Chapter 96: Pathology of Lymphomas

The classification of malignant lymphoma has been fraught with controversy during much of the 20th century, with much needed consensus reached during the past two decades. A detailed discussion of the history of lymphoma classification is beyond the scope of this chapter and can be found elsewhere.¹

From Thomas Hodgkin’s description in 1832 of what became known as Hodgkin’s disease² to the first half of the 20th century, several types of lymphomas with distinctive morphologic and clinical features were described using a variety of terms, including lymphoma, lymphosarcoma, reticulum cell sarcoma, and giant follicular lymphoma.¹ However, many of the terms were not used uniformly, resulting in significant misunderstanding, particularly between pathologists and clinicians. Starting in the 1930s, several attempts were made to classify lymphomas and provide some uniformity of diagnosis, culminating in the Rappaport classification, initially published in 1956, which divided lymphomas based on growth pattern, cell type, and stage of differentiation.^3,4 Most importantly, this classification demonstrated clinical relevance, showing that lymphomas with a nodular pattern had a better prognosis than diffuse lymphomas.

In the 1960s and 1970s, an explosion of studies on the immune system had a profound effect on our understanding of lymphocyte biology and had a consequent effect on our understanding of malignant lymphomas. Normal lymphocytes could now be classified into distinct lineages (B, T, and natural killer [NK]), which could be determined by expression of lineage-specific surface antigens and eventually by genetic analysis of B- and T-cell receptors.^5,6 Several new lymphoma classification schemes were developed to incorporate the new immunologic data, the most important being the Kiel classification⁷ (used primarily in Europe) and the Lukes and Collins classification⁸ (used primarily in North America). By the 1970s, at least five classification schemes were widely used in different parts of the world. At the same time, clinical studies were beginning to show that some patients with aggressive lymphomas could be cured with combination chemotherapy.⁹ Oncologists needed to interpret results of clinical trials performed in different institutions, a situation made difficult by the use of different classification schemes that were not easily translated among themselves.

The problem was addressed by the United States National Cancer Institute, which convened a large group of investigators to determine which classification scheme was best at predicting clinical outcome of lymphoma. None of the classification schemes was identified as predicting clinical outcome better than the other schemes. Therefore, pathologists were advised to continue using one of the six classification schemes studied, and a “Working Formulation” was developed so that oncologists could translate clinical data derived in different institutions using different classification schemes.¹⁰ Lymphomas were divided into 10 categories based solely on morphologic features. To help clinicians deal with a large number of lymphoma subtypes, the lymphomas were further grouped into three clinical prognostic groups (clinical grades). Although the Working Formulation was not intended to be a standalone classification scheme, it was used as such by many institutions, particularly in North America.

However, increasing phenotypic and genotypic data were further defining several distinctive lymphoma subtypes. The Working Formulation lumped different lymphomas into broad categories that were obscuring the distinctive features of the newly described entities. The Working Formulation categories were based solely on morphologic features and were not able to incorporate new immunologic and molecular genetic data that were recognizing new types of lymphomas.

In the 1980s and 1990s, several new lymphoma entities were identified based on new immunologic and molecular genetic data. Although attempts were made to incorporate these new entities into the existing classification schemes,¹¹ problems with uniformity between different institutions persisted. A desire to eliminate the continued confusion ultimately led to a new approach to lymphoma classification proposed by the International Lymphoma Study Group that used all available information, including morphology, immunophenotype, genetic and clinical features, to define a list of distinctive entities that could be uniformly diagnosed by hematopathologists. The proposal was published in 1994 and was known as the Revised European-American Lymphoma (REAL) classification.¹² Importantly, this classification identified entities that had distinctive clinical features and could be reproducibly diagnosed by expert hematopathologists.¹³

In the late 1990s, a new World Health Organization (WHO) classification for lymphoproliferative disorders was being developed, based on the REAL classification. First published in 2001 (and revised in 2008), the WHO classification represented a consensus between an international group of more than 50 experienced hematopathologists, including contributions from a clinical advisory committee of hematologists and oncologists experienced in treating lymphomas.¹⁴ The WHO classification (Table 96–1) identified several major categories, including precursor lymphoid neoplasms, mature B-cell neoplasms, mature T- and NK-cell neoplasms, and Hodgkin lymphoma.

Distinctive lymphoma entities were identified based upon a combination of morphologic, immunophenotypic, genetic, and clinical features. The 2008 WHO classification includes several provisional entities and categories of unclassifiable neoplasms with features intermediate between two distinct entities. This allows the classification to retain flexibility so that new data that further identify distinct diseases within these entities can be incorporated. In distinction from the Working Formulation, lymphomas were not classified based on clinical outcome. The WHO classification agreed that each type of lymphoma that was identified by pathologic and clinical features could have a spectrum of clinical aggressiveness, and that lumping distinct entities into groups based on clinical outcome would inhibit the development of targeted therapeutic approaches. Therefore, the WHO classification represents a complete change from the Working Formulation, with the emphasis on pathologic classification rather than classification based on survival characteristics.

Genome-wide expression studies have been instrumental in further delineating distinctive subtypes of lymphomas of clinical relevance. Using complementary DNA microarray technology, the expression of thousands of genes at the mRNA level can be studied simultaneously and compared to other tumor samples.¹⁵ Such studies have (1) defined more than one distinct entity in what was previously characterized as a morphologically homogeneous category, (2) identified distinct gene expression patterns that each encompasses a disease that may demonstrate morphologic heterogeneity, and (3) identified new surface molecules and signaling pathways that could provide targets for new therapeutic approaches. The impact of the new gene expression profiles is detailed in the sections on separate lymphomas below.

The WHO classification attempts to correlate each lymphoma to normal lymphocyte biology by postulating a cell of origin for each neoplasm. This correlation is particularly well-suited for B-cell lymphomas in which several distinct stages of normal B-cell development can be identified (Fig. 96–1) but is not as satisfying for T and NK neoplasms. Briefly, B-cell development begins in the marrow with precursor B lymphoblasts that differentiate into naïve B cells that circulate in the blood. The lymph node is the primary site where B cells encounter antigen, where naïve B cells colonize primary follicles and in mantle zones of secondary follicles (Figs. 96–2,96–3,96–4). Upon antigen stimulation, these cells undergo blast transformation and enter the germinal center reaction in the late primary immune response and the secondary immune response. In the germinal center, cells downregulate BCL2 (Fig. 96–5) and initially transform into intermediate-size cells (follicular B blasts), then into large centroblasts, and finally into small centrocytes (Figs. 96–6 and 96–7).¹⁶ Cells that survive the germinal center upregulate BCL2 and either differentiate into short-lived plasma cells through an immunoblast stage or differentiate into memory cells that populate follicular marginal zones or recirculate in the blood. Several B-cell lymphomas can be correlated with these stages of development (Fig. 96–8) and are mentioned in the sections on separate lymphomas below.

Determining a benign from malignant lymphoid infiltrate often can be difficult because malignant lymphocytes in many lymphomas closely resemble their benign counterparts. Therefore, diagnosis commonly rests on demonstrating a combination of an abnormal architectural pattern, an abnormal immunophenotype, and evidence of lymphoid monoclonality. As a result, several ancillary special studies have become instrumental in the diagnosis and classification of lymphoma, requiring special handling of the biopsy material (Table 96–2). Whenever a diagnosis of lymphoma is considered clinically, the surgeon should perform an open biopsy of the largest involved lymph node. The lymph node should be removed intact whenever possible, because assessment of architecture is extremely important in the diagnosis and classification of lymphomas. The lymph node should be sent immediately to the pathology laboratory in the fresh state, at which time the pathologist allocates the tissue for fixation for routine histology and for special studies.

Automated flow cytometry on single-cell suspensions prepared from tissue samples is extremely helpful in demonstrating B-cell clonality by surface immunoglobulin light chain restriction. It also determines the expression pattern of surface markers helpful in subclassifying lymphomas, particularly lymphomas of small B cells.¹⁷ A wide variety of antibodies that can be used on formalin-fixed tissue now are available, allowing accurate diagnosis and subclassification of lymphoma in most cases.

Molecular genetic techniques to determine B- or T-cell monoclonality or lymphoma-specific chromosomal translocations include polymerase chain reaction (PCR), Southern blot, fluorescence in situ hybridization (FISH), and cytogenetic analysis.¹⁸ PCR and FISH can be performed on formalin-fixed tissue. Results from molecular genetic testing should be interpreted in conjunction with the morphologic and immunophenotypic data, as some benign reactive lymphoid proliferations show evidence of lymphoid monoclonality.¹⁹

The diagnosis of lymphoma has become more complex than the diagnosis of other malignancies because diagnosis of lymphoma rests on correlation of morphologic features with immunophenotype and genetic data in many cases. Because of the complexity of diagnosis and the relative infrequency of lymphoma in general pathology practice, a second review by a hematopathologist with expertise in lymphoma pathology is recommended. The second review can have a significant impact on the clinical management of patients.²⁰

Whereas open biopsy of an involved lymph node is the most useful diagnostic procedure, core-needle biopsies and fine-needle aspiration can play a role in limited situations. Core-needle biopsy might be helpful in the diagnosis of deep-seated disease in the abdomen, allowing the patient to avoid a laparotomy. However, a definitive diagnosis by core biopsy is not always possible, necessitating an open biopsy. Fine-needle aspiration is not helpful in primary diagnosis of lymphoma,²¹ but it may be helpful in detecting recurrence of a previously diagnosed lymphoma or in ruling out a nonhematolymphoid lesion causing lymphadenopathy. Although automated flow cytometry can be used in conjunction with cytologic examination to provide additional information for lymphoma diagnosis and classification, tissue biopsy generally is required before commencement of therapy.

Lymphoblastic leukemia/lymphoma represents a malignancy of lymphoblasts, either of B or T lineage. They can present in the marrow (leukemia) or with predominant tissue involvement (lymphoma), but they are considered single-disease entities. Most cases of acute lymphoblastic leukemia are of B lineage, whereas most cases of lymphoblastic lymphoma are of T lineage, with the mediastinum being a common site of involvement. The morphologic features are the same regardless of site or lineage, consisting of small- to intermediate-size cells with finely dispersed nuclear chromatin, inconspicuous nucleoli, and scant cytoplasm (Fig. 96–9). Assessment of lineage and distinction from minimally differentiated acute myeloid leukemia require immunophenotypic data and may require molecular genetic analysis of B- and T-cell receptors. Lymphoblastic neoplasms are distinguished from other lymphomas by the expression of terminal deoxynucleotide transferase, which is specifically expressed at the lymphoblast stage of development.

The 2008 WHO classification includes several categories of B-lymphoblastic leukemia/lymphoma characterized by recurrent genetic abnormalities.¹⁴ Many of these are associated with distinct clinical or pathologic features, have prognostic implications, or are considered biologically distinct entities.

CHRONIC LYMPHOCYTIC LEUKEMIA/SMALL LYMPHOCYTIC LYMPHOMA

Chronic lymphocytic leukemia is a neoplasm of mature B lymphocytes characterized by blood and marrow involvement and commonly associated with lymph node involvement (Chap. 92). Small lymphocytic lymphoma is the nonleukemic form of the disease. Lymph nodes involved by chronic lymphocytic leukemia show a diffuse infiltrate of small mature lymphocytes admixed with prolymphocytes and paraimmunoblasts, which characteristically form ill-defined nodules known as proliferation or growth centers (Figs. 96–10 and 96–11). The B cells have a characteristic immunophenotype, demonstrating CD5 and CD23 expression and dim expression of CD20 and clonal immunoglobulin light chain. Studies have divided chronic lymphocytic leukemia into two distinct subtypes with distinct clinical behavior (Chap. 92). The type with the more favorable prognosis expresses mutated variable regions of the immunoglobulin heavy-chain (IGH) genes, whereas the other subtype expresses unmutated IGH genes. The IGH gene mutation status is reflected in differences in gene expression.^22,23 The gene encoding the zeta-associated protein of 70 kDa (ZAP-70) is one of these genes, which generally is expressed by leukemia cells that express unmutated IGH genes and hence can be used to discriminate between the two subtypes.²⁴ Certain cytogenetic abnormalities also correlate with clinical aggressiveness.²⁵

Some cases of chronic lymphocytic leukemia/small lymphocytic lymphoma demonstrate plasmacytic features but are distinct from an entity known as lymphoplasmacytic lymphoma, which is characterized by a prominent component of plasmacytic lymphocytes and plasma cells (Fig. 96–12). These cases typically do not express CD5, less often involve blood, and often are associated with a monoclonal immunoglobulin M serum protein that can cause hyperviscosity or cryoglobulinemia (Waldenström macroglobulinemia). Somatic mutations in MYD88 are a commonly recurring and highly specific feature of Waldenström macroglobulinemia.²⁶

MANTLE CELL LYMPHOMA

Mantle cell lymphoma most commonly involves lymph nodes, but it can involve extranodal sites, including the gastrointestinal tract, as a clinical variant known as lymphomatous polyposis (Fig. 96–13). It typically is composed of a uniform population of small lymphocytes with cleaved nuclei and a virtual absence of large transformed cells (Fig. 96–14).^27,28 It most commonly has a diffuse growth pattern, but it can show a nodular or, more rarely, a mantle zone pattern (Fig. 96–15). The postulated cell of origin is the B cell of the inner mantle zone. The lymphoma cells coexpress CD5, as does chronic lymphocytic leukemia, but mantle cell lymphoma can be distinguished by lack of CD23 expression and expression of cyclin D1 (Fig. 96–16). Cyclin D1 expression results from the chromosomal translocation t(11;14)(q13;q32) characteristic of mantle cell lymphoma. Gene expression data have demonstrated a subset of mantle cell lymphomas that are cyclin D1–negative.²⁹ Some of these cyclin D1–negative cases have chromosomal translocations involving the cyclin D2 gene.³⁰ SOX11 expression is a highly specific immunohistochemical marker for mantle cell lymphoma, and can identify cases that are negative for cyclin D1.³¹ Overall, patients with mantle cell lymphoma have a median survival of approximately 3 years, but gene expression data that determine tumor cell proliferation are able to identify patient subsets that differ in median survival by more than 5 years²⁹ (Chap. 100).

Cyclin D1–positive lymphocytes in mantle zones can be an incidental finding in reactive lymphoid follicles, a condition referred to as in situ mantle cell lymphoma. These appear to have an indolent behavior and do not require treatment.³²

FOLLICULAR LYMPHOMA

Follicular lymphoma is a proliferation of cells that correspond to normal germinal center cells,³³ retaining expression of germinal center markers (BCL6, CD10), and demonstrates a follicular architecture (Fig. 96–17) imparted by nodular aggregates of CD21-positive follicular dendritic cells. Follicular lymphomas are composed of a variable mixture of centrocytes (small cleaved cells) and centroblasts (large noncleaved cells). They can be divided into three grades (grades 1 to 3) based on the number of centroblasts present. The most common is grade 1 (0–5 centroblasts per high-power microscopic field), previously known as follicular small-cleaved-cell lymphoma (Fig. 96–18). Both grade 1 and grade 2 tumors are indolent, and distinguishing between them is not required. Grade 3 follicular lymphoma (>15 centroblasts per high-power microscopic field) can be further divided into grade 3A (mixture of centroblasts and centrocytes) (Fig. 96–19) and grade 3B (solid sheets of centroblasts). Data have shown some molecular genetic differences between 3A and 3B cases, but further study is required because no significant clinical impact has been demonstrated.^34,35 Follicular lymphoma can have an accompanying diffuse component, and identification of a diffuse area of large cells (diffuse large B-cell lymphoma) indicates transformation to a more aggressive disease. Approximately 90 percent of follicular lymphoma demonstrate the t(14;18)(q32;q21) involving rearrangement of the BCL2 gene, leading to the constitutive expression of the antiapoptotic BCL2 protein. Although BCL2 protein expression does not help distinguish follicular lymphoma from other lymphomas, it is a helpful feature in distinguishing it from reactive follicles that are BCL2-negative (Fig. 96–20).

Follicular lymphoma in situ is an entity where BCL2-positive germinal centers are present in an otherwise reactive lymph node. When it is distinguished from partial involvement by follicular lymphoma, follicular lymphoma in situ has a very low rate of progression to overt follicular lymphoma.³⁶

MARGINAL ZONE B-CELL LYMPHOMAS

Marginal zone lymphomas are characterized by a proliferation of small lymphocytes, commonly with abundant pale cytoplasm (called monocytoid B cells) and plasmacytic features. The postulated cell of origin of these lymphomas is the postgerminal center B cell of the marginal zone at various anatomic sites. Marginal zone lymphomas can be divided into three distinct types based on site of presentation: (1) extranodal marginal zone lymphomas of mucosa-associated lymphoid tissue (MALT), (2) splenic marginal zone lymphomas,³⁷ and (3) nodal marginal zone lymphomas (Fig. 96–21).³⁸ This classification is supported by distinctive cytogenetic abnormalities in each entity. Extranodal lymphomas of the MALT type are the most common and arise in mucosal sites subject to longstanding chronic inflammation (Fig. 96–22), including chronic infection, the prototypical example being chronic Helicobacter pylori infection of the stomach.³⁹ At early stages of development, many of these lymphomas respond to treatment with antibiotics to eradicate H. pylori, whereas later changes, including cases with chromosomal translocations activating genes involved in nuclear factor-κB (NF-κB) signaling,⁴⁰ lead to antigen-independent growth (Chap. 101).

DIFFUSE LARGE B-CELL LYMPHOMA

Diffuse large B-cell lymphoma (DLBCL) is characterized by a diffuse infiltrate of large B cells that can resemble centroblasts or immunoblasts (Figs. 96–23 and 96–24). The 2008 WHO classification identifies several types of large B-cell lymphoma, the most common type being DLBCL not otherwise specified, which constitutes 25 to 30 percent of all non-Hodgkin lymphomas.

Gene-expression data show that DLBCL is a heterogeneous disease consisting of at least three entities having distinct gene-expression profiles based on cell of origin: (1) cases with an expression profile similar to germinal center B cells (GCBs), (2) cases expressing genes typical of activated B cells (ABCs), and (3) cases with a different pattern referred to as “unclassifiable” that are neither GCB-type nor ABC-type (Fig. 96–25).⁴¹ Importantly, clinical differences were apparent, with GCB-type cases having a significantly better prognosis compared to the other two types, even when clinical prognostic markers are considered (Chap. 98). Further studies confirmed these differences in the current era of therapy (including anti-CD20 antibody therapy), and identified nonneoplastic cells in the microenvironment as important contributors to patient survival.⁴² New therapies with selective activity in these subtypes of DLBCL are under development.⁴³ It has been suggested that division of DLBCL into clinically distinct groups may be determined by the expression profile of a limited number of genes using routine immunohistochemistry.⁴⁴ However, such an approach to classification is limited by problems of reproducibility of immunohistochemical staining and interpretation.⁴⁵ The study of gene expression patterns of a small number of genes using formalin-fixed paraffin-embedded material may provide a rapid and accurate method to classify DLBCL.⁴⁶

Rearrangement of the MYC gene is present in 5 to 10 percent of DLBCLs, and is associated with an inferior prognosis.⁴⁷ Approximately half of these cases will also have a rearrangement involving the BCL2 gene, referred to as a “double-hit” lymphoma. Such double-hit lymphomas have a very poor prognosis.⁴⁸ MYC protein expression is present in approximately 30 percent of DLBCLs, which may be independent of gene rearrangement. Concurrent expression of MYC and BCL2 in DLBCL is associated with an inferior prognosis.⁴⁹

Mediastinal large B-cell lymphoma is a distinct subtype of DLBCL that has been separately identified in the WHO classification.⁵⁰ Patients with mediastinal lymphomas typically are younger than those with conventional DLBCL. The histology shows large cells with abundant cytoplasm associated with diffuse fibrosis (Fig. 96–26). Gene expression studies have demonstrated an expression profile that is distinct from conventional diffuse large B-cell lymphoma and shares some features with classic Hodgkin lymphoma (Fig. 96–27).^51,52 Indeed, the 2008 WHO classification recognizes that some cases of mediastinal lymphomas can have features that are intermediate between DLBCL and classical Hodgkin lymphoma.¹⁴

BURKITT LYMPHOMA

Burkitt lymphoma is a highly aggressive lymphoma characterized histologically by a diffuse infiltrate of intermediate-size cells with a high mitotic rate. The lymphomas commonly have significant spontaneous cell death (apoptosis), which results in a “starry sky” appearance caused by numerous macrophages that have engulfed the apoptotic debris (known as tingible body macrophages) (Figs. 96–28 and 96–29). The postulated cell of origin is the early follicular B blast cell of the germinal center. Virtually all cases of Burkitt lymphoma are characterized by chromosomal translocations involving the MYC gene on chromosome 8. The MYC gene most commonly is translocated to the IGH gene on chromosome 14, resulting in t(8;14)(q24;q32), but it also can involve the light-chain genes on chromosomes 2p12 (κ) and 22q11 (λ). A diagnosis of Burkitt lymphoma can be suggested based on morphologic examination alone but should be supported by immunophenotypic data (positive for CD20, CD10, and BCL6; negative or focally weakly positive for BCL2; growth fraction near 100 percent as determined by Ki67 stain) and confirmed by molecular testing for MYC translocations whenever possible.

Gene-expression studies have shown that Burkitt lymphoma has a consistent gene-expression signature, but that there is not always correlation between the diagnosis based on gene-expression profiling and the diagnosis based on standard diagnostic testing.^53,54 To reflect this, the 2008 WHO classification recognizes a provisional entity of B-cell lymphoma, unclassifiable, with features intermediate between DLBCL and Burkitt lymphoma.¹⁴ Many of these cases represent “double-hit” lymphomas, which carry a MYC gene rearrangement and another chromosomal rearrangement, often involving the BCL2 gene.⁴⁸

T cells and NK cells share several immunophenotypic and functional features; therefore, these neoplasms are grouped together in the WHO classification. These lymphomas make up 10 to 15 percent of non-Hodgkin lymphomas in Western countries, with a higher incidence in Asia. Mature T-cell lymphomas comprise a heterogeneous group of neoplasms, the most common subtype being the peripheral T-cell lymphoma (PTCL) not otherwise specified.

PTCLs typically grow in a diffuse pattern that effaces normal nodal architecture or, more rarely, show expansion of the interfollicular areas. They show a diverse cytologic spectrum, with most cases showing a mixture of large- to intermediate-size cells and occasional cases showing predominantly small cells (Figs. 96–30 and 96–31). Cell type has no prognostic relevance. A reactive background consisting of eosinophils, plasma cells, and macrophages may be present, in which case the diagnosis of Hodgkin lymphoma may be entertained. Immunophenotypic data cannot prove clonality as in B-cell lymphomas, but evidence of an aberrant T-cell phenotype supports a diagnosis of T-cell lymphoma. Molecular techniques to demonstrate clonal rearrangement of T-cell receptor genes can be helpful in confirming the diagnosis. Gene-expression profiling has helped to delineate biologic and prognostic groups within PTCL, not otherwise specified.⁵⁵ Angioimmunoblastic T-cell lymphoma is a mature T-cell lymphoma that typically presents with systemic symptoms and polyclonal hypergammaglobulinemia, and arises from a distinct subset of helper T cells, the follicular helper T cell.⁵⁶

Anaplastic large cell lymphoma (ALCL) represents a unique subtype of T-cell lymphoma, particularly common in children. ALCL can show significant morphologic variability but typically is composed of large pleomorphic cells characterized by the presence of “hallmark” cells with horseshoe- or kidney-shaped nuclei and a perinuclear eosinophilic region (Fig. 96–32).⁵⁷ Partial involvement of lymph nodes can be limited to the sinuses, with obliteration of nodal architecture in later stages. ALCL is characterized by uniform, strong expression of CD30 (Fig. 96–33). The majority of cases express one or more T-cell antigens and demonstrate clonal T-cell receptor gene rearrangement.⁵⁷ ALCL is divided into two entities based on the expression of anaplastic lymphoma kinase (ALK) (Fig. 96–34). ALK-positive ALCL is most often seen in the first 3 decades of life and has a favorable prognosis compared to ALK-negative ALCL.^58,59 Expression of ALK is the result of chromosomal translocations involving the ALK gene on chromosome 2p23, the most common translocation being the t(2;5)(p23;q35) involving the nucleophosmin gene on chromosome 5.⁶⁰ ALK-negative ALCL is recognized as a provisional entity that is distinct from ALK-positive ALCL and PTCL, not otherwise specified.⁶¹ Recently, additional genetic abnormalities have been identified in ALK-negative ALCL, involving DUSP22 and TP63 genes.^62,63

Other types of mature T/NK cell lymphomas are uncommon and include enteropathy-associated T-cell lymphoma (an aggressive T-cell lymphoma typically arising in the small bowel from a background of celiac disease) and extranodal NK/T-cell lymphoma, nasal type (an aggressive Epstein-Barr virus [EBV]-associated neoplasm commonly involving the nasal cavity). A detailed description of these specific lymphoma subtypes is beyond the scope of this chapter and can be found in the WHO classification.¹⁴

Hodgkin lymphoma consists of two distinct clinicopathologic entities: classical Hodgkin lymphoma (including four subtypes) and nodular lymphocyte predominant Hodgkin lymphoma (Chap. 97).

CLASSICAL HODGKIN LYMPHOMA

The neoplastic cell of classical Hodgkin lymphoma is the Reed-Sternberg cell, first described more than 100 years ago.^64,65 It is a large cell with two or more nuclei or nuclear lobes, each of which contains a large eosinophilic nucleolus (Fig. 96–35). The presence of Reed-Sternberg cells alone is insufficient for a diagnosis of Hodgkin lymphoma, because cells with similar morphology can be seen in a variety of non-Hodgkin lymphomas and benign reactive conditions.⁶⁶ For a diagnosis of Hodgkin lymphoma, diagnostic Reed-Sternberg cells must be found in an appropriate background consisting of a variable polymorphous reactive infiltrate of inflammatory and accessory cells.⁶⁷

Reed-Sternberg cells are derived from B cells in the vast majority of cases of classical Hodgkin lymphoma, as determined by clonal rearrangement of IGH genes.⁶⁸ However, Reed-Sternberg cells have lost most of their B-lineage antigens, including expression of immunoglobulin. Reed-Sternberg cells express CD30 in almost all cases of classical Hodgkin lymphoma and express CD15 in the majority (Figs. 96–36 and 96–37).⁶⁷ They typically are negative for CD45 (leukocyte common antigen) and positive for B-cell marker CD20 in 20 to 40 percent of cases, usually of variable intensity in a minority of cells. Classical Hodgkin lymphoma is associated with EBV in 20 to 40 percent of cases and is thought to play a role in the pathogenesis of these cases.⁶⁹ Reed-Sternberg cells express many cytokines and several members of the tumor necrosis factor receptor family (e.g., CD40, CD30).⁷⁰ The cytokines are thought to play a role in the recruitment of reactive infiltrate and to contribute to Reed-Sternberg cell proliferation and survival. The tumor necrosis factor receptor family members can be activated by ligands expressed by the surrounding reactive infiltrate, leading to proliferation and survival.

The most common subtype of classical Hodgkin lymphoma is the nodular sclerosis variant. The variant is characterized by the presence of broad collagen bands dividing the tumor into nodules and by the presence of “lacunar” cells, mononuclear Reed-Sternberg variants that typically show retraction artifact so that the cells appear to be in lacunae (Fig. 96–38). These cells are found within a reactive infiltrate that typically includes prominent eosinophils and lymphocytes.

The second most common subtype is the mixed cellularity variant, which is characterized by Reed-Sternberg cells in a mixed inflammatory background without the broad collagen bands seen in nodular sclerosis (Fig. 96–39). Mixed cellularity cases are more commonly associated with EBV compared to the nodular sclerosis variant.

The lymphocyte-rich and lymphocyte-depleted subtypes of classical Hodgkin lymphoma are the least common, each representing approximately 5 percent of all cases. The lymphocyte-rich variant has a small number of Reed-Sternberg cells in a background of small lymphocytes with absent or rare eosinophils and neutrophils, typically in a nodular pattern. It is easily confused with nodular lymphocyte predominant Hodgkin lymphoma, so immunohistochemical stains to determine the immunophenotype of the Reed-Sternberg cells is required to make the distinction.^71,72 It can rarely have a diffuse growth pattern.

In the past, the lymphocyte-depleted variant had been divided into reticular and diffuse fibrosis types. The diffuse fibrosis variant is characterized by a hypocellular infiltrate with prominent diffuse nonbirefringent sclerosis accompanied by rare Reed-Sternberg cells and a minor reactive inflammatory component. The reticular variant showed an increased number of large atypical cells, commonly with bizarre multinucleated cells, with a minor reactive component. It now is recognized that the vast majority of these cases are cases of ALCL or DLBCL, and as such the diagnosis of the reticular variant of lymphocyte-depleted Hodgkin lymphoma is rare and should be made only in the presence of definitive supportive immunophenotypic data.

NODULAR LYMPHOCYTE PREDOMINANT HODGKIN LYMPHOMA

Nodular lymphocyte predominant Hodgkin lymphoma has several pathologic and clinical features that are distinct from classical Hodgkin lymphoma.⁷³ The malignant cells are known as lymphocyte predominant (LP) cells, large cells with a single nucleus that contains multilobated or folded features. They often are referred to as “popcorn” cells because they resemble popped kernels of corn (Fig. 96–40). Nucleoli typically are smaller than the nucleoli seen in classical Reed-Sternberg cells. They differ from Reed-Sternberg cells in classical Hodgkin lymphoma in that they retain expression of CD45 and B-lineage markers (CD20, immunoglobulin) and are negative for CD15 and CD30 (Fig. 96–41).⁷⁴ As the name implies, the cells have a complete or partial nodular architectural pattern with a background consisting primarily of lymphocytes. Histiocytes are also a common feature, but neutrophils and eosinophils are absent or rare.