, some cutaneous T-cell lymphomas, natural killer (NK) neoplasms, a subset of B-cell lymphomas (DLBCL, BL), and embryonal carcinoma. CD30 expression in neoplasms has become of increasing importance with the development and successful treatment with anti-CD30 monoclonal antibodies. Brentuximab vedotin is an anti-CD30 chimeric IgG1 monoclonal antibody used to target CD30 neoplasms particularly in the setting of a relapse after first line therapy has failed.
Anaplastic lymphoma kinase (ALK) is a 220-kDa tyrosine kinase receptor belonging to the insulin receptor superfamily. Initially described in ALCL, the t(2;5)(p23;q35) created a fusion gene from the nucleophosmin gene and the tyrosine kinase receptor gene (ALK), and the resulting chimeric gene encoded a constitutively activated tyrosine kinase that is a potent oncogene. Additional translocations were subsequently discovered; at least 15 different ALK fusion proteins have been described. Although initially described in ALCL, numerous other malignancies express ALK either as activated fusion proteins derived from chromosomal rearrangements or as mutationally activated ALK proteins (such as the activating mutations described in neuroblastoma). Staining may be cytoplasmic, nuclear, or membranous, and different staining patterns correlate with specific translocations.
Normal ALK immunohistochemical expression is seen in rare scattered neural cells, endothelial cells, and pericytes in brain of adults. ALK expression has been reported in cases of B-cell lymphomas, non-small cell lung cancer (NSCLC), rhabdomyosarcomas, glioblastomas, melanomas, inflammatory myofibroblastic tumors, esophageal squamous cell carcinomas, and systemic histiocytosis. Crizotinib (PF-02341066), a receptor kinase inhibitor, has been used to treat ALCL and NSCLC cell lines that harbor ALK translocations. Ceritinib and alectinib are newer ALK inhibitors with several others currently in development. The ALK1 antibody is used in lymphoma diagnosis, whereas other highly sensitive antibodies (including D5F3, 5A4) are used to detect positive lung cancer and other nonhematopoietic tumor cases. Although the highly sensitive antibodies can be used for lymphoma and hematologic cases, the ALK1 antibody is not sensitive enough to use in the nonhematologic tumors.
Terminal Deoxynucleotidyl Transferase
Terminal deoxynucleotidyl transferase (TDT) is a DNA polymerase that catalyzes the addition of deoxynucleotides to the 3′-hydroxyl terminus. TDT is expressed in precursor B and precursor T lymphocytes during early differentiation; it generates antigen receptor diversity in both cell lines by synthesizing nongerm line elements at the ends of rearranged immunoglobulin heavy chain and T-cell receptor (TCR) gene segments, respectively. Staining may be nuclear and membranous or with paranuclear dot-like positivity.
Nonmalignant TDT expression is seen in sections of the thymus containing precursor T cells, and in bone marrow sections containing early B-cell precursors, so-called hematogones. The malignant blasts of acute T or B lymphoblastic leukemia/lymphoma express TDT and a subset of acute myeloid leukemias and hematodermic CD56+/CD4+ neoplasms. Nonhematopoietic malignancies expressing TDT include some pediatric small round blue cell tumors, Merkel cell carcinoma, and small cell lung carcinoma.
CD43 is a sialomucin expressed on hematopoietic precursors and is thought to play a role in regulation of hematopoiesis. In adults, CD43 occurs both on bone marrow hematopoietic stem cells and on mature white blood cells in the periphery with the exception of resting B lymphocytes. CD43 is also found on tissue macrophages, dendritic cells, smooth muscle cells, epithelium, and endothelium. CD43 expression has been reported on myeloblasts, lymphoma cells, and metastases of solid neoplasms. The aberrant expression of CD43 on B cells can be useful in identifying B cell lymphomas. Staining is cytoplasmic.
Cyclin D1 (BCL1/PRAD1/CCND1) is a transcriptional regulator protein that complexes with the cyclin dependent kinases that maintain regulation of G1 to the S phase of the cell cycle. The CCND1 gene is found at chromosome 11q13. Most notably, mantle cell lymphoma (MCL) is associated with the translocation involving this gene and the immunoglobulin heavy chain locus: t(11;14)(q13;q32). In addition to MCL, a subset of cases of PCM and hairy cell leukemia (HCL) have been found to express cyclin D1. Focal areas of weak cyclin D1 expression have been described in the proliferation centers of CLL. Some cases of DLBCL express cyclin D1 without any significant impact on prognosis. Nonneoplastic lymphoid cells do not express cyclin D1. Overexpression of cyclin D1 is not restricted to hematopoietic malignancies and has been associated with breast carcinoma and NSCLC. Scattered nonneoplastic endothelial cells will often express cyclin D1 and can be used as an internal positive control. Staining is nuclear.
Immunohistochemical antibodies detect an increase in the antiapoptotic BCL2 protein, often resulting from the translocation of the BCL2 gene to a position behind the enhancer elements of the Ig heavy chain gene t(14;18)(q32;q21). In addition to inhibiting apoptosis, the resulting BCL2 oncogene may also block chemotherapy-induced cell death. Staining is membranous.
BCL2 expression may be used as one means of differentiating benign follicular hyperplasia from the neoplastic counterpart FL; however, BCL2 expression should not be interpreted as evidence of malignancy. Intrafollicular T cells, T and B cells in the interfollicular areas, primary follicles, and mantle zone B cells normally express BCL2. Expression of BCL2 is not limited to lymphomas and is commonly encountered in nonhematopoietic malignancies as well. Loss of BCL2 expression may be seen in T-cell lymphomas.
T-Cell Associated Antigens
CD2 is a 50- to 55-kDa transmembrane glycoprotein that is found on both T cells and NK cells. The CD2 genes are found on chromosome 1 and are thought to play a role in antigen independent adhesion and in T-cell activation. In T-cell development, it is thought to appear after CD7. The staining is cytoplasmic.
CD2 expression is seen on both immature (precursor) T cells and mature (peripheral) T cells. Aberrant loss of CD2 expression is seen in a subset of T-cell lymphomas. CD2 expression is usually seen in T acute lymphoblastic leukemia and more rarely may be seen on the myeloblasts of acute myeloid leukemia. CD2 expression in mast cells is considered aberrant and supportive of a diagnosis of mastocytosis.
CD3 is a T-cell antigen composed of four distinct subunits (ε, γ, δ and ζ) that span the membrane and are associated with the TCR and is considered a first-line antigen in identification of T-cell lineage. CD3 staining is membranous and cytoplasmic.
CD3 is first found in the cytoplasm of developing T cells, cytoplasmic CD3 (cCD3). As T cells mature, the CD3 antigen moves to the surface. Similarly, the neoplastic counterparts show a similar distinction with cytoplasmic CD3 on precursor T-cell neoplasms and surface CD3 seen on peripheral “mature” T-cell neoplasms. Although most widely used as a T-cell lineage specific antigen, CD3 expression (cytoplasmic and membranous) has been reported on B-cell lymphomas, particularly those with expression of EBV.
CD4 and CD8 T-cell surface molecules play a role in T-cell recognition and activation by binding to their respective class II and class I major histocompatibility complex (MHC) ligands on an antigen-presenting cell (APC). CD4 has the additional role of stabilization of the TCR complex; it is also a major target of HIV. Staining for both CD4 and CD8 is both cytoplasmic and membranous.
Early T-cell precursors express both CD4 and CD8 simultaneously, and then with maturation lose one of these markers. CD4 staining is seen in the T helper class of T cells, the predominant population in the T-cell compartment. CD8 staining is seen in the cytotoxic T-cell population and nonneoplastic sinusoids of the spleen. The majority of T-cell lymphomas express CD4, not CD8, and are thought to be derived from the T helper lineage. Cytotoxic T-cell lymphomas expressing CD8, not CD4, are a proportionally smaller group of lymphomas. Rarely, lymphomas may have aberrant loss of both these antigens; however, a small subset of nonneoplastic T cells will also be “double negative” for these two markers, the γ/δ T cells, and this should not be misinterpreted as aberrant antigen loss.
CD5 is a 67-kDa type I glycoprotein that is thought to attenuate signals arising from the crosslinking of the TCR and its antigen on the MHC of APCs. Staining is membranous.
CD5 is expressed by the majority of peripheral (or mature) T cells; the loss of this marker may be seen as one of the first findings in a developing T-cell lymphoma. However, this finding has also been reported in reactive populations of T cells. Expression of CD5 has also been used to distinguish the neoplastic thymocytes of thymic carcinoma from a benign thymoma. Although a T-cell lineage antigen, CD5 is famously coexpressed aberrantly on certain B-cell lymphomas: CLL/small lymphocytic lymphoma (SLL) and MCL. More rarely, cases of MZL, DLBCL, and cases of FL may express this T-cell antigen as well. In addition, a small nonneoplastic subset of B cells, the B1a cells, normally expresses this T cell marker.
CD7 is a 40-kDa glycoprotein member of the immunoglobulin gene family. CD7 antigen is thought to be involved in signal transduction, proliferation, and adhesion. The CD7 monoclonal antibody, CBC.37, shows expression on the majority of peripheral T cells (i.e., mature T cells), NK cells, and precursor T cells (immature T cells). Staining is membranous.
CD7 is one of the earliest markers in T-cell development. Although a T-cell lineage antigen, CD7 expression may be seen in small populations of fetal bone marrow B cells, and myeloid precursor cells; however, this is subsequently lost early with differentiation. As with CD5, loss of CD7 antigen may be seen in the setting of a T-cell lymphoproliferative process as well as in a benign reactive setting. Aberrant CD7 expression has been described in the myeloblasts of acute myeloid leukemia, where it has been associated with Fms-like tyrosine kinase-3 internal tandem duplication (FLT3/ITD) mutation, and significantly shorter disease-free/postremission survival.
The T-cell compartment is composed of CD4 expressing T helper cells, and CD8 expressing T cytotoxic cells. T-cell intracytoplasmic antigen (TIA1), granzyme B, and perforin are markers used to identify the cytotoxic T cells that induce lysis of their targets by using these granule-associated cytotoxic proteins. These cytotoxic markers can also be seen in NK cells and granulocytes. Expression of TIA1 can be detected in all cytotoxic cells, whereas granzyme B and perforin expression can be detected in high levels only in activated cytotoxic cells. Expression is cytoplasmic.
CD56 and CD57 are frequently used to identify the NK cell lineage; however, they are frequently found expressed in other, nonhematopoietic tissues. CD56, or neuronal cell adhesion molecule (N-CAM), is a NK cell marker that is also expressed in the central and peripheral nervous systems. Most cases of PCM (~70%) also express this marker. Similarly, CD57 is a marker expressed on NK cells as well as other T cells; it is occasionally used in the diagnosis of NLPHL to visualize the small T cells ringing the LP cells. The staining of CD56 (monoclonal antibody 123C3) and CD57 are membranous.
NK cells express CD2, CD7, CD8, CD56, and CD57. They are positive for cytoplasmic CD3, but they not surface CD3 and do not typically express CD5. The neoplastic counterpart: extranodal NK/T-cell lymphomas express CD2, cytoplasmic CD3, CD56, and in most cases EBV.
Additional markers for NK cells include killer inhibitory receptors (KIRs). KIR expression is identified using monoclonal antibodies specific for CD94, CD158a, and CD158b.
βF1 staining identifies a portion of the beta subunit of the T cells that carry the α/β TCR. T cells with the α/β TCR normally represent the majority of T cells (i.e., 95%). Similarly, the majority of T-cell lymphomas express the α/β TCR. Staining is membranous.
Negativity for βF1 stain implies that the T cells carry the alternative TCR subunits: γ/δ. Small subsets of normal γ/δT cells are found in the skin, splenic red pulp, mucosal associated lymphoid tissue (MALT), and in the medulla of the thymus. γ/δ T cells have a distinct immunophenotypic profile so that CD2 and CD3 are positive but CD4 and CD8 are negative (double negative T cells), and CD5 is likewise usually negative. The neoplastic counterparts also express the γ/δ TCR: hepatosplenic T-cell lymphoma and subcutaneous panniculitic type T-cell lymphoma (SPTCL). However, caution should be used in interpreting this negative staining as evidence of a γ/δ T-cell derivation, as NK cells also are negative for β F1, staining with CD56 in this instance helps to identify these cells.
Immunohistochemical staining for the δ subunit has become available but is technically challenging. Its use is not widespread, but when appropriate, staining can confirm γ/δ origin in embedded tissues.
Lymphoid Enhancer-Binding Factor 1
Lymphoid enhancer-binding factor 1 (LEF1) is associated with a gene that encodes a transcription factor belonging to a family of proteins that share homology with the high mobility group protein-1. The protein can bind to an important site in the TCR-alpha enhancer, inducing enhancer activity. It is also involved in the Wnt signaling pathway and may function in hair cell differentiation and follicle morphogenesis. Mutations in this gene have been found in somatic sebaceous tumors. This gene has also been linked to other cancers, including androgen-independent prostate cancer.
LEF1 is a nuclear stain and staining can be variable in intensity, but any staining, even weak, is considered positive. Aside from hematopoietic cells, there are only a few positive cells types. Colon cancer, but not normal colon, shows expression, and pancreatic tumors and normal pancreas show expression. There is normal expression in hair follicles in skin/adnexal structures.
In hematopoietic tissues, LEF1 is a pan-T-cell antigen. It has the same distribution and number of cells positive as typical T cell stains including CD3. LEF1 is aberrantly expressed in the majority of cases of CLL/SLL. It is positive in >90% cases of CLL/SLL. It is not considered to be expressed in any significant number of non-CLL small B-cell lymphomas. A recent report suggests that LEF1 staining may be seen in 5% to 8% of cases of MCL. It is expressed on a subset of DLBCL and can be expressed in Richter syndrome.
Additional Markers Useful for Evaluation of Hodgkin Lymphoma
Immunoglobulin (Ig) D is an immunoglobulin with a delta heavy chain present on the surface of B lymphocytes or as a soluble form in plasma. IgD and IgM are coexpressed on the surfaces of most peripheral B cells. IgD is normally expressed in resting mantle cells and can be useful in identifying primary follicles. In addition, the small dark blue lymphocytes in the nodules of NLPHL are typically of mantle cell type and are positive for IgD. IgD expression in LP cells has been reported in a subset (27%) of NLPHL cases with an extrafollicular distribution of LP cells, and a relatively T-cell–rich background. In contrast, IgD expression is rarely seen in T-cell/histiocyte–rich large B-cell lymphoma (TCHRLBCL).
PD1 (CD279) and PDL1
Programmed cell death protein 1 (PD1; CD279) is a cell surface receptor of the immunoglobulin superfamily. PD1 expression is primarily seen in follicular T-helper cells (TFH). Expanded extrafollicular PD1 immunoreactive cells have been reported in TFH cell derived PTCLs such as angioimmunoblastic T-cell lymphomas (AITLs), other PTCLs, Rosai-Dorfman disease, and some reactive viral lymphadenitis. The pattern of immunoreactivity is membranous. A recent study by Menter and coworkers demonstrated that the ligand of PD1, PDL1, is expressed in HRS cells of 70% evaluated CHL cases, 54% of NLPHL, and approximately a third of PMLBCLs and DLBCLs.
CD45 or leukocyte common antigen (LCA) is a 200-kDa transmembrane glycoprotein expressed on most hematopoietic cells, with the exception of erythroid cells and a subset of plasma cells. The protein is a member of the protein tyrosine phosphatase family and plays an essential role in T- and B-cell receptor signaling. The neoplastic cells in NLPHL, also known as popcorn cells or lymphocytic and histiocytic (LP) cells because of their distinctive morphology, generally express CD45. CD45 is typically absent of HRS in CHL. CD45 has a strong membranous pattern of immunoreactivity. It is often difficult to discern CD45 expression in tumor cells owing to strong immunoreactivity of surrounding cells in CHL. One should look for areas without adjacent cells to determine if the tumor cells are CD45 immunoreactive.
Establishing the Diagnosis of Lymphoma: Differential Diagnostic Considerations
Conventional microscopic findings, coupled with the clinical history, is still the standard for making a lymphoma diagnosis. Although modern lymphoma classification emphasizes relatedness of the neoplastic process to normal cell counterparts of the immune system (i.e., B-cell versus T-cell versus NK-cell), microscopic cellular features and patterns of proliferation frame the differential diagnosis. This starting point for pathologists is all-important since the cost-effective selection of definitive special study methods, in particular IHC, depends on their acumen. It is usually more critical for the patient that the diagnosis of lymphoma per se be correct than that its classification be precise. There are many processes, both benign and malignant, which can microscopically and sometimes even clinically simulate lymphoma. IHC can be a very powerful tool in resolving differential diagnostic puzzles; however, the correct markers must be brought to bear.
Low-Grade Lymphoma Versus Chronic Lymphoid Hyperplasia
The most common differential diagnostic quandary consists of the often difficult distinction of low-grade (B-cell) malignant lymphoma versus chronic immune hyperplasia, usually in an older patient. The overriding principle in this setting is not to make the diagnosis of lymphoma unless one is absolutely certain, because in most cases, early detection is not effective in eradicating such lymphomas in any event, and an erroneous diagnosis of malignancy may take a long time, if ever, to be suspected clinically. As was mentioned in the Introduction, the pathologist’s strength is in recognizing the microscopic and IHC hallmarks of a benign immune reaction. Functional immune compartments are distinct and recognizable, including follicle centers, mantle zones, paracortex, sinuses, and medullary cords. Special cellular compartments may also be recruited and expanded, such as plasmacytoid dendritic cells, granulomas, and abscesses. Appropriate IHC markers that label these compartments may be of use in confirming the diagnosis of benign hyperplasia, for example, B-cell marker CD20 and CD10 or BCL-6 for follicle centers and CD3 for the T cells of the paracortex. A diffuse, uniform cellular composition is the hallmark of a neoplastic process. Various combinations of IHC markers can be used to confirm the diagnosis—for example, expression of one or more B-cell markers with only a small minority of interspersed T cells. Abnormal coexpression of certain markers, such as CD43 and/or CD5, or demonstration of immunoglobulin light chain restriction can add certainty to the diagnosis and assist further in proper classification of the lymphoma ( Fig. 6.2 ). Even with such evidence caution must be exercised since rare benign conditions have been identified with light chain restriction and CD43 coexpression.
When select compartments are disproportionately expanded, it raises the question of the early appearance of a particular form of low-grade lymphoma. Such cases can be extremely difficult to diagnose. For example, FL versus follicular hyperplasia is perhaps the most commonly recognized conundrum within this grouping. When even meticulous evaluation of conventionally stained sections fails to resolve this issue, IHC for BCL2 protein and for Ki67 is almost always conclusive ( Fig. 6.3 ). As is always the case, the interpretation of such IHC studies requires experience and skill. It is not a matter of just “positive” or “negative” staining but rather a complex pattern recognition challenge. In rare cases even further study is warranted, such as polymerase chain reaction (PCR) for clonality testing. A related differential diagnosis addresses the sometimes subtle distinction between follicular hyperplasia with the variant features of progressively transformed germinal centers and the early appearance of NLPHL. Here IHC is the definitive ancillary method. CD20 and BCL6 highlight scattered individual tumor cells against a background of small mantle zone B lymphocytes, and collarettes of small T cells surrounding the tumor cells ( Fig. 6.4 ).
The IHC studies indicated depend on the differential diagnosis ( Table 6.1 ). For example, when medullary cord regions of a lymph node are selectively expanded by a plasmacytoid cellular proliferation, staining for kappa and lambda light chains is likely to be informative. In situ hybridization for light chain mRNA has become a widely popular substitute for IHC because this method eliminates the problem of background plasma immunoglobulin as a visually interfering phenomenon ( Fig. 6.5 ). The recognition of the earliest involvement of sampled lymphoid tissues by MZL or CLL/SLL can be almost impossible without powerful ancillary studies. Sometimes, the effort is triggered by detection of a small population of light chain restricted cells in flow cytometry or by suspicious clinical findings. Indeed, the distinction of these two processes can itself be difficult. Coexpression of CD5, LEF1, and CD23 weighs strongly in favor of CLL/SLL (see Table 6.1 ). Because MZL has essentially no disease-specific markers, in comparison to other types of low-grade B-cell lymphoma, its diagnosis is often based on exclusion of alternative types. In some cases of very subtle, early nodal involvement by MZL, only molecular probe determinations can detect the process.
|Diagnosis||Primary Evaluation||Additional Evaluation|
High-Grade Malignant Lymphoma Versus Acute Immune Hyperplasia
Fortunately, only rarely do benign immune reactions simulate high-grade lymphoma. When they do, it represents an emergency for clinical management. The histologic hallmark of aggressive lymphomas is the predominance of large, sometimes bizarre appearing cells with high proliferation. In the setting of normal host immunity, there are only a few processes that can instigate benign mimics of this type. Viral infections, and in particular those of the Herpes family, can induce an intense immune proliferation featuring expanses of large immunoblastic cells. Primary EBV infection (infectious mononucleosis) is the classic example that is most often encountered in tonsillectomy specimens. Detection of this virus with IHC for EBV latent membrane protein or with in situ hybridization for EBV-encoded RNA (EBER) is effectively conclusive ( Fig. 6.6 ). Herpes simplex, human herpesvirus 6 (HHV-6), and varicella viruses can also less commonly be the culprit for mimicry of high-grade lymphoma. In the setting of abnormal host immunity , there are a number of processes, some yet only poorly understood, that can microscopically mimic high-grade lymphoma. EBV-driven B-cell proliferations in the setting of immunosuppression can mimic either high-grade B-cell non-Hodgkin lymphoma or CHL. Zones of necrosis and plasma cellular maturation are features that may tip off the wary pathologist as to the benign nature of such cases; however, clinical history is the most essential means to avoid an erroneous diagnosis of lymphoma. Highly atypical expanses of T-cell proliferation are seen in Kikuchi-Fujimoto histiocytic necrotizing lymphadenitis. The demonstration of a CD8-positive population of large T-cell immunoblasts in combination with histiocytes in the absence of neutrophils is the key to recognizing this mysterious entity, which may be a relatively mild, self-limited autoimmune disorder. The autoimmune lymphoproliferative syndrome (ALPS) is a congenital disorder relating to the absence of cellular Fas receptor, permitting T cells to escape normal apoptotic destruction. The resulting accumulation of these T cells in both peripheral blood and lymphoid tissues produces a microscopic appearance resembling a PTCL. However, these cells are negative for both CD4 and CD8 while staining for pan-T markers, such as CD3 and CD2.
Cases leaving no doubt as to their neoplastic nature may yet pose challenges to the pathologist with a different form of mimicry, that of other forms of malignancy that can be mistaken for lymphoma. This differential diagnostic grouping is the subject of the final section in this chapter.
Immunohistochemical Evaluation of Small B-Cell Lymphomas
In most circumstances, initial evaluation of lymph nodes can determine whether the histologic and architectural pattern represents general categories of lymphoid processes. In general, these can be divided into groups including reactive processes, small B-cell lymphomas, large cell lymphoma, Hodgkin lymphoma, or other considerations. This categorization is simplified, as there may be circumstances in which one must consider overlap between these groups.
In this section, an overview of small B-cell lymphomas predominantly with indolent clinical behavior is covered; these account for about 40% of all B-cell lymphoma. The most common lymphomas considered in this evaluation include FL (grades 1 and 2), CLL/SLL, MZL, and MCL (see Table 6.1 ). Other more rare subtypes of small B-cell lymphomas including HCL, LPL, and plasma cell disorders are covered in the latter portion of the section.
The diagnosis of lymphoma should be arrived at by a combination of findings including clinical, cytologic, histologic, genetic, and other laboratory results. In many cases, the combination of clinical and histologic features are distinctive enough to make a diagnosis; however, immunophenotype may occasionally reveal unusual antigen expression, uncover important clinical or clinicopathologic features, or add clarity to a difficult diagnostic problem.
General Immunohistochemical Approach to Diagnosis of Small B-Cell Lymphoid Lesions
Histologic findings are the cornerstone of diagnosis of small B cell lymphoid proliferations. However, because of considerable overlap between entities and small samples with limited architectural features, a panel of immunohistochemical stains is of considerable benefit in the evaluation of lymphoid lesions.
Foremost is evaluation of CD20 and CD3. These stains allow identification of amount and distribution of B and T cells in almost all cases. In cases of small B-cell lymphomas, CD20 is markedly increased and likely labels much of the tissue present. In partial involvement or unusual cases, CD3 positive T cells may be more numerous. Staining for CD5 (1) allows identification of abnormal CD5 coexpressing B cells and (2) acts as a secondary pan-T-cell antigen (in case of a possible T cell neoplasm). The combination of BCL2 and BCL6 staining adds several benefits. Most classically, distinction of reactive from neoplastic follicles can be determined with normal follicles positive for BCL6 and negative for BCL2, whereas the majority of abnormal follicles of low-grade FL are positive for BCL6 and BCL2. Cyclin D1 is another critical stain, which is primarily used to identify or exclude MCL. LEF1 can be of great benefit in evaluation of small B-cell lymphomas. It is a nuclear stain, allowing precise localization. Although it is a pan-T-cell antigen, it is coexpressed on 90% to 95% of cases of CLL/SLL.
CD43 can be of benefit in the evaluation of small B-cell lymphoma. Normal B cells do not express CD43. However, many small B cell lymphomas are positive for CD43, including most cases of CLL/SLL and MCL. Approximately 20% to 30% of MZL are positive for CD43, and approximately 50% of LPL. CD43 can be useful in the differential diagnosis of lymphoma types as well: (1) FL is almost never positive, and (2) if the differential diagnosis includes CLL/SLL or MCL, then a negative result would favor other diagnoses such as MZL or FL.
In the authors’ experience, CD10 is only of limited benefit and is better replaced by BCL6 in most cases. In addition, CD23 expression is also considered of limited benefit. Although its primary use is to distinguish CLL/SLL from MCL, there are subsets of cases of each type that show aberrant expression patterns, reducing the utility of this stain. Furthermore, expression of LEF1 would typically provide support for a diagnosis of CLL/SLL and make MCL unlikely. The expression or lack of cyclin D1 would trump the results of CD23 expression in any case. In addition, there is variable expression of CD23 in other small B-cell lymphomas (FL, MZL), which does not add diagnostic clarity but adds another potentially confusing result. The use of light chain staining for kappa and lambda in small B-cell lymphomas is controversial. In most cases, if there is clear evidence of lymphoma, then demonstration of light chain restriction is unnecessary. If there is no evidence of plasma cell differentiation, light chain expression and interpretation is positive only in a limited number of cases, using the standard immunohistochemical staining in most laboratories. It can be useful in cases with plasma cell differentiation; it can demonstrate the presence or absence of light chain restriction in such a population.
Finally, Ki67 can be quite useful in the evaluation of lymphoid lesions including small B-cell lymphomas. The overall pattern of staining can be instructive and quite distinctive for a specific diagnosis. The low proliferation seen in follicles is characteristic of low-grade FL; it can even be used in cases where there is a lack of abnormal BCL2 expression. Furthermore, it can be used to highlight the follicular colonization in MZL and the proliferation centers of CLL/SLL.
General Prognostic and Therapeutic Issues
Prognosis and therapy in small B-cell lymphomas is typically well defined. In most cases of small B-cell lymphoma, there is a relatively good prognosis, with indolent disease and a gradual increased risk over time of transformation to more aggressive disease. Histologic evidence of transformation to a large B-cell lymphoma, to a blastic B-cell malignancy, or to other histologic transformations is associated with more aggressive disease and a poor outcome.
In general, genetic abnormalities, except as defining the disorder, are not specifically associated with outcome. However, there are general findings that do apply. Abnormalities associated with p53 tumor suppressor gene (deletion at 17p13) have been associated with a poor outcome. Similarly, a complex karyotype beyond the typical genetic abnormalities implies genetic instability in a lymphoma and can portend a stormy clinical course.
At present, for almost all B-cell lymphomas, rituximab, an anti-CD20 humanized monoclonal antibody, is included as part of their therapy. This therapeutic option is based on the almost universal expression of CD20 by mature B-cell malignancies. As a secondary agent, or in rare cases as a primary agent, anti-CD22 therapies (epratuzumab; inotuzumab ozogmicin) have also been used. Although not in general use at present, anti-BCL2 therapies (oblimersen, an antisense BCL2 oligodeoxynucleotide) have been used in some trials and may find use as primary or salvage therapies for CLL/SLL and other lymphoma types.
FL is a distinctive B-cell lymphoma of abnormal cells of follicular origin. In cases of low-grade FL (grades 1 and 2), there is typically an architectural arrangement of abnormal follicular structures present. However, in small samples, the characteristic architecture may not be appreciated. In exceptional cases, there may be a partly or predominantly diffuse pattern in otherwise unremarkable FL. The immunophenotype is distinctive and helpful in the diagnosis. The lymphoma cells express pan-B-cell antigens (CD19, CD20, CD22, PAX5) with expression of germinal center associated antigens, such as BCL6. It should be noted that CD10 can also be seen in benign and malignant germinal centers. However, the loss of CD10 expression in the neoplastic cells of FL is not uncommon. Other stains that are associated with the germinal center, with high degrees of sensitivity and specificity, include HGAL and GCET1. HGAL may be of benefit in identifying FL in bone marrow samples, as it retains more robust staining compared with BCL6 in decalcified marrow samples. The vast majority of low-grade FL are positive for BCL2, and this expression is most often due to the presence of the t(14;18) IGH/BCL2 translocation with overexpression of BCL2. Proliferation rate in low-grade FL will be <5% to 25% within follicles. As in normal follicles, follicular dendritic cells will be highlighted by CD21 staining. Follicular dendritic cells (FDCs) will also be variably positive for CD23, and the FL cells may also express CD23. FL lack expression of cyclin D1, CD5, CD43, and only rarely express cytoplasmic light chains by IHC or in situ staining with clear evidence of monoclonality.
Atypical immunophenotypic findings in low-grade FL include expression of CD5 or CD43 (rare cases; less than 1% to 2%), or have an increased proliferation rate with an apparent low-grade histology. These latter cases may have a more aggressive clinical course than usual low-grade FL, and a diagnosis of “FL grade 1–2/3 with increased proliferation rate” should be noted. In situ follicular neoplasia (formerly follicular lymphoma in situ ) is an indolent proliferation of neoplastic cells that have only a limited number of the molecular “hits” required to develop FL. They only rarely progress, but they must be distinguished from partial lymph node involvement by FL. In situ follicular neoplasia is characterized by involvement of a single or a few follicles by abnormal lymphocytes. They are brightly positive for CD10 and BCL2, with a low proliferation index by Ki67. This is in contrast to normal follicles, which are negative for BCL2, and dim positive for CD10 and FL, which is typically dim positive for BCL2 and dim positive or negative for CD10.
Pediatric FL are most often seen in younger patients in the head and neck region and most often have large, blastoid-appearing follicles (grade 3/3). They typically lack BCL2 expression in contrast to usual type of FL grade 3.
Prognostic and Therapeutic Studies
Follicular lymphoma international prognostic index (FLIPI) has been shown to strongly correlate with prognosis in FL. Histologic progression in FL can occur in 5% to 60% of cases and is associated with a poor prognosis. Progression includes genetic instability with secondary genetic and molecular events, associated with a poor prognosis including chromosomal gains of 7, 12q13-14, and 18q and chromosomal losses del6q, 9p21, and 17p13 by array comparative genomic hybridization (aCGH). Poor prognosis is associated with gains of 7, 12q13-14, 18q (oncogenes or dosage effects) as well. Transformation of FL to large-cell lymphoma (occur in 5% to 60% cases at 10 years) is of critical clinical significance in the overall prognosis.
Potential pitfalls in FL arise from some inconsistent immunohistochemical findings and variations with different grades of FL. It is important to be advised that not all FLs are BCL2 protein expression positive. These will typically have other standard features of FL, including an abnormal low proliferation rate by Ki67. IgH PCR can be negative in a significant number of cases of FL. FL grade 3 can be challenging, as they are more commonly BCL2 negative, and will have a high proliferation rate, comparable to hyperplastic processes.
Chronic Lymphocytic Leukemia/Small Lymphocytic Lymphoma
CLL/SLL is a relatively indolent B-cell lymphoma, with frequent bone marrow and peripheral blood involvement. The appearance of CLL/SLL cells is that of small round lymphocytes with dense mature chromatin and scant cytoplasm. There is almost always a subset of cells with large round nuclei, prominent nucleoli, and increased amounts of cytoplasm (prolymphocytes or paraimmunoblasts). Mitotic figures are very rare. Larger cells will sometimes cluster, forming vague nodules called variously proliferation centers or pseudofollicles. The typical immunophenotype for CLL/SLL includes expression of CD5, LEF1, CD23, CD19, CD43, and BCL2. Most cases express CD20, although weak, and even focal expression may be seen in a subset of cases. They will most typically have a proliferation rate of less than 10%, although increased proliferation may be seen in pseudofollicles/proliferation centers by Ki67 staining. CLL/SLL lacks expression of CD10, cyclin D1, and only rarely express cytoplasmic light chains by IHC or in situ staining with clear evidence of monoclonality. There are typically no appreciable FDC networks unless there are preserved normal follicles present.
Atypical immunophenotypic findings in CLL/SLL include lack of staining for CD5 or CD23, focal expression of cyclin D1, and rarely, evidence of plasma cell differentiation. CLL/SLL without CD5 expression is estimated to be present in approximately 2% of cases. It should be noted that other CD5-negative lymphomas should be ruled out carefully before diagnosing CD5-negative CLL/SLL. Lack of CD23 expression is rare. In this circumstance, because of the close overlap with the immunophenotype of MCL, it is appropriate to exclude MCL carefully, by either cyclin D1 staining or genetic studies for the t(11;14) of MCL. Likewise, CD5-positive MZL may not express CD23 and should be considered in the differential diagnosis. Focal expression of cyclin D1 has been noted in CLL/SLL; it is typically seen in proliferation centers. It is a rare occurrence and has no specific impact on diagnosis. Likewise, these cases will not have evidence of the t(11;14) IGH/CCND1 of MCL. Mutations in TP53 , NOTCH1 , SF3B1 , and BIRC3 have been identified in CLL/SLL and are associated with a poor prognosis.
Monoclonal B-cell lymphocytosis (MBL) is considered to be a precursor lesion to CLL/SLL. It shares immunophenotypic features of CLL/SLL, and the distinction of the two entities requires correlation with degree of tissue involvement and in the case of peripheral blood, absolute lymphocyte counts.
Prognostic and Therapeutic Studies
Increased proliferation centers that are large or confluent and have a high proliferation index by Ki67 are considered an independent poor prognostic factor in SLL. Immunoglobulin heavy chain variable region (IGHV) mutation status has been shown to be associated with prognosis in CLL. Cases with unmutated heavy chain variable gene regions are associated with a poor prognosis compared with those with mutated IGHV genes. CD38 expression correlates partly with IGVH status and prognosis, but it may be independent of IGHV in some cases. Immunohistochemical or flow cytometric staining for ZAP-70, a tyrosine kinase that plays a role in signal transduction, has been shown to correlate with prognosis in CLL/SLL. ZAP-70 expression by both flow and IHC has been shown to correlate with IGHV mutation status and with prognosis. Overexpression of p53 by IHC suggests an aggressive clinical course.
Poor risk cytogenetics include 17p deletions (5% to 8%), 11q deletions (20%), trisomy 12 (15%), complex karyotype, mutated IGHV genes, and expression of CD38. NOTCH1 mutations (10%) are associated with a poor prognosis. Del 17p is associated with rapid disease progression, drug resistance, and overall short survival. The t(14;19) is a rare finding but is associated with a relatively aggressive clinical course. Del13q14, no expression of ZAP70, negative CD38, and mutated IGHV are genetic and laboratory findings associated with a good prognosis. Splicing factor 3b subunit 1 (SF3B1) is associated with fludaribine refractory disease in some cases.
Pitfalls in the diagnosis of CLL/SLL are mostly due to variations in immunophenotype, as mentioned earlier. It should be recognized that variation in CD5, CD23, CD20, and BCL2 may be seen in individual cases. Likewise, focal staining for cyclin D1 and some cases with high proliferation by Ki67 can be seen ( Fig. 6.7 ). In all cases, exclusion of other lymphoma types, such as MCL, CD5-positive MZL, CD5-positive LPL, and other rare lymphoma types is prudent.
Mantle Cell Lymphoma
Of the predominantly small B-cell lymphomas, MCL is distinctive. Although there are exceptions, in most cases MCL has a poor prognosis despite more aggressive therapeutic approaches, although some recent improvement in outcome has been achieved. Histologic patterns of diffuse, nodular, and mantle zone types are described. In conventional cases, the cytologic appearance is that of small lymphocytes, with slightly irregular nuclei, dense, mature chromatin, and scant cytoplasm. Notably, the cells typically lack nucleoli, and in most cases, large lymphocytes are rare. In the background, bland histiocytes with round or ovoid nuclei and ample pale or pink cytoplasm are seen. Occasional mitotic figures may be seen. MCL expresses pan-B-cell antigens (CD19, CD20, CD22) with expression of CD5, CD43, BCL2, and cyclin D1 ( Fig. 6.8 ). The cornerstone of diagnosis of MCL, cyclin D1 staining is seen in almost all abnormal cells, and variation in intensity of this nuclear stain is typical in conventional MCL, with more uniform strong staining seen in the blastoid variant (later). MCL lacks staining for CD10, BCL6, CD23, and only rarely expresses clear evidence of monoclonality using cytoplasmic light chains by IHC or in situ staining. No FDC networks are seen with CD21 staining in diffuse cases; the nodular or mantle zone patterns may have residual FDC networks in follicles.
Atypical immunohistochemical findings include coexpression of CD10 (very rare) and CD23 (can be seen in up to 25% of cases, with mostly focal and/or weak staining). Lack of CD5 expression may be seen in very rare cases. There are rare literature reports of “cyclin D1 negative MCL.” These cases are thankfully rare (≪1%); often coexpress cyclin D2, cyclin D3, or cyclin E; and have a histologic appearance and clinical course comparable with that of typical MCL. The majority of these cases are SOX11 positive, which can be helpful in their diagnosis. BCL6 expression can be seen in about 12% of cases of MCL, with MUM1 expression seen in 35%. In situ mantle cell neoplasia is a limited involvement, limited stage disease, which is localized typically to the mantle zones of otherwise unremarkable nodes. These are identified by routine staining for cyclin D1 (and/or SOX11). Only rarely do these cases progress to overt MCL.
Prognostic and Therapeutic Studies
The proliferation rate in MCL by Ki67 is often less than 30%, and cases with more than 30% are associated with a worse prognosis. TP53 overexpression, as a surrogate for loss of p53 tumor suppressor activity, is associated with poor prognosis in MCL. SOX11 staining in MCL lymphoma has been suggested as a prognostic marker, but this is controversial. Although most indolent MCL cases are SOX11 positive, this marker does not adequately distinguish between indolent and aggressive cases.
Not all cyclin D1–positive neoplasms are MCL. A subset of plasma cell neoplasms is cyclin D1 positive with some positive for the t(11;14). HCL lacks the t(11;14), but it has weak cyclin D1 expression in most cases. A small subset of DLBCL has expression of cyclin D1 but in the absence of the t(11;14).
Although most cases of MCL are associated with aggressive clinical behavior, a small subset has indolent behavior. The basis for recognition of these cases has not been established. There are also cases of in situ MCL, which have only limited lymph node involvement, confined to mantle zones, that are not always associated with progressive disease.
LPL is a rare, and indolent B-cell lymphoma with proliferations of small lymphocytes, lymphoplasmacytic cells, and plasma cells in varying proportion. In almost all cases, there is bone marrow and peripheral blood involvement, with varying degrees of involvement of spleen and lymph nodes. The morphologic and immunophenotypic distinction of LPL from MZL (both nodal and extranodal) can be difficult. When associated with a monoclonal protein of IgM type, the preferred diagnosis is Waldenström macroglobulinemia (WM)/LPL.
The immunophenotype of LPL is not distinctive; LPL will express pan-B-cell antigens CD19, CD20, CD79a, and PAX5. BCL2 is positive in almost all cases. There is often expression of IgM by neoplastic lymphoid and plasma cells. Both plasma cells and lymphoplasmacytic cells will express plasma cell associated antigens CD38, VS38c, and CD138, with monoclonal expression of light chains. Rare cases are positive for CD5 (17% to 43%), with some positive for CD10 (16%) and many positive for CD23 (58%). LPL does not express cyclin D1 or BCL6. Most cases of LPL are positive for CD25.
Although not entirely specific, an increase in mast cells has been noted in LPL/WM. Methods to highlight increases in mast cells (including CD117 and/or tryptase staining) could provide diagnostic support in difficult cases.
MYD88 L265P mutations are identified in more than 90% of cases of LPL and very rare cases of other small B-cell lymphomas. In general, in the appropriate clinical and histologic context, the presence of this mutation is strongly supportive of a diagnosis of LPL. CXCR4 mutations are found in about 30% of cases.
Prognostic and Therapeutic Studies
No specific immunohistochemical findings are associated with prognosis in LPL. A variety of genetic findings can be seen in LPL; cases with del6q have been associated with a worse prognosis. Gains in 4q and 8q occur in 12% and 10% of WM/LPL and are not commonly reported in MZL.
The primary pitfall in LPL and WM/LPL is the clear diagnosis of this entity. This diagnosis can be difficult, and correlation of clinical, histologic, immunophenotypic, and genetic information is critical.
Nodal Marginal Zone Lymphoma
Nodal marginal zone lymphoma (NMZL) is a low-grade B-cell lymphoma that is predominantly composed of small, mature appearing B cells. It is expected to have predominantly lymph node involvement to distinguish it from extranodal MZL (covered later); it does not have prominent splenic involvement, as this would suggest a diagnosis of splenic marginal zone lymphoma (SMZL; see later discussion). NMZL may have evidence of follicular colonization, a marginal zone pattern, monocytoid B-cell differentiation and plasma cell differentiation, although these features are variable in individual cases. The immunophenotypic findings in NMZL are not entirely specific, and they are mostly derived from a lack of staining seen in other specific types. Distinction from LPL is difficult. NMZL will typically express pan-B-cell antigens including CD19, CD20, PAX5, and CD79a. Expression of BCL2 is seen in almost all cases. CD43 coexpression by B cells is seen in 50% of cases. Proliferation assessment by Ki67 is generally low (less than 10%), and if above 30% to 40%, it raises the possibility of a variant of DLBCL. NMZL lacks expression of CD5, CD10, BCL6, and cyclin D1, which is in contrast to cases of MCL, CLL/SLL, and FL. Rare cases may express CD5, but without a careful correlation of clinical, histologic, and genetic features, distinction from CLL/SLL can be difficult.
Pediatric NMZL is a distinctive variant associated with young males (2 to 18 years). These cases are almost always stage I at presentation, are seen in head and neck locations, and have an excellent prognosis, even with conservative management.
Although trisomies of some chromosomes can be identified, none of the translocations seen in extranodal marginal zone are seen in NMZL. Genetic findings include trisomies 3, 7, 12, and 18; structural rearrangement of chromosome 1; and gains in several regions of chromosome 3. The presence of translocations specific for other lymphoma types would preclude a diagnosis of NMZL. Several gene mutations have been identified in NMZL, including MLL2 (also known as KMT2D, 34%), PTPRD (20%), NOTCH2 (20%), and KLF2 (17%).
Prognostic and Therapeutic Studies
No specific pathologic testing to predict prognosis is known for NMZL. Features that are prognostic include age and stage of disease. FLIPI has been shown to strongly correlate with prognosis in NMZL.
Extranodal Marginal Zone Lymphoma
Extranodal marginal zone lymphoma (EMZL) is an indolent lymphoma found in diverse sites and with variable morphology. They arise at a number of extranodal sites including stomach, lung, small intestine, salivary gland, thyroid, skin, and other sites. They recapitulate some features of MALT and are often referred to as “MALT lymphomas.” EMZL often arise in a background of chronic antigenic stimulation. As such, the presence of reactive germinal centers and polyclonal plasmacytosis may be seen in association with EMZL. A distinctive feature of MZL is follicular colonization; it is characterized by invasion of reactive germinal centers by neoplastic lymphoid cells. The cellular composition of EMZL is predominantly small lymphocytes. Frequently, these will have increased amounts of pale cytoplasmic (e.g., monocytoid B cell appearance). There are admixed larger transformed lymphocytes, which typically number less than 30% of the overall cellularity. In addition, varying degrees of plasma cell differentiation may be present. Dutcher bodies and Russell bodies may be seen in these plasma cells. When extensive, the degree of plasma cell differentiation may raise the possibility of plasmacytoma. It should be noted that some authors believe that many extramedullary plasmacytoma are simply localized manifestations of EMZL with extensive plasma cell differentiation.
The immunophenotype of EMZL is often based on exclusion of other lymphoma types. The neoplastic lymphocytes express pan-B-cell antigens (CD20, CD79a, PAX5) and are positive for BCL2 in the vast majority of cases. They will have a low proliferation rate (<10% in most cases) by Ki67. They will lack expression of CD10, cyclin D1, and BCL6. Rare cases (<5%) will express CD5 and tend to be in nongastric sites. Follicular colonization can be highlighted by the use of BCL6 staining or Ki67, highlighting follicles (BCL6 positive, Ki67 high), and they can be disrupted by lymphoma cells (BCL6 negative, Ki67 low) ( Fig. 6.9 ).
EMZL can be difficult to distinguish from LPL. In LPL, increased mast cells, marrow, and peripheral blood involvement, a large amount of serum IgM are more common and may be helpful in distinguishing the two. In addition, clear evidence of CD20 and PAX5 positive plasma cells or lymphoplasmacytic cells would favor a diagnosis of LPL.
A broad range of genetic abnormalities, including trisomies, and a variety of translocations are seen in MZLs ( Table 6.2 ). These have different frequencies depending on the sites. The presence of t(11;18) (API2-MALT1) or t(1;14)(BCL10-IGH) are seen in gastric EMZL, and these cases tend not to respond to Helicobacter pylori eradication therapy. Trisomies of chromosome 3 and 18 are also seen frequently in EMZL, but these are not entirely specific to EMZL.
|Anatomic Site||Infectious Agent||Translocation||Gene||Frequency|
|Intestine||Campylobacter jejuni a||t(11;18)(q21;q21)||API2-MALT1||15%|
|Ocular adnexa||Chlamydia psittaci a||t(3;14)(p14.1;q32)||FOXP1||20%|
|Skin||Borrelia burgdorferi a||t(14;18)(q32;q21)||MALT1||14%|
Lymphoepithelial lesions are prominent in many EMZL. This is due to the homing present in the neoplastic cells to epithelial structures. The use of pan-keratin stains (such as AE1/AE3) may be useful in highlighting the lymphoepithelial lesions of EMZL ( Fig. 6.10 ).
H. pylori has been linked to the development of EMZL in the stomach. It serves as the initial and sustaining antigenic event in both the reactive lymphoid proliferation and subsequent proliferation of lymphoma cells. This critical pathophysiologic mechanism in several cases is also critical, as a large portion of these lymphomas will respond favorably to H. pylori eradication therapy. The literature has stated that immunohistochemical staining for H. pylori is not necessary in routine cases ( Fig. 6.11 ). However, given the prevalence and relative impact on diagnostic and prognostic decisions, the use of the immunohistochemical stains for H. pylori in cases with a suspicion of lymphoma is warranted.
Hairy Cell Leukemia
HCL is a rare, low-grade B-cell lymphoma, with prominent involvement of spleen, bone marrow, and peripheral blood. It only rarely involves lymph nodes, most commonly in the splenic hilum or intraabdominal region. The morphology is of small to intermediate-sized lymphocytes with round nuclei and moderate to large amounts of pale cytoplasm. In peripheral blood, the abnormal lymphocytes will have delicate, hair-like projections, giving the disorder its name. In histologic preparations, HCL will often have a round centrally placed nucleus with a rim of pale or clear cytoplasm; this has been referred to as a “fried egg” appearance.
In most circumstances, HCL is identified by flow cytometry of either blood or bone marrow samples. The flow cytometric immunophenotype is distinctive, and shows expression of pan-B-cell antigens includes CD20 (bright) and CD19 with restricted light chain expression. HCL will also have a characteristic pattern of expression of CD103, CD11c, CD22, and CD25. FMC7 is positive, as is CD123. By IHC, CD123 is rarely identified, but it is seen by more sensitive flow cytometry. Most cases lack CD5 and CD10 expression, although rare cases of either or both have been reported.
Primary diagnostic immunohistochemical studies of HCL in tissue include expression of tartrate-resistant acid phosphatase (TRAP) and DBA.44 ( Fig. 6.12 ). Annexin-1 is positive in almost all cases, but can be difficult to interpret due to extensive normal staining of background elements. Cyclin D1 is positive in a large percentage of cases of HCL and can be useful in discriminating between HCL and other lymphoma types; it stains less intensely and less uniformly than in cases of MCL. T-bet lacks sensitivity and specificity but, if positive, could support a diagnosis of HCL. Immunohistochemical staining for mutation for BRAF V600E is highly sensitive and specific for the diagnosis of HCL.
No specific cytogenetic abnormalities have been associated with HCL. Most cases of HCL will harbor a BRAF V600E mutation. Although not necessary for primary diagnosis of HCL, it may be of benefit in cases with a differential diagnosis of similar or related entities, such as SMZL, HCL-variant (HCL-V), or splenic diffuse red pulp small B-cell lymphoma (SDRPL). A small subset of BRAF mutation negative cases of HCL have IGVH4-34 abnormalities. These cases have a poor prognosis compared with typical HCL.
Splenic Marginal Zone Lymphoma
SMZL is uncommon and accounts for less than 1% of all non-Hodgkin lymphoma (NHL). SMZL is an indolent lymphoma of small B cells that prominently involves splenic white pulp, bone marrow, and peripheral blood. It is purported to arise from cells of the splenic marginal zone. The morphology has features of MZLs in other sites including monocytoid differentiation, follicular colonization, plasma cell differentiation, and admixed large cells.
Immunophenotype in SMZL includes expression of pan-B-cell antigens including CD20, PAX5, and CD19. In general, SMZL lacks expression of CD5, CD10, CD23, cyclin D1, CD43, and CD123. Some cases show TRAP expression but will lack expression of DBA.44, in contrast to HCL. CD103 is positive by flow cytometry, but this marker is not presently available for paraffin IHC. Most cases coexpress IgM and IgD. Annexin A1 is uniformly negative. CD5 expression can be seen in approximately 5% of cases. Ki67 staining will reveal a low proliferation rate. However, follicular colonization is highlighted by BCL6 or Ki67 staining. Normal residual follicular structures will show staining for both BCL6 and Ki67, with disruption of normal follicular architecture by SMZL cells that are negative for both markers. Cases with overexpression of p53 may be associated with an aggressive clinical course.
Deletion 7q is a relatively common genetic abnormality in SMZL, but its frequency varies with the method of detection. The presence of 7q deletion makes diagnoses of CLL/SLL, MCL, HCL, and FL unlikely. However, other less well-defined splenic B-cell lymphomas (SDRPL, HCL-V; later) have 7q deletion rates comparable to SMZL, making distinction by the criterion alone impossible.
KLF2 is mutated in 10% to 40% of SMZL, thus representing the most frequently mutated individual gene in this lymphoma. Second, the NOTCH pathway is affected in approximately 30% to 40% of SMZL, and the nuclear factor κ-B (NFK-B) pathway is mutated in ~25%.
Hairy Cell Leukemia-Variant
HCL-V is an exceedingly rare lymphoma, with prominent splenic, bone marrow, and peripheral blood involvement. Although relatively indolent in clinical course, it is more aggressive than conventional HCL, and some cases may be aggressive and refractory to typical treatment. In peripheral blood, HCL-V cells will be intermediate to large in size with prominent nucleoli and moderate to large amounts of pale cytoplasm. Cytoplasmic projections may be apparent in peripheral blood smears. In tissue, the cells will be intermediate to large in size, although the nucleoli may be less apparent than in peripheral blood. Similar to conventional HCL, they will have a diffuse red pulp pattern of involvement in the spleen.
Most cases of HCL-V are identified by flow cytometry of peripheral blood or bone marrow. HCL-V will be positive for B cell markers, such as CD20, CD19, and PAX5. They also express CD11c, CD22, CD103, DBA.44, and FMC-7, with bright expression of immunoglobulin light chains and CD20. In contrast to most cases of conventional HCL, HCL-V will lack expression of cyclin D1, CD123, CD25, TRAP, and annexin A1.
No specific cytogenetic abnormalities are identified in HCL-V. At present, in contrast to conventional HCL, HCL-V does not have evidence of BRAF mutations. Mutations of MAP2K1 are seen in the majority of cases of HCL-V and is also seen in case of HCL with IGVH4-34.
Splenic Diffuse Red Pulp Small B-Cell Lymphoma
SDRPL has only recently been identified as a distinctive entity. Traditionally, because of the overlapping features with SMZL, it has been considered a “red pulp variant” of SMZL. It has similar histologic and clinical findings as SMZL. However, because of comparable peripheral blood and bone marrow findings, SDRPL can only be diagnosed by splenectomy, since the diagnosis is based on the architectural pattern in spleen.
Immunohistochemical staining highlights SDRPL as a B-cell lymphoma (CD20 positive), and lacking specific markers of other lymphoma types. BCL6, TRAP, CD10, CD23, annexin A1, and cyclin D1 are negative. DBA.44 is positive in most cases (88%), suggesting that this marker is not useful in distinguishing small B cell lymphomas of the spleen. CD123 is reported as being negative in SDRPL. Most cases are positive for IgG (67%) with a minority positive for IgD (27%). CD5 has been noted in rare cases. Proliferation rate, as measured by Ki67, is less than 10%. CD43 is reported as negative. TP53 is seen in a subset of cases (29%), but its overall impact on prognosis is not clear. SDRPL has similar frequency of 7q deletion to SMZL.
Plasma Cell Neoplasms
Plasma cell disorders encompass a broad range of neoplasms of differing clinical aggressiveness. A number of types of lymphomas may have plasmacytic or plasmablastic differentiation and are covered in other sections. This section covers manifestations of PCM and extraosseous plasmacytomas. PCM is an aggressive disorder of post-GCB derivation, with end differentiation into plasma cells. It is notable for patchy systemic distribution with prominent involvement in bone and bone marrow. The term plasmacytoma can be used in two ways: (1) as a localized extraosseous tissue manifestation of PCM or (2) as a single isolated mass without associated systemic disease. These two scenarios can be difficult to distinguish with only pathology findings. In this context, isolated plasmacytoma is used to refer to the non-PCM related type.
Plasma cell differentiation by IHC can be evaluated by the presence of expression of CD79a, CD138, CD38, VS38c, and MUM1. Another hallmark is the strong expression of restricted cytoplasmic lights chains (kappa or lambda) in almost all cases. Evaluation of heavy chains (IgA, IgG, IgM, IgD) may be of benefit in some cases, but it is not highly specific in a diagnostic or therapeutic context. In most cases, mature neoplastic plasma cells will not express pan-B-cell antigens CD20, PAX5, CD19, or express CD45. Most cases of PCM (~70%) express CD56. Isolated plasmacytoma will typically not express CD56, p53, CD117, or cyclin D1. If any of these markers were present, this would be more suggestive of tissue involvement by PCM.
A subset of PCM will express CD20. In some of these cases, there is concurrent presence of a t(11;14) translocation identical to that seen in MCL. PCM with t(11;14) will express cyclin D1, CD19, CD20, and PAX5, along with expression of plasma cell markers CD138 and MUM1. Differentiation from MCL can be difficult, and in many cases, the presence of clinical features of myeloma may be necessary to make a conclusive diagnosis. Occasional cases of myeloma without the t(11;14) will express cyclin D1; this marker in a case with clear plasma cell differentiation can exclude other alternative diagnoses such as LPL.
Large B-Cell Lymphomas and Other Aggressive B-Cell Lymphomas
The 2016 WHO classification includes many types of large B-cell lymphomas, which includes classification of entities based on clinical features, morphology, histologic features, immunohistochemical features, genetics, and combinations of these features. Sorting of this large category often requires careful correlation of a variety of clinical and pathologic findings. In some cases, there are no specific therapeutic differences, but there are a number of subcategories associated with differences in prognosis. IHC is key in identifying many of the different types of large B cell lymphomas.
In almost all cases, primary identification is made by a histologic pattern along with the presence of pan-B-cell antigens, most often CD20. In the appropriate context, the combination of large cells, in a diffuse pattern, with clear expression of CD20 may be “adequate” for identification of DLBCL. Subclassification of specific types is dependent on a variety of additional antibody combinations that allow for specific subtyping.
Modern lymphoma classification relies on the presence of subsets within the general category of DLBCL. In older systems, morphologic findings, such as immunoblastic, centroblastic, or anaplastic features were noted. However, groupings based on morphologic findings alone lacked sufficient reproducibility to prognosticate in this large group of disorders. More recent classifications have taken gene expression array findings to group DLBCL into categories GCB derived, activated B cell type (ABC), and a third group that includes PMLBCL. The current WHO classification requires identification of GCB and non-GCB types. GCB type lymphomas are associated with a better prognosis using current therapeutic regimens such as R-CHOP (rituximab, cyclophosphamide, doxorubicin hydrochloride, vincristine sulfate, and prednisone). ABC types of lymphomas have a more aggressive clinical course; they appear to be dependent upon dysregulation of the NFK-B pathways.
Gene expression arrays are not currently applied in routine diagnosis, and as such, a number of immunohistochemical systems have been proposed to act as surrogates for gene expression arrays. Although the Hans classifier was considered to be the benchmark for some time, subsequent studies have shown that it is lacking in sensitivity and specificity necessary to accurately predict gene expression array patterns. As such, other systems (tally classifier, Choi classifier) have been proposed as being more accurate ( Fig. 6.13 ).