Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
Classical Hodgkin's lymphoma (CHL) is a clonal, malignant lymphoproliferation originating from germinal-center B cells. In contrast to most other lymphomas, the malignant cells usually represent only a small minority (0.1% to 2%) of the total cellular population of involved tissues. A histopathologic diagnosis of CHL is based on the identification of diagnostic Reed-Sternberg (RS) cells in an appropriate inflammatory background. Although many cases of CHL can, in principle, be diagnosed on the basis of morphology alone, current diagnostic criteria include the characteristic immunophenotype of the neoplastic population. RS cells and variants express the CD30 and CD15 antigens in the majority of cases, lack the common leukocyte antigen CD45, and show an inconsistent and heterogeneous expression of lineage-specific lymphoid markers.
In contrast to non-Hodgkin's lymphomas (NHLs), the classification of Hodgkin's lymphoma (HL) has remained remarkably constant ( Box 28-1 ). Notwithstanding the significant progress in delineating the antigenic profile and the origin of the neoplastic cells of HL, the majority of cases are classified much as they were 30 years ago, after the development of the Rye classification. This underlines the paramount importance of morphology for the correct diagnosis of this neoplasm.
Nodular lymphocyte–predominant Hodgkin's lymphoma
Classical Hodgkin's lymphoma
Nodular sclerosis classical Hodgkin's lymphoma
Lymphocyte-rich classical Hodgkin's lymphoma
Mixed-cellularity classical Hodgkin's lymphoma
Lymphocyte-depleted classical Hodgkin's lymphoma
The change in terminology from Hodgkin's disease to Hodgkin's lymphoma was first proposed in the Revised European American Lymphoma (REAL) classification and reflects a better understanding of the nature and histogenesis of this lymphoproliferation. After decades of controversy as to whether HL represents an infectious, immunologic, or neoplastic disorder, the success of radiotherapy and multimodal chemotherapy in treating HL confirmed its neoplastic nature in a pragmatic way. Finally, the molecular analysis of single isolated RS cells firmly established the clonality as well as the B-cell origin of HL and its derivation from germinal-center B cells, which, together with the establishment of Hodgkin cell lines, brought significant insights into the pathogenesis of the disease. In diagnostic terms, the last 2 decades have led to clarification concerning the separation of diffuse large B-cell lymphoma (DLBCL) variants from CHL and the recognition of true biologic gray-zone cases between the two entities, which were included in the fourth edition of the WHO Classification of Tumours of Hematopoietic and Lymphoid Tissues as B-cell lymphoma unclassified, with features intermediate between DLBCL and CHL. On the other hand, nodular lymphocyte–predominant HL (NLPHL) was defined more precisely and separated from lymphocyte-rich classical HL (LRCHL), which had first been introduced as a new subtype of CHL in the third edition of the WHO classification. In addition to its unique inflammatory background, CHL today is separated from other B-cell neoplasms (including NLPHL) based on the extensive loss of the B-cell transcriptional program in the neoplastic cells, as evidenced by the loss or downregulation of most B-cell–specific transcription factors and surface antigens. This biologic key feature also serves as a useful tool in distinguishing CHL from morphologically similar entities.
HL accounts for approximately 15% to 25% of all malignant lymphomas. With current diagnostic criteria, approximately 95% of HLs fall into the CHL category; the remaining cases are NLPHLs. The age-adjusted annual incidence rate of HL in Western countries is approximately 2 to 4 per 100,000 population. In recent decades, a decrease in incidence has been observed among older adults; however, this can be attributed mainly to the misdiagnosis of NHL as HL in previous years. In contrast, there has been a slight increase of the nodular sclerosis subtype in young adults. The incidence of CHL is higher in affluent, industrialized nations than in developing countries.
In industrialized countries, CHL shows a bimodal age distribution: the first, higher peak occurs in early adulthood (15 to 35 years of age), with a second peak in those older than 55 years. CHL is rare in children and is an exceptional occurrence in those younger than 3 years. The overall male-to-female ratio is approximately 1.5:1. Although males predominate in childhood cases and among older adults, the male-to-female ratio is balanced or even slightly reversed in the early adulthood peak. The distribution of histologic subtypes varies with age. Although the nodular sclerosis subtype predominates in young adults, especially in females, there is a higher percentage of the mixed-cellularity subtype in children and older patients. Developing countries show a distinct epidemiologic pattern: the first incidence peak occurs in childhood, with a predominance of the mixed-cellularity subtype, and there is no peak in young adulthood.
Nodular sclerosis CHL in young adults is associated with a higher socioeconomic status and smaller family size, factors thought to favor a delayed exposure to common childhood viruses. Paradoxically, the early adulthood cases show the lowest frequency of Epstein-Barr virus (EBV) positivity, the only infectious agent so far shown to be associated with CHL. EBV is most often identified in CHL in children and older adults and in association with human immunodeficiency virus (HIV) infection.
Familial cases of HL have been reported frequently, and siblings of HL patients have a moderately increased risk for developing the disease. A markedly increased risk for CHL was reported for monozygotic twins in a single study. There seems to be a weak but consistent increase in the relative risk for CHL among certain human leukocyte antigen (HLA) types. Interestingly, carriers of HLA-A*02, known to be involved in the cytotoxic response to latency proteins of EBV, are underrepresented in EBV-positive CHL.
The risk for developing CHL is increased in patients with some types of immunodeficiency. HIV-infected persons have a 6-fold to 20-fold increased risk for developing CHL. In the setting of acquired immunodeficiency syndrome (AIDS), CHL shows a predominance of unfavorable histologies (mixed-cellularity and lymphocyte-depleted) and advanced-stage disease, and it is almost universally associated with EBV. The incidence of CHL seems to increase in the first months after initiation of highly active antiretroviral therapy.
A significantly increased risk for CHL has been described in recipients of solid organs and allogeneic bone marrow grafts. In either setting, CHL occurs late after transplantation, is almost always EBV positive, and is mainly of the mixed-cellularity subtype. The clinical presentation and outcome of CHL in allograft recipients is similar to that in non-immunocompromised patients (see later).
The cause of CHL is still an open question, but owing to the unique epidemiologic and clinical features of the disease, an infectious cause has long been suspected. Detection of EBV in the neoplastic cells of a significant proportion of CHLs has confirmed the involvement of an infectious agent. However, as evidenced by EBV-negative cases, EBV is not mandatory for the development of the disease and probably represents only a cofactor in positive cases.
After the recognition of EBV as the cause of infectious mononucleosis and endemic Burkitt's lymphoma, several lines of evidence soon pointed to a possible relationship with CHL. Tissues involved by infectious mononucleosis frequently contain cells that are morphologically indistinguishable from RS cells. Individuals with a history of infectious mononucleosis have an approximately threefold increased risk for developing CHL, and increased antibody titers or an altered antibody pattern in response to EBV infection has been demonstrated in patients with CHL. The first direct evidence of the possible involvement of EBV in CHL was reported by Poppema and colleagues, who detected Epstein-Barr nuclear antigen (EBNA) by immunohistochemistry in the lymph node of a patient with chronic EBV infection and a “CHL-like” morphology. Subsequently, clonal EBV was demonstrated in tissues involved by CHL with Southern blot analysis. DNA in situ hybridization confirmed that EBV nucleic acids were indeed present in the neoplastic cells of a significant proportion of cases. Currently, immunohistochemistry for the EBV latent membrane protein-1 (LMP-1) and non-radioactive in situ hybridization for EBV-encoded early RNAs (EBERs) are the methods of choice for the detection of EBV in routinely fixed, paraffin-embedded tissues. In situ hybridization for EBERs, short non-translated RNA molecules that are abundantly expressed in viral latency (approximately 10 6 to 10 7 copies/cell), is regarded as the most sensitive method, although LMP-1 immunohistochemistry provides almost equivalent results in CHL. Conventional polymerase chain reaction (PCR) is not recommended because it does not allow the visual identification of infected cells. In healthy virus carriers, approximately 1 to 50 in 10 6 B cells are latently infected with EBV. Consequently, in situ hybridization reveals that in a portion of the PCR-positive cases of CHL, EBV is present only in latently infected, small “bystander” lymphocytes that are not part of the malignant clone and do not express LMP-1. EBV-positive CHL shows a latency type II pattern of viral gene expression. In addition to EBERs and LMP-1, the neoplastic cells express EBNA1 and LMP-2A and LMP-2B but lack EBNA2. Latency in CHL is tightly controlled, and there is usually no evidence of lytic infection, even in immunosuppressed individuals. EBV is clonal in the neoplastic cells of HL, as shown by Southern blot analysis of the terminal repeat region of the EBV genome. This indicates that the infection takes place before clonal expansion and implicates a direct role for EBV in the transformation process. The EBV genome is present in episomal (non-integrated) form, and each infected cell contains multiple copies. In EBV-associated HL, tumor cells in multiple involved anatomic sites contain the virus, as do tumor recurrences.
LMP-1, which is strongly expressed by the neoplastic cells in HL, is the only viral protein with proven oncogenic properties. LMP-1 induces an array of phenotypic changes in infected cells, including upregulation of activation antigens and antiapoptotic genes, as well as induction of various cytokines; it can also induce an RS-like phenotype in germinal-center B cells. LMP-1 functions as a constitutively activated tumor necrosis factor (TNF) receptor and induces a variety of downstream effects, such as constitutive activation of nuclear factor-κB (NF-κB) and JAK/STAT signaling pathways and upregulation of antiapoptotic genes. Of note, virtually all cases of CHL with so-called crippling IGH mutations are EBV positive, emphasizing the importance of EBV-related genes for rescuing CHL precursor cells from apoptosis. However, there is no clear-cut correlation between EBV positivity and tissue expression of genes known to be regulated by LMP-1 in vitro, and the gene-expression profiles of HRS cells are very similar irrespective of EBV status, indicating that the main impact of EBV infection of CHL pathogenesis may lie in the early stage of disease development and in the interaction with the immune system.
Two common strains of EBV, types A and B (or 1 and 2), exist in a worldwide distribution. EBV type A transforms lymphocytes more efficiently and is found in the majority of CHL cases in non-immunosuppressed patients. Type B is encountered more often in HIV-associated CHL but has also been found in a higher proportion of cases from Latin America, indicating that geographic variations in the prevalence of viral strains can influence their distribution in EBV-associated lymphomas.
In addition to these two viral strains, a number of variations of the EBV genome have been described. Among these, deletions in the carboxyterminal part of the LMP-1 gene have received the greatest attention. Although initial findings suggested an increased occurrence in cases with aggressive morphologic features, further studies demonstrated that the deletion variant is not associated with any specific histologic or prognostic features; its frequency in EBV-associated CHL reflects its prevalence in healthy carriers from the same geographic region, with the possible exception of higher frequencies of deletion variants in HIV-associated CHL. Of note, CHL cases with EBV type B contain the LMP-1 deletion very frequently, indicating that this deletion might be necessary for the transforming capacity of this less virulent strain.
The frequency of the association between EBV and CHL in non-immunosuppressed patients is influenced by histologic subtype, age at presentation, geographic and ethnic origin, and possibly socioeconomic factors. Mixed-cellularity CHL is EBV positive in approximately 75% of cases, whereas EBV positivity in the nodular sclerosis subtype ranges from 10% to 25%. Reported rates for both lymphocyte-depleted and lymphocyte-rich CHL are variable, probably reflecting differences in diagnostic criteria. However, when stringent criteria are used, lymphocyte-depleted CHL is usually EBV positive. Children and older adults are more likely to have EBV-positive CHL, as are males. Developing countries generally show a higher incidence of EBV-associated disease, reaching almost 100% in childhood cases. A history of infectious mononucleosis and atypical prediagnosis EBV serology are associated with an increased risk for EBV-positive, but not EBV-negative CHL, further supporting the pathogenetic role of EBV.
Based on observations in EBV-positive Burkitt's lymphoma cell lines, it has been proposed that the virus might be lost during the malignant progression of CHL once it is no longer required for propagation of the malignant clone. Although a single study reported detection of defective, rearranged EBV in a fraction of CHLs that were EBER negative, results from other groups have not provided evidence for this “hit and run” theory. Searches for other candidate viruses, such as human herpesvirus 6 or 8, have failed to provide evidence for their involvement in the pathogenesis of CHL.
Clinically, CHL manifests first in lymph nodes in about 90% of cases, usually as a slowly growing, painless mass. Cervical (75%), axillary, and inguinal nodes are the most frequently involved sites. Asymptomatic patients are often diagnosed with mediastinal disease detected on a routine chest radiograph. Symptoms related to specific organ involvement, such as superior vena cava syndrome, bone pain, or neurologic symptoms, can occur. Retroperitoneal lymphadenopathy and splenic involvement are frequent, whereas extra-axial lymph nodes (mesenteric, perigastric, epitrochlear, preauricular, popliteal nodes) are rarely involved. Bone marrow involvement is relatively rare in CHL, occurring in approximately 5% of non-immunosuppressed patients according to recent large series. Disease manifestations on both sides of the diaphragm and B symptoms indicate a higher risk for bone marrow involvement.
The distribution of disease correlates with histologic subtype. The nodular sclerosis type typically occurs above the diaphragm, most frequently involving the lower cervical, supraclavicular, and mediastinal nodes and contiguous structures. Patients frequently show an enlarged mediastinal silhouette and present with bulky mediastinal disease (greater than one third of the intrathoracic diameter) in approximately 50% of cases. The spleen and bone marrow are involved in 10% and 3% of cases, respectively. The mixed-cellularity subtype of CHL presents more frequently in stages III and IV and with B symptoms. It occurs more often below or on both sides of the diaphragm; there is splenic involvement in 30% of cases and bone marrow infiltration in 10%. Bulky mediastinal disease is uncommon. Owing to its recent delineation as an independent subtype, relatively few clinical data are available for LRCHL. It has similarities with NLPHL in terms of clinical symptoms and disease distribution. Seventy percent of patients with LRCHL present with stages I and II, and bulky disease and B symptoms are infrequent. Because of its rarity and changes in diagnostic criteria, clinical features of lymphocyte-depleted CHL are not well described. A predominance of older adult patients with higher-stage disease and predominant involvement of abdominal organs and bone marrow has been reported. CHL in patients with HIV infection frequently presents as advanced disease and involves unusual sites. The bone marrow is frequently infiltrated and occasionally represents the primary diagnostic manifestation of the disease.
Approximately 30% to 40% of patients with CHL present with B symptoms. Although B symptoms are more frequent in advanced stages of disease, they can also occur in early stages, possibly due to inflammatory cytokines produced by the tumor. The so-called cyclic Pel-Ebstein type of fever is rare. Other symptoms include generalized pruritus and occasionally pain in involved nodes upon alcohol ingestion.
Laboratory findings are mostly non-specific and include leukocytosis, elevated erythrocyte sedimentation rate, and increased lactate dehydrogenase. Eosinophilia can be observed in approximately 20% of patients, and lymphopenia is present in advanced disease stages. However, most patients with CHL exhibit demonstrable defects in cell-mediated immunity, regardless of stage. A decrease in CD4-positive cells in the peripheral blood is already detectable at an early stage. T lymphocytes from HL patients show a weaker response to T-cell mitogens and a decrease in cytokine production upon stimulation. Clinically, this anergy can manifest as an increased susceptibility to infections and a lack of reactivity in the tuberculin skin test. It is still unclear whether these immune abnormalities are preexistent and may contribute to disease development or are secondary phenomena, possibly due to immunosuppressive cytokines secreted by the neoplastic population.
Although CHL is almost always a node-based lymphoma, practically any site and organ of the body can be involved during the course of the disease or, much rarer, as a primary manifestation. Although the spleen, bone marrow, and liver are frequently involved in advanced disease, isolated involvement of these organs is rare. Thymic involvement is frequent in mediastinal disease. Rarely, if lymph node involvement is not apparent, CHL can simulate other neoplasms of the thymus radiologically, prompting surgical removal of the thymic mass. The affected thymus gland frequently shows cystic degeneration. CHL is rare in mucosa-associated lymphoid tissue, including the tonsils and Waldeyer's ring. Apart from the liver, the lung is probably the most frequently involved non-lymphoid organ, usually as an extension of mediastinal disease. Primary pulmonary disease is rare. The same is true for primary gastrointestinal CHL. Further, many cases from earlier series probably represent NHLs such as polymorphic large B-cell NHL or enteropathy-type T-cell lymphoma. Interestingly, primary CHL of the gastrointestinal tract was initially reported in association with a history of inflammatory bowel disease or immunosuppression. These cases would today probably be classified as EBV-associated lymphoproliferation with Hodgkin-like features, specifically as EBV-associated mucocutaneous ulcer. Many other extranodal primary sites of CHL have been reported, mainly in anecdotal form. An especially problematic organ is the skin, owing to the morphologic and phenotypic overlap between CHL and CD30-positive lymphoproliferative disorders of the skin, including cutaneous anaplastic large-cell lymphoma (ALCL) and lymphomatoid papulosis. Only a few well-documented cases of apparently primary cutaneous CHL have been reported. Central nervous system involvement with CHL is very rare but may occur as primary disease.
Become a Clinical Tree membership for Full access and enjoy Unlimited articles
If you are a member. Log in here