Pediatric Lymphoma

Mechanisms of Transformation.

EBV has been associated with the pathogenesis of CHL, which shows EBV infection of the Reed-Sternberg cell in 40% of the cases in the Western world and 90% in Central America. Importantly, EBV can immortalize human B cells in vitro, and Reed-Sternberg cells are clonally infected by EBV, which suggests that viral infection occurs early in the pathogenesis of the disease. EBV positive Reed-Stenberg cells express EBNA1, LMP1, and the LMP2a viral gene, which is characteristic of the viral latency II stage. Mechanistically, LMP1 mimics an active CD40 receptor and can activate the nuclear factor κB (NF-κB) pathway, which is constitutively active in Hodgkin lymphoma (see later). LMP2, a second viral gene product, mimics B-cell receptor signaling and might play a role in rescuing Reed-Sternberg cells from apoptosis with immunoglobulin-crippling mutations. Cytogenetically, Reed-Sternberg cells often show aneuploidy and chromosomal abnormalities, with 20% of the cases harboring chromosomal translocations involving the immunoglobulin loci, a hallmark of many B-cell lymphomas.

The most prominent molecular abnormality found in Hodgkin lymphoma is the constitutive activation of the NF-κB pathway. NF-κB functions as an important survival-signaling pathway in B cells in response to the activation of members of the tumor necrosis factor (TNF) receptor family (CD30, CD40). In Hodgkin lymphoma, somatic mutations in different elements of the NF-κB pathway keep it constitutively activated. Thus the REL gene, which encodes an NF-κB component, is amplified in the of 30% of the cases. Gains of NF-κB–inducing kinase (NIK), a positive regulator of the alternative NF-κB pathway, are also commonly found in Reed-Sternberg cells, and translocations and amplifications resulting in overexpression of the BCL3 gene in Reed-Sternberg cells has been described. Moreover, mutations of the gene-encoding NF-κB inhibitors IKBA and IKBE are found in 10% to 20% of cases of Hodgkin lymphoma. Most significantly TNFAIP3, which encodes A20, a ubiquitin-modifying enzyme involved in the negative regulation of NF-κB signaling, harbors loss of function mutations and deletions in almost 45% classic Hodgkin lymphomas. Notably, TNFAIP3 mutations and deletions are more common in EBV-negative tumors and are rarely found in lymphocyte-predominant Hodgkin lymphomas. Finally, activation of TNF signaling by the microenvironment and activation of NF-κB by the LMP1 EBV oncogene may also contribute to aberrant NF-κB signaling.

Gene amplifications involving the JAK2 locus have been described in about 20% of CHL cases, suggesting a role for JAK2 activation in the pathogenesis of this disease. In addition, rearrangements of JAK2 including a recurrent SEC31A–JAK2 translocation have been described in rare cases of Hodgkin lymphoma. Moreover, loss of function mutations in SOCS1 , a negative regulator of Janus kinase (JAK)–signal transducer and activator of transcription (STAT) signaling are present in about 50% of classic and lymphocyte-predominant Hodgkin lymphomas and are associated with increased levels of JAK signaling.

Translocations involving the major histocompatibility complex (MHC) class II transactivator gene CIITA , resulting in impaired MHC class II expression, are found in 15% of CHL cases and seem to be associated with unfavorable prognosis. In contrast to CHL, the malignant cells of the lymphocyte-predominant type of Hodgkin lymphomas are always EBV-negative, and there is little information on the molecular lesions implicated in the pathogenesis of this group. Chromosomal translocations involving the BCL6 oncogene can be detected in 30% of cases. BCL6 is a zinc-finger transcription repressor that functions as a master regulator of germinal-center formation and drives B-cell transformation. Thus Bcl6 -null mice fail to generate germinal centers in response to immunization, and deregulated expression of Bcl6 induces B-cell lymphomas in mice. Mechanistically, aberrant expression of BCL6 in lymphoma hijacks the role of BCL6 in the control of B-cell activation, differentiation, deoxyribonucleic acid (DNA) damage response, cell-cycle arrest, and apoptosis. Finally, about 40% of lymphocyte-predominant Hodgkin lymphoma tumors show mutations in SOCS1, but mutations in TNAIP3 and NFKBIA are rare despite strong NF-κB activity.

Cellular Microenvironment: Cytokines and Chemokines.

The cellular and cytokine microenvironment surrounding the lymphoid cells in Hodgkin lymphoma plays an essential role in the pathogenesis of this disease. In particular, the expression of soluble factors and their receptors by malignant cells and the reactive microenvironment seems not only to mediate the inflammatory characteristics observed in the histology of Hodgkin lymphoma, but also to contribute to the proliferation and survival of the malignant clone. Thus the cellular microenvironment supports and is supported by a network of cytokines secreted in autocrine and paracrine loops that are essential for the proliferation of Reed-Sternberg cells and the maintenence of a favorable inflammatory environment rich in regulatory T cells (Tregs) and eosinophils.

Both interleukin (IL) 13 and the IL-13 receptor IL-13RA1 are expressed in Hodgkin lymphoma and constitute an important autocrine loop for Reed-Sternberg cells. Other cytokines expressed in Hodgkin lymphoma and thought to influence survival of these cells include IL-4, IL-6, IL-7, IL-9, and IL-15.

The vast majority of cells in a tumor biopsy of Hodgkin lymphoma are not malignant cells but represent an inflammatory-like cellular infiltrate composed mainly of CD4+ T lymphocytes intermixed with macrophages, eosinophils, plasma cells, and fibroblasts. Most lymphocytes present in this infiltrate are Tregs, which play an important role in protecting the tumor's Reed-Sternberg cells from cytotoxic T cells involved in antitumor immune surveillance.

In contrast with the abundance of Tregs, Th1 CD4+ T cells and CD8+ cytotoxic T cells are rare in Hodgkin lymphoma biopsies and are not detected in the immediate proximity of Reed-Sternberg cells. The recruitment of cells involved in immune tolerance and the exclusion of lymphoid cells responsible for antitumor immune responses are explained by the expression of Reed-Sternberg cells of pro-Th2–associated cytokines, such as IL-4 and IL-13, and anti-Th1 or CD8 cytokines, such as IL-10 and transforming growth factor (TGF) β. The secretion of these immunosuppressive cytokines creates an immune-privileged microenvironment, allowing the tumor cells to avoid immune surveillance and T-cell–mediated apoptosis.

One of the most prominent features of Reed-Sternberg cells is the expression of CD30. Although coexpression of CD30 and its respective receptor CD153 initially suggested an autocrine mechanism promoting the proliferation of Reed-Sternberg cells via NF-κB, it is now well established that the activation of CD30 is primarily a CD153-independent process. Two additional members of the TNF receptor family, CD40 and receptor activator of nuclear factor κB (RANK), are expressed in Reed-Sternberg cells. Activation of CD40 seems to be mediated by the expression of CD40 ligand (CD40L) on T lymphocytes surrounding the Reed-Sternberg cell. Importantly, soluble CD40L induces proliferation and blocks CD95-induced apoptosis in Reed-Sternberg cells. Other factors that help rescue Hodgkin and Reed-Sternberg cells from an immunologic attack include the expression of the PD1 ligand and secretion of IL-10, TGF-β, and galectin 1 by the Reed-Sternberg cell, as well as inhibition of cytotoxic T cells by regulatory T cells (Tregs). Activation of RANK and osteoprotegerin, a member of the TNF-receptor superfamily, is triggered by the autocrine expression of RANK ligand (RANKL) in Reed-Sternberg cells. Activation of RANK promotes NF-κB activation and contributes to the maintenance of an inflammatory microenvironment by promoting interferon gamma (IFN-γ) and IL-13 secretion. Overall, the activation of TNF receptors by different mechanisms seems to play an important role in promoting the survival of the Reed-Sternberg cell.

Prognostic Factors/Stratification of Treatment.

As with other malignancies, prognostic factors are useful as tools for defining risk groups for stratification of treatment intensity. For patients with Hodgkin lymphoma this translates into both the choice of chemotherapeutic regimen and the use of radiotherapy. The stage of disease is a strong prognostic factor and widely used for risk stratification that has been used in all major cooperative groups. Apart from the stage of disease, a number of factors have been reported to be associated with the risk of treatment failure, including the presence of B symptoms, histologic subtype, bulky disease, gender, anemia, ESR, and evidence of latent EBV infection, some of which have lost significance with improvements in therapy. The prognostic value of these variables has not been consistent across studies, in part related to treatment prescribed. For example, mediastinal bulky disease—defined as a mass larger than one third of the maximum chest diameter—was associated with increased risk for relapse in some studies but not in the German-Austrian trial DAL-HD90. However, in this protocol patients with a larger residual mass at the end of chemotherapy received a boost dose in addition to the involved field irradiation. Similarly, male gender had an adverse impact on outcome in that trial. For boys, however, procarbazine was partially replaced by etoposide, and girls received a full dose of procarbazine. Race and ethnicity have been identified as risk factors in some pediatric cancers. In a single-center retrospective review, African-American children had an inferior event-free survival (EFS) rate but an identical overall survival (OS) as compared with white children.

The risk-group definitions used in selected clinical trials are described in Table 53-4 . In general in addition to stage, disease bulk and presence of B symptoms are the factors used. The kinetics of disease response to chemotherapy, measured by CT and/or PET scan is used in many cooperative group trials as an important additional tool for the stratification of treatment intensity including the use of radiation (see below).

Treatment and Outcomes.

In the 1960s extended-field radiation with doses of 35 to 44 Gy were shown to be curative in a significant portion of patients with Hodgkin lymphoma. In children, the negative sequelae of high-dose extended-field radiation quickly became apparent. Soon thereafter the MOPP combination was created by Devita and coworkers to treat patients with Hodgkin lymphoma. The basic rationale was to combine single drugs with proven efficacy but different mechanisms of action and resistance that had few overlapping toxicities to maximize antitumor effect and limit side effects caused by use of moderate doses of the individual drugs. The second important combination chemotherapy regimen for the treatment of Hodgkin lymphoma was doxorubicin (Adriamycin), bleomycin, vinblastine, and dacarbazine (ABVD), which was developed by Bandanna and coworkers. Variations of these two regimens have formed the cornerstone of almost all subsequent chemotherapy regimens.

Combined-modality therapy using both chemotherapy and radiation has emerged as the standard of care for most children and adolescents diagnosed with Hodgkin lymphoma. The use of six cycles of MOPP chemotherapy combined with involved-field radiotherapy (IFRT) was pioneered by the Stanford investigators in children with pathologically staged disease. This pivotal trial has served as the model for combined-modality treatment programs in children.

Subsequent studies evaluated regimens with modifications incorporated designed to minimize risks of late effects while maintaining efficacy. Changes have included both the substitution of drugs the use of which is associated with lower risks of specific long term toxicities as well as limiting the cumulative dose of drugs that are associated with risks of late sequelae in a dose dependent manner. Some modifications have been gender specific. Table 53-5 gives an overview of chemotherapy combinations. With most of these multiagent combinations, comparable high disease-free survival rates have been achieved.

The balance of efficacy and late risk of the combined-modality of chemotherapy and radiotherapy adopted by the German-Austrian study group has been investigated over a series of trials ( Table 53-6 ). Modifications investigated have included reduction of the cumulative dose of alkylating agents and replacement of mechlorethamine of MOPP with doxorubicin (vincristine [Oncovin], prednisone, procarbazine, and doxorubicin [OPPA]) and cyclophosphamide (cyclophosphamide, vincristine [Oncovin], procarbazine, and prednisone [COPP]); reduction of radiotherapy (volume, dose, and number of patients who receive radiotherapy); and replacement of procarbazine by etoposide for boys. The seventh of the sequential studies built on findings of the previous six trials. In this study the primary aim was to evaluate modifications of therapy to maintain good outcomes for boys but to do so with less gonadotoxic therapy. In this study procarbazine was replaced with higher dose etoposide in the first two cycles and with dacarbazine in the subsequent two to four cycles. The outcomes for boys and girls on this study were superimposable, supporting the efficacy of this regimen.

The second strategy that is of critical importance in trying to maximize efficacy and minimize toxicity is the stratification of treatment intensity based on the patient's risk for relapse. Cooperative group trials of pediatric and adolescent Hodgkin lymphoma have generally divided patients into low-, intermediate-, and high-risk groups (see Table 53-4 ). In addition to stratification by stage, studies have also evaluated modifications of therapy based on early disease response. Many reports from these therapeutic studies are confined to a distinct risk group of patients. Comparisons among trials confined to patient subsets need to be done with caution, because the risk group definitions vary both based on the patient characteristics at the time of initial staging and the criteria of disease response.

Table 53-7 summarizes treatment strategy and results of therapeutic studies in the patient subgroup defined as low risk. Patients at low risk in general are those with stage I and II disease without B symptoms and without bulky disease. These patients have been shown to have excellent outcomes with sequential trials using two to four cycles of multiagent chemotherapy and low-dose involved-field radiation. These regimens have been designed to minimize exposure to anthracyclines, alkylators, and epipodophyllotoxins in order to minimize risks of cardiac dysfunction, infertility, and secondary acute myelogenous leukemia (AML), respectively.

Table 53-8 summarizes the treatment strategy and results of selected therapeutic studies in patients considered to have intermediate- or high-risk disease. Chemotherapy for these groups generally consists of combinations of ABVD and MOPP or hybrid regimens. Most regimens contain four to six cycles of chemotherapy followed by involved-field radiation with some including high-dose radiation to bulky or residual sites of disease.

Prognostic information for patients with Hodgkin lymphoma is based on information available at the time of the child's initial evaluation. Additionally the importance of the kinetics of disease response to chemotherapy has been shown to be a predictive factor for this disease, as it has been for pediatric leukemias. The premise is that those with more robust responses to initial chemotherapy have more favorable disease and that responsiveness can be included in the factors that are used to decide the intensity of therapy, including whether irradiation is indicated. The definition of a favorable response to chemotherapy, including the timing of assessment of response, the intensity of the chemotherapy regimen used to get the response, and the tools used to measure response have varied widely between study groups and over time. Table 53-9 describes the criteria for disease response evaluations in recent cooperative group trials.

Cross-sectional imaging can define response based on a percentage decrease in size. In Hodgkin lymphoma, a significant proportion of patients have residual masses at the time of disease response assessment and at the end of therapy. Functional imaging methods such as gallium scanning and PET are able to assess metabolically active tissue associated with residual disease ( Fig. 53-5 ). A series of studies primarily performed in adults with Hodgkin lymphoma where therapy was not changed based on interim radiographic results have shown that PET scans performed after two cycles of chemotherapy are highly predictive of outcome. The accuracy of PET to predict treatment outcome is higher early in the course of treatment as compared with the end of chemotherapy. Both the German-Austrian study GPOH-HD2002 and the Children's Oncology Group (COG) studies of intermediate- and high-risk disease addressed the strength of early PET obtained after two cycles of chemotherapy; however the results are still pending.

Several cooperative group studies have focused on modifications of regimens to avoid gender-specific toxicities. The COG evaluated a strategy of stratifying patients with high-risk disease based both on response and gender. All patients received escalated bleomycin, etoposide, doxorubicin (Adriamycin), cyclophosphamide, vincristine, procarbazine, and prednisone (BEACOPP) therapy for 4 cycles. Patients with a rapid early response, defined as a 70% tumor reduction and having a negative gallium scan after four cycles, were then assigned to subsequent therapy based on gender. Boys received two additional cycles of ABVD designed to minimize risks of infertility and involved field radiation. Girls received four cycles of COPP/doxorubicin (Adriamycin), bleomycin, and vinblastine (ABV) and no radiation in order to avoid the risks of secondary breast cancer. Those with disease that met the criteria for slow early response received an additional 4 cycles of BEACOPP. Seventy-seven percent of girls avoided radiotherapy, whereas 68% of boys avoided alkylating agents during consolidation. The EFS and OS rates remained excellent.

The German GPOH-HD2002 study also evaluated sex-based modifications of therapy but did not use patient response for treatment allocation. Boys received two cycles of dose-intensified vincristine, etoposide, prednisone, and doxorubicin (Adriamycin; OEPA) therapy followed by two (for those with intermediate-risk disease) or four (for those with high-risk disease) cycles of cyclophosphamide, vincristine (Oncovin), prednisone, and dacarbazine (COPDAC) designed to attempt to limit gonadal toxicity. Girls received the same number of cycles but of OPPA followed by COPP. EFS and OS rates did not differ between the two sexes with the regimens employed.

The Pediatric Oncology Group (POG) evaluated response-based modifications of therapy without modifications for sex in patients with intermediate- and high-risk disease. In this study patients were evaluated for response after three cycles of therapy. Those with a rapid early response defined as a greater-than-70% reduction in disease and gallium negative, received 21 Gy IFRT. Those who were slow responders received an additional two cycles followed by radiation. With this strategy the 5-year EFS rate was the same in both groups, and the OS rate was 95%.

Whether radiotherapy can be omitted from treatment for at least some children and adolescents with Hodgkin lymphoma is one of the important questions in recent and ongoing clinical trials. Three cooperative group trials in the low-risk group each evaluated the removal of IFRT in a subset of patients with largely successful results. In the Stanford-St.Jude-Boston protocol, patients with a complete response, defined as greater than 75% reduction of the sum of the perpendicular of all lesions after two cycles of vinblastine, doxorubicin (Adriamycin), methotrexate (methotrexate), and prednisone (VAMP) chemotherapy were nonrandomly assigned to no radiation. All patients received four cycles of VAMP chemotherapy. Fifty three percent of patients met criteria for complete response. The 2-year EFS rates of this group were no different from those who did receive radiation, at 89% and 92% respectively. The GPOH-HD 2002 study also nonrandomly assigned patients at low risk with a complete response after two cycles, defined as more than 95% volume reduction, to no radiation. Thirty three percent of patients met this criterion, did not receive radiation, and had an outcome identical to those who did receive IFRT, with EFS rates at 5 years of 93% and 91% respectively. The Children's Cancer Group (CCG) study 5942 randomly assigned patients who obtained a complete remission (CR), defined as 70% mass reduction of tumor volume and negative gallium scan, to radiation or no radiation after four cycles of COPP/ABV therapy. The EFS rates for randomized patients at 3 years were significantly different at 97% versus 91%; however, the OS rate at 3 years in both groups was 100%.

Omission of radiation has also been investigated in subgroups of patients with intermediate- and high-risk disease with mixed results. In the POG trial 8725, patients with stages IIB, IIIA2, IIIB, and IV were randomized to receive eight alternating cycles of MOPP/ABVD chemotherapy followed by total nodal irradiation (21 Gy) or no radiotherapy in cases of CR. There was no difference in EFS or OS between patient groups who did or did not receive radiation. In the GPOH-HD95 trial patients who achieved a complete response received no further therapy, whereas those who did not received 20 to 35 Gy of IFRT. In the intermediate- and high-risk groups, omission of radiation was associated with a significantly decreased EFS rate of 77% for those treated with chemotherapy only versus 92% for those who received combined modality therapy.

The CCG5942 trial included subjects with all disease stages. Based on their risk-group assignment, patients received either four or six cycles of COPP/ABV or six cycles of the mutiagent ABC regimen chemotherapy. Patients with at least a 70% tumor-volume reduction and negative gallium scan at the completion of chemotherapy were randomized to receive or not receive IFRT with 21 Gy. This randomized trial was stopped because of a significantly higher number of relapses in the no-IFRT arm. Most of the relapses were at sites of initial disease. Patients who did not receive IFRT had an EFS rate at 3 years from the time of randomization of 87% compared with 92% ( P = .057) for patients treated with IFRT in the intent-to-treat analysis. Importantly, longer term follow-up study of this cohort has shown that the OS in the two arms is no different. At 10 years postrandomization, the EFS of those treated or not treated with IFRT were 91.2% and 82.9% ( P = .004), respectively, whereas the OS of the two groups were 97.2% and 95.5% ( P = .5), respectively.

Radiation remains a critical component of therapy for the majority of children and adolescents with Hodgkin lymphoma. Radiotherapy planning involves decision making about both dosage and field, with the goals of maximizing efficacy but limiting late effects. Radiation fields have evolved over time ( Fig. 53-6 ). IFRT, used in the majority of contemporary protocols, generally involves radiation of the entire lymph-node region that includes the pathologically enlarged node. One major implication of the change from mantle-field radiation to IFRT for cervical disease is the exclusion of the axilla, which significantly reduces the exposure of lung and breast tissue. Investigations of further limitations to radiation fields are ongoing. The dosages in most contemporary studies used in combination with chemotherapy are generally no greater than 21 Gy, with some studies using higher dosages to sites of bulky or residual disease.