Part 2: Hematologic Malignancies

This chapter is dedicated in memory of James B. Nachman, MD.

INTRODUCTION

Acute lymphoblastic leukemia (ALL) is the most common malignancy of childhood and adolescence, with approximately 3500 new cases diagnosed annually in North America in patients 21 years of age and younger. ALL is defined as the presence of 25% or more malignant lymphoblasts (“blasts”) in the bone marrow. The diagnosis of ALL is often initially suspected by the presence of circulating blasts in the peripheral blood that often occurs in the setting of concomitant anemia and thrombocytopenia. Leukemia cells may also infiltrate extramedullary sites (eg, central nervous system [CNS], testes, lymphatic system, solid organs). Patients with extramedullary lymphoid masses, but less than 25% blasts in the bone marrow, are considered to have lymphoblastic lymphoma. ALL may arise from precursor B cells (B-ALL) or T cells (T-ALL).

The incidence of childhood ALL peaks between 2 and 3 years of age (Table 445-1) and is slightly more common in males than females. The incidence of leukemia is substantially higher in white children compared to black children, with the highest rates in children of Hispanic ethnicity. With the exception of Down syndrome and rare familial leukemia predisposition syndromes, causal factors of childhood ALL remain largely unknown. Epidemiologic studies have also investigated various environmental factors (eg, viral infections, parental smoking, living near power lines) potentially associated with increased risk of childhood ALL, although definitive associations have not been proven. Exposure to very high doses of ionizing radiation, chemicals such as benzene, and certain chemotherapies have been associated with an increased risk of acute myeloid leukemia (AML), but not ALL.

Remarkable progress has been made in the treatment of children with ALL during the past 6 decades. This success is largely attributable to improved understanding of key biologic and genetic features of childhood ALL that has informed risk stratification of patients, as well as to appropriate therapeutic intensification for children with high-risk disease. Enhanced supportive care has also decreased morbidity and mortality and contributed to improved outcomes. In the 1960s, fewer than 10% of children with ALL were cured. In 2016, more than 90% of all children with newly diagnosed ALL can be cured with modern frontline therapy (Fig. 445-1). Relapse-free survival exceeds 95% for a subset of patients with the most favorable genetic features and treatment response. Patients are generally considered cured if they remain in remission more than 5 years from completion of therapy.

BIOLOGY

In normal T- and B-cell development, lymphocyte maturation and differentiation occurs independently of antigenic stimulation. Subsequent interaction of T and B cells with their specific antigens results in lymphocyte proliferation and differentiation, resulting in (1) effector T cells capable of cytotoxicity, cytokine secretion, and B-cell help, (2) long-lived memory T cells for sustained antigen recognition, and (3) plasma cells of B-cell origin involved in antibody secretion.

Eighty percent of cases of childhood ALL are of B-cell lineage, whereas 15% are of T-cell lineage. T-cell ALL comprises 10% of ALL in children ages 1 to 9 years and 20% of ALL in adolescents and young adults. Approximately 98% of B-ALL results from arrest at an immature B-precursor cell stage, which confers a distinct immunophenotype (Fig. 445-2). Rarely, mature B-cell leukemias can occur, which are a form of Burkitt lymphoma and require different treatment (discussed in detail in Chapter 447). The various stages of B-cell development cannot be judged by cell morphology but require classification by patterns of cell surface antigen expression. The earliest stage in B-cell development is the pro-B cell, which is characterized by surface expression of the histocompatibility complex HLA-DR, lack of CD10 (CALLA, common ALL antigen), and lack of cytoplasmic immunoglobulins. The slightly more mature pre-B cell expresses cytoplasmic immunoglobulin and may or may not express CD10. Most childhood ALL arises from pre-B cells, while the pro-B immunophenotype is most often seen in infants with ALL (Fig. 445-2). In T-ALL, cases are evenly distributed among the early, mid, and late thymocyte states classified by the presence or absence of various T-cell antigens, including CD2, CD4, CD5, CD7, CD8, and cytoplasmic CD3. Rarely, ALL cases may show a mixed phenotype (most commonly T/myeloid or B/myeloid) or a natural killer cell immunophenotype.

Many leukemia-associated somatic alterations are prognostic with respect to clinical outcomes (Table 445-2). Characterization of the cytogenetics and other molecular mutations within ALL cells are thus important factors for assignment of patients to appropriately intensive therapy. To determine cytogenetic abnormalities in childhood ALL, chromosomal metaphase preparations of leukemic cells (“spreads”) are analyzed to identify potential abnormalities in chromosome number and structure. It is now well known that certain genetic abnormalities cannot be detected by routine cytogenetic analysis. An additional cytogenetic assay called fluorescent in situ hybridization (FISH) uses dye-conjugated probes to identify specific chromosomal losses, gains, or translocations (rearrangement or exchange of genetic material between nonhomologous chromosomes) within leukemic cell spreads. Newer molecular assays using DNA- or RNA-based array and sequencing technologies can now also identify cytogenetically cryptic point mutations or gene fusions, many of which have prognostic significance and therapeutic implications.

Most cases of B-ALL have detectable chromosomal abnormalities by either standard cytogenetic or molecular techniques. These abnormalities may include gains or losses of whole chromosomes, structural abnormalities involving individual chromosomes, or translocations involving 2 or more chromosomes. The 2 most common genetic aberrations in childhood B-lineage ALL are (1) hyperdiploidy (usually 51–67 chromosomes) and (2) translocation between chromosomes 12 and 21 that results in ETV6-RUNX1 fusion. High hyperdiploidy and t(12;21) each comprise 20% to 25% of alterations in younger children with B-ALL. Both abnormalities are most common in children 1 to 9 years of age and are associated with a good prognosis. Hypodiploidy (< 44 chromosomes) occurs rarely in childhood ALL, although this subtype is associated with a poor prognosis. Survival analyses have further demonstrated increasingly inferior survival with increasing chromosomal loss; worse outcomes occur for children with near-haploid (24–31 chromosomes) or low hypodiploid (32–39 chromosomes) ALL.

Translocation between chromosomes 9 and 22 resulting in BCR-ABL1 fusion (the Philadelphia chromosome) is another uncommon genetic abnormality (occurring in 3–4% of childhood B-ALL) that was historically associated with a very poor prognosis prior to addition of tyrosine kinase inhibitor therapy (discussed below). The incidence of Philadelphia chromosome-positive (Ph⁺) ALL increases during adolescence and accounts for a major subset of adult ALL.

A new subtype of B-ALL known as BCR-ABL1–like or Philadelphia chromosome–like (Ph-like) was recently identified. Ph-like ALL is characterized by an activated kinase gene expression profile similar to that observed in Ph⁺ ALL, although the BCR-ABL1 fusion is not present. The Ph-like subset comprises 10% to 20% of B-ALL occurring in children and adolescents and is associated with a variety of genetic alterations that activate kinase signaling. Children with Ph-like ALL have poor clinical outcomes when treated with standard chemotherapy, but these leukemias may be amenable to treatment with the addition of kinase inhibitors (discussed below), similar to the approach now taken in Ph⁺ ALL.

Another common translocation in childhood B-ALL is between chromosomes 1 and 19, which results in TCF3-PBX1 fusion. Although the t(1;19) was historically associated with intermediate outcomes, it is no longer considered prognostic in the era of modern risk-based chemotherapy. Another recurrent translocation involves rearrangement of KMT2A (formerly MLL) on the long arm of chromosome 11 (at 11q23) with 1 of more than 75 known gene partners. KMT2A-rearranged ALL occurs in approximately 75% of infants less than 1 year of age and carries an extremely poor prognosis even with intensive chemotherapy. The incidence of KMT2A-rearranged ALL is much lower in noninfants, accounting for 5% to 6% of B-ALL cases occurring in children and adolescents. While older children with KMT2A-rearranged ALL do not fare quite as poorly as infants, they are treated as high risk with more intensive chemotherapy, given their inferior outcomes. Finally, intrachromosomal amplification of part of chromosome 21 (iAMP21) occurs in approximately 2% of children with B-ALL. This genetic alteration is also considered high risk, and children with iAMP21 ALL are also treated more intensively. Other less common recurrent genetic abnormalities also occur in childhood B-ALL, some of which have prognostic significance.

In T-ALL, it is uncommon to detect leukemia-associated abnormalities based on routine cytogenetic analyses. However, FISH and other molecular testing often reveal recurrent abnormalities. For example, mutations in NOTCH1 occur in approximately 50% of cases of childhood T-ALL and appear to be associated with more favorable prognoses. The prognostic significance of most recurrent alterations in T-ALL remains unknown, however, so they are not presently used for risk stratification.

CLINICAL PRESENTATION

In many cases of childhood ALL, production of the various types of blood cells other than lymphocytes is seriously impaired. Children may present with decreased hemoglobin and hematocrit (anemia), manifesting as pallor or fatigue. Decreased platelets (thrombocytopenia) can lead to bruising and spontaneous bleeding. Low white blood cell (WBC) counts can lead to serious bacterial, fungal, or viral infections manifesting with fever, particularly in the context of prolonged neutropenia or lymphopenia. The triad of fever, pallor, and bruising is a common presentation of ALL, particularly in younger children. Some children also initially present with bone pain or refusal to walk due to leukemic infiltration of the bone marrow. Many patients with ALL present with palpable lymphadenopathy, hepatomegaly, or splenomegaly due to infiltration of organs by leukemic blasts. Children and adolescents with ALL less commonly present with respiratory compromise due to the presence of a large anterior mediastinal mass, which is more common in patients with T-ALL. Large mediastinal masses causing respiratory distress are considered a medical emergency and often require immediate intervention. The presence of lytic bone lesions diagnosed by radiography or magnetic resonance imaging (MRI) is another rare presentation of ALL. Prior to performing a biopsy of a lytic bone lesion, a complete blood count and a careful review of the peripheral blood smear should be performed.

DIAGNOSIS

Many patients with ALL have evidence of circulating blasts that can be seen on a peripheral blood smear, but the specific diagnosis must be made by flow cytometry analysis. ALL blasts are characterized by a high nuclear-to-cytoplasmic ratio, clumpy nuclear chromatin, and few small nucleoli. Even if peripheral blasts are present, performing a bone marrow aspiration is highly recommended to obtain optimal leukemia specimens for flow cytometric determination of immunophenotype, cytogenetic and FISH analyses, as well as other molecular genetic testing. Classification of leukemia immunophenotype and specific cytogenetic translocations are necessary for the risk stratification and treatment of childhood ALL. Most cases of childhood ALL are of B-cell or T-cell origin that can be readily identified by flow cytometric delineation of specific cell surface antigens. However, both B-ALL and T-ALL may rarely have aberrant concomitant expression of myeloid antigens consistent with mixed phenotype acute leukemia, which can be a diagnostic dilemma and is generally associated with an inferior treatment response.

As part of the routine diagnostic evaluation, a lumbar puncture (LP) should be performed to assess whether blasts are present in the cerebrospinal fluid (CSF). Ideally, the initial LP (or spinal tap) should be performed by an experienced physician using sedation for younger patients to minimize the possibility of a traumatic tap that may alter proper risk stratification. Patients with < 5 WBCs per μL in CSF and no evidence of microscopic ALL involvement on a cytocentrifuged (cytospin) specimen are classified as CNS negative (CNS 1). In a nontraumatic LP, the presence of ≥ 5 WBC/μL in the CSF and the presence of leukemic blasts on cytospin are needed to make a diagnosis of overt CNS involvement (CNS 3). Patients with < 5 WBC/μL in the CSF, but with detectable leukemic blasts on cytospin, are classified as CNS 2. Patients who present with neurologic dysfunction (eg, cranial neuropathies, altered mental status, back pain) should undergo brain or spine computed tomography (CT) or MRI prior to an LP to assess potential CNS infiltration with leukemia and to ensure safety in performing the procedure. Rarely, patients can present with retinal hemorrhages or with vision changes due to leukemic infiltration of the optic nerves. Any patient with ocular symptoms should also be evaluated immediately by formal ophthalmologic exam and MRI of the brain and orbits. Urgent radiotherapy can preserve vision in some patients with acute-onset vision loss and optic nerve infiltration with leukemia. Testicular examination to assess for masses and asymmetry should be performed in all males at diagnosis and during therapy as testicular infiltration with leukemia may occur, more commonly at relapse than at diagnosis. Testicular biopsy is indicated in boys with suspected testicular involvement if the findings persist after the first course of therapy. If confirmed, testicular radiotherapy may be incorporated into the treatment plan.

PROGNOSTIC FACTORS

A number of prognostic factors are utilized for risk stratification of patients and allocation to treatment of appropriate intensity (Table 445-2). Prognostic factors can generally be divided into 3 groups: (1) host-related factors, (2) disease-related factors, and (3) treatment response factors.

Age is an important prognostic factor in childhood B-ALL. Generally, children ages 1 to 9 years have better prognoses and outcomes than those younger than 1 year or older than 10 years of age at time of diagnosis. This difference is in part related to the significantly higher incidence of favorable cytogenetic subtypes of ALL (eg, hyperdiploidy, ETV6-RUNX1 translocation) that comprise 50% to 60% of recurrent genetic alterations in B-ALL in younger children. As previously mentioned, infants less than 1 year of age with KMT2A-rearranged ALL have inferior clinical outcomes.

Presenting WBC count is another important prognostic factor in B-ALL. Many pediatric oncology consortia utilize a cutoff of 50,000 WBC/μL at diagnosis to determine risk group assignment. Extreme hyperleukocytosis (WBC > 200,000/μL) is often associated with unfavorable leukemia cytogenetics and generally portends a poorer clinical prognosis. Overt CNS involvement of leukemia is a less favorable prognostic factor, and testicular involvement with leukemia in males is also considered high risk.

The combination of age and WBC is currently used for initial risk group allocation in patients with B-ALL. Standard risk B-ALL is defined as age > 1 and < 10 years and a WBC of < 50,000/μL, while high-risk B-ALL is defined as age ≥ 10 years and/or initial WBC ≥ 50,000/μL. Children with CNS positivity (CNS 3) are also generally considered high risk. Hepatosplenomegaly and other extramedullary involvement at diagnosis have no independent prognostic significance.

Race and ethnicity have been evaluated as prognostic factors in many studies, but are not considered prognostic with modern anti-ALL therapy. However, emerging data demonstrate that specific genetic variations and mutations that confer increased chemosensitivity or are associated with increased toxicity are more common in certain racial and ethnic groups of patients.

In the past, pretherapy factors such as age, initial WBC count, and blast cytogenetics were used to determine risk group and treatment assignment. However, rapidity of response to induction chemotherapy is now recognized as the most important factor determining clinical outcome. Several large clinical trials have demonstrated that children who achieve remission with no detectable minimal residual disease (MRD) at the end of induction therapy have significantly better outcomes compared to those with higher levels of MRD. Therefore, children with ALL are currently assigned to an initial induction treatment based on age, WBC count, and immunophenotype. Treatment is subsequently modified at the end of induction based on depth of achieved remission (as measured by MRD) and by leukemia-associated genetic results (Fig. 445-3).

In some studies, the morphologic response in the peripheral blood following a 7-day treatment prophase with prednisone alone and a single dose of intrathecal chemotherapy has demonstrated prognostic significance. Patients with fewer than 1000 blasts/μL in peripheral blood on day 8 had a cure rate greater than 70%, whereas patients with greater than 1000 blasts/μL had a cure rate of approximately 30%. Historically, patients with 25% or greater blasts in the bone marrow on day 8 of induction therapy (“slow early response,” or SER) also had a significantly worse prognosis compared to patients with less than 25% blasts on the day 8 bone marrow evaluation. Subsequent treatment of SER patients with more aggressive treatment strategies has resulted in improved outcomes in some studies.

Morphologic assessment of response by light microscopic examination of post-treatment bone marrow tissue remains an imprecise technique, as it reliably allows detection of only relatively high levels of residual ALL. Newer, more sensitive genetic- and flow cytometry–based MRD assays are capable of detecting submicroscopic amounts of residual leukemia. Because ALL is a malignant clonal disorder, each cell in the clone has a unique immunoglobulin gene rearrangement. This rearrangement can be identified by polymerase chain reaction (PCR) and sequenced to determine the percentage of cells in peripheral blood or marrow with the unique leukemia-associated rearrangement. Polymerase chain reaction–based MRD techniques have a sensitivity of 0.01%, meaning that such assays can identify 1 ALL cell among 10,000 normal cells. In addition, most ALL cells have unique immunophenotypes that can be distinguished from normal B or T cells by flow cytometry. Flow cytometric MRD analyses also have a sensitivity of approximately 1 in 10,000 cells (0.01%). Regardless of the detection assay used, patients with high levels of MRD at the end of induction therapy have worse outcomes than those patients with very low or undetectable levels. New assays using next-generation sequencing technologies are currently in clinical development and have demonstrated even greater sensitivity for MRD detection in childhood ALL.

INITIAL MANAGEMENT

With rare exception, emergent treatment with chemotherapy is not necessary and may actually be contraindicated. Patients often present with severe metabolic abnormalities, hematologic abnormalities, or infectious complications that must be adequately managed before antileukemia treatment can be safely initiated. Initial assessment of the newly diagnosed patient should begin with a careful determination of the vital signs. A patient with ALL and a large mediastinal mass may present with respiratory distress due to tracheal compression; if untreated, this can progress to respiratory arrest. Sedation of such patients carries a significant risk of cardiac arrest due to impediment of venous blood return to the heart. Many patients with ALL will present with fever that is commonly associated with neutropenia (absolute neutrophil count < 500/μL), which puts them at high risk for bacterial sepsis. In patients with fever and neutropenia, broad-spectrum antibiotic coverage should be promptly initiated after blood and any other clinically indicated cultures are obtained.

Tumor lysis syndrome (TLS) refers to the release of intracellular by-products from leukemic cells, including uric acid and phosphorus. Patients with ALL and high WBC counts or large extramedullary masses often present with an elevated blood uric acid level. In rare cases, patients may present with significant renal compromise due to uric acid nephropathy. In any patient with ALL and an initial high uric acid level, renal function (serum creatinine) should be carefully assessed, and medications such as allopurinol should be started to lower the uric acid level. Even if uric acid levels are initially normal, they may rise to high levels as chemotherapy is initiated and leukemia cells break down. All patients should thus receive therapy to prevent the development of uric acid nephropathy even if uric acid levels are not elevated at presentation. For patients with high tumor burdens and/or hyperuricemia, use of rasburicase (urate oxidase) is recommended instead of allopurinol. Rasburicase degrades uric acid to harmless metabolites, and blood levels of uric acid fall precipitously following a single dose. Short-term dialysis may rarely be required for patients with renal dysfunction who do not show rapid improvement in renal function with intravenous fluid hyperhydration, and initiation of allopurinol or rasburicase. Serum phosphate levels, although rarely significantly elevated at diagnosis, may rise as treatment is initiated. This hyperphosphatemia can lead to hypocalcemia and precipitation of calcium phosphate in the renal tubules. Addition of medications to bind excess phosphorus can be used, such as aluminum hydroxide or sevelamer.

Patients with ALL are often anemic at diagnosis. Development of anemia in ALL is often a chronic process. For severely anemic patients, packed red blood cell (PRBC) transfusions should be approached with caution. Small aliquots of PRBCs can initially be given to raise the hemoglobin concentration to 7 to 8 g/dL. If necessary, a diuretic may be given to control fluid volume. In patients with very high WBC counts, increasing hemoglobin levels beyond 7 to 8 g/dL may significantly increase viscosity and increase risk of thrombosis.

Patients can present with hyperleukocytosis that requires urgent treatment. Extreme leukocytosis (WBC > 200,000/μL) is usually asymptomatic but may be associated with CNS symptoms and/or thrombosis. A diagnosis of leukemia can often be established rapidly by peripheral blood flow cytometry in these circumstances, and treatment with corticosteroids can be initiated to lower the WBC. In these situations, patients should be very carefully monitored for TLS. Leukopheresis can also be performed for temporary reduction of WBC count, but the overall additional efficacy of this procedure remains unknown.

Patients with ALL (most commonly T-ALL) can present with a large anterior mediastinal mass and risk of airway compromise. Sedation and anesthesia should be avoided in such situations. Initiation of steroid therapy will generally shrink the mass rapidly, but may increase the risk for TLS. Low-dose radiation to a small field around the airway can also be considered. Following the initiation of radiation, significant improvement in respiratory status is usually seen within 12 hours.

TREATMENT STRATEGY

Ten-year overall survival rates now exceed 90% for childhood ALL (see Fig. 445-1). This remarkable success can be attributed to several fundamental observations and principles from early treatment eras that have continued to provide the foundation for contemporary therapy. One of the first essential observations from early treatment eras was the recognition that cure requires the sustained delivery of therapy beyond an initial remission. Additional advances have included (1) the establishment of the efficacy of combination chemotherapy, (2) the recognition of the critical need for presymptomatic treatment of the CNS sanctuary site, and (3) the value of later intensification of therapy to eradicate remaining residual disease, especially among children with slow initial treatment responses. More recently, risk-directed treatment approaches have been adopted and refined.

Modern treatment for childhood ALL consists of multiagent chemotherapy administered in several phases: induction, consolidation, intensification, and maintenance phases of therapy (Table 445-3). The induction phase of treatment generally lasts 1 month, and over 95% of children achieve a remission at the end of this phase. The next phases of treatment (consolidation and intensification) are designed to intensify systemic and CNS-directed therapy to eradicate any residual disease. The final and longest phase of treatment (maintenance) is designed to prevent disease recurrence. In general, treatment is 2 to 3 years in duration. Some earlier ALL studies treated boys for longer than girls because male gender was associated with inferior outcomes. Treatment duration is now the same for boys and girls, as contemporary regimens have abrogated the prognostic impact of male gender.

Among children with newly diagnosed ALL, specific treatment approaches have generally been developed based on immunophenotype (B-ALL vs T-ALL) and age (infants < 1 year of age vs older children). Unique genetic subtypes of ALL, such as Ph⁺ ALL, are treated with separate regimens that include a tyrosine kinase inhibitor. Several studies have now shown that outcomes are improved for adolescents and young adults treated with pediatric protocol-inspired regimens. Accordingly, many “pediatric” ALL clinical trials now treat patients up to 30 years of age.

In an ongoing effort to reduce the risk for late treatment effects, there has been a growing trend to minimize the use of cranial radiotherapy. CNS prophylaxis for the majority of children now consists of intrathecal chemotherapy and systemic agents that penetrate the CNS, such as dexamethasone and high-dose methotrexate. Radiotherapy tends to be limited to those children who have CNS 3 leukemic involvement at diagnosis. Some treating groups, however, have eliminated the use of radiotherapy entirely.

The current paradigm for systemic treatment of patients with B-ALL begins by assigning patients to either a standard or a high-risk group based on age and initial WBC count. Within the Children’s Oncology Group (COG), children ages 1 to 9 years who have an initial WBC count of less than 50,000/μL are assigned to a standard risk group and receive a 3-drug induction that includes vincristine, asparaginase, and dexamethasone. All other patients are considered high risk and receive a 4-drug induction that also includes an anthracycline such as daunomycin. High-risk patients ages 1 to 9 years have been shown to benefit from dexamethasone as the induction steroid, whereas older patients have not derived the same benefit and receive prednisone.

Patients are reclassified at the end of induction based on their initial risk group, blast cytogenetics, and early response as assessed by flow cytometric or PCR-based measurement of MRD during and at the end of induction (Fig. 445-3). Patients with hypodiploidy (< 44 chromosomes in the blast population) are considered very high risk regardless of initial risk group or early response and often undergo hematopoietic stem cell transplantation (HSCT) in first remission. Patients with Ph⁺ ALL receive a tyrosine kinase inhibitor, which inhibits the BCR-ABL1 fusion protein, in conjunction with intensive chemotherapy. End-induction MRD is the most significant predictor of outcome, and patients who have MRD ≥ 0.01% are treated more intensively following induction. Patients with rapid responses and MRD < 0.01% at the end of induction are assigned to treatment regimens based on initial risk group and blast cytogenetics. Standard-risk rapid responders with either the ETV6-RUNX1 translocation or hyperdiploidy with extra copies of chromosomes 4 and 10 and no extramedullary disease have a projected event-free survival (EFS) of greater than 95% and are treated with a lower-risk regimen with less exposure to anthracyclines and no exposure to cyclophosphamide.

Several advances have been made in the treatment of standard-risk B-ALL patients, and their 5-year EFS now exceeds 90%. Survival rates approaching 100% are observed in certain very favorable low-risk subsets, including standard-risk patients with favorable cytogenetics and rapid MRD responses. For standard-risk patients who are MRD positive at the end of induction, improved outcomes can be achieved with intensification of postinduction therapy.

A recent trial for children with high-risk ALL compared dexamethasone to prednisone during the induction phase of treatment and high-dose methotrexate to escalating lower doses of intravenous methotrexate without leucovorin rescue in interim maintenance and demonstrated the superiority of high-dose methotrexate among all age groups. The superiority of dexamethasone for 14 days during induction was also observed in high-risk patients less than 10 years of age. This modified augmented Berlin-Frankfurt-MÜnster (BFM) treatment regimen has now become a standard in COG high-risk ALL protocols. The prognostic role of MRD response has also been further defined. Persistent MRD on day 29 of induction is the most robust prognostic factor, but later time point MRD also significantly influences prognosis such that patients who are MRD positive at the end of induction, but achieve MRD negativity at the end of consolidation or its equivalent, have favorable outcomes with EFS of approximately 80%. In contrast, patients with MRD ≥ 0.01% at the end of consolidation fare quite poorly, with estimated EFS rates of 39%. Many treating groups consider HSCT or alternative treatment approaches after remission is achieved for these patients.

Infants less than 1 year of age with ALL pose a unique challenge. The majority of patients less than 6 months of age have KMT2A (formerly MLL) gene rearrangements, most commonly a translocation between chromosomes 4 and 11, and have poor outcomes. Infants without KMT2A gene rearrangements have a significantly better outcome. Currently, infants with ALL are treated with intensive chemotherapy for 2 years. Benefits for HSCT have not been established, and current trials are focusing on the inclusion of targeted therapy directed at the unique biological pathway activation in this subtype of ALL, including epigenetic modulators.

Substantial recent advances have also occurred in the treatment of T-ALL. Whereas T-ALL has historically been more challenging to treat, outcomes now approach those observed in B-lineage disease with several international treating consortia reporting overall survival rates of approximately 90%. Intensive therapy, very similar to that employed in high-risk B-ALL, is required to achieve cure, and recent clinical trials are investigating new agents targeting unique underlying biological pathways. Recent trials have also demonstrated that, like B-ALL, cranial radiation can be safely eliminated in the majority of patients. The superiority of dexamethasone versus prednisone in reducing relapse risk has also been demonstrated. The kinetic pattern of disease regression is unique in T-ALL with slower initial responses. Similar to B-ALL, the persistence of MRD at the end of consolidation is a negative prognostic factor and an indication for treatment intensification. Unlike B-ALL, presenting WBC count is not prognostically significant in T-ALL, so children receive a common high-risk 4-drug induction regardless of initial WBC or age at diagnosis.

While most children with ALL are now cured, relapses may occur in favorable-risk patients and are often unpredictable. The current focus in ALL therapy has therefore centered upon further refining risk prediction. Comprehensive genomic analyses have been pivotal in this regard and have identified new subtypes of ALL, such as Ph-like ALL, where targeted agents are being introduced in an effort to improve cure rates. Efforts to further optimize the delivery of standard chemotherapy agents and to introduce promising new immunotherapeutic agents for these children who are at the highest risk for disease recurrence are presently underway. Finally, for rare leukemia subtypes, such as Ph⁺ ALL and infant ALL, international collaborations to address the unique challenges among these populations are in progress.

RECURRENT DISEASE

Despite numerous advances in the treatment and supportive care for children with ALL, 2% to 3% will die of infectious or other toxic complications, 1% will develop a second malignant neoplasm, and 10% to 15% will relapse. Recurrent or relapsed ALL remains a significant challenge, as fewer than 50% of patients will survive long term. Relapse can occur in the bone marrow or extramedullary sites, with CNS and testes being most common, or in both locations. The most important prognostic factors for recurrent ALL are the site and timing of the recurrence relative to the initial diagnosis. In general, isolated bone marrow relapse carries a worse prognosis compared to extramedullary or combined relapse. Early relapse (< 36 months from initial diagnosis) has a worse outcome than late relapse. Relapses of T-ALL are particularly challenging to treat and have very few long-term survivors.

The general treatment strategy for early isolated or combined bone marrow and extramedullary relapse is to reinduce remission with intensive chemotherapy. Once remission is achieved, allogeneic HSCT is generally recommended for early marrow relapses as this has been shown to increase chances of cure when compared to ongoing treatment with chemotherapy alone. Late isolated extramedullary relapse carries a more favorable prognosis and is generally treated with chemotherapy and local therapy (treatment directed to the extramedullary site). Treatment for late bone marrow and early extramedullary relapse varies, and most cooperative groups utilize early MRD response to determine whether to continue intensive chemotherapy alone or to pursue allogeneic HSCT after a second remission is achieved.

Current intensive therapy regimens have reached the limit of tolerability with or without HSCT for children with relapsed ALL, and further improvements in outcomes are likely not possible without development of new treatment approaches. Several new promising immunotherapeutic agents have thus been developed in recent years. Monoclonal antibody therapies (some conjugated to drug toxins) directed against leukemic cell surface antigens have shown promising early results. Bi-specific T-cell–engaging (BiTE) antibodies that simultaneously target a cell surface antigen on leukemic blasts and a target antigen on cytotoxic T cells have also been developed. This bridging effect facilitates more direct contact of effector T cells to kill leukemia cells. Similarly, chimeric antigen receptor (CAR) T-cell therapy has been developed where autologous T cells are harvested and engineered to express a chimeric receptor targeting leukemic cell surface antigens. Very promising initial response rates of 90% have been observed with CD19-redirected CAR T-cell therapy in children with relapsed or refractory B-ALL, but the long-term success of such therapies is not yet known. These and other new immunotherapies are under active investigation in clinical trials in children with relapsed and refractory B-ALL.

COMPLICATIONS AND LATE EFFECTS

Survival rates for children with newly-diagnosed ALL now exceed 90%. However, this achievement requires 2 to 3 years of intensive chemotherapy, which can come at a significant cost in terms of acute and long-term toxicities. There are many ongoing efforts to better understand the spectrum of acute and late treatment-related effects of therapy and to improve the quality of ALL survivorship.

Infectious complications are the most common acute toxicities in children undergoing ALL therapy and the most common cause of treatment-related deaths. During treatment, patients with ALL often have extremely low lymphocyte counts and are at risk for pulmonary or disseminated infection with Pneumocystis jirovecii (formerly Pneumocystis carinii), which can be fatal. All patients with ALL thus receive prophylactic trimethoprim/sulfamethoxazole therapy 2 to 3 days per week. Supportive care guidelines routinely include surveillance for infection during all phases of treatment with prompt initiation of broad-spectrum antimicrobial therapy for fever and neutropenia. Such guidelines and treatment modifications have been developed to minimize the risk for infectious complications among the most vulnerable ALL populations, including infants and children with Down syndrome.

One of the most significant causes of acute and long-term morbidity in patients undergoing ALL therapy is osteonecrosis, occurring in 1% to 2% of young children and up to 15% to 20% of adolescents. Osteonecrosis is disabling in many children and most commonly affects weight-bearing joints. It frequently occurs in multiple joints and requires surgical intervention in more than 25% of cases. While the onset of osteonecrosis is typically during treatment, symptoms can persist long-term. Risk factors for osteonecrosis include steroid exposure (especially dexamethasone), adolescent age, female sex, and high body mass index. Recently, several host genetic polymorphisms in a diversity of biological pathways have been linked to a higher risk of osteonecrosis. Results from a recently completed COG trial for high-risk ALL suggest that alternate week dosing of dexamethasone is 1 successful strategy to reduce the risk of osteonecrosis, while preserving treatment efficacy. Other studies are assessing potential contribution of other common ALL agents, such as methotrexate and asparaginase, to development of osteonecrosis.

Cardiac late effects from anthracyclines can be observed in ALL survivors and have been the focus of many recent studies. Contemporary ALL protocols have limited the cumulative exposure to anthracyclines compared to past regimens; however, even lower exposures can lead to asymptomatic echocardiographic changes. One successful strategy to mitigate these effects has been use of the cardioprotectant dexrazoxane, which is an iron-chelating agent designed to prevent free radical formation. This agent has been shown to protect children from anthracycline-related heart damage without compromising the effectiveness of ALL therapy or increasing the risk for secondary leukemia.

One of the most significant concerns about ALL therapy is the short- and long-term impact on neurocognitive outcomes. Early treatment regimens in the 1970s and 1980s commonly included higher doses (24 Gy or higher) of cranial radiotherapy, which has been associated with a risk for secondary brain tumors, decline in cognitive functioning, and endocrinopathies. Cranial radiation is now delivered at much lower doses (12–18 Gy) on contemporary treatment regimens, and radiotherapy has been eliminated altogether in the vast majority of children. This has raised the question of late neurocognitive outcomes among children treated with systemic and intrathecal chemotherapy alone. Most studies have shown that, while the toxicities are fewer than observed with higher doses of cranial radiation, late effects still occur. Recent studies have shown, for example, that children who have been exposed to intrathecal chemotherapy and higher doses of methotrexate are at risk for attention problems and impairment in executive functioning, although their intelligence is generally unchanged from that of the general population. As the number of adult long-term survivors of childhood ALL continues to increase, concerted efforts to minimize the risk for treatment-related complications and late effects are underway.

FUTURE DIRECTIONS

It is an exciting time in the treatment of childhood ALL with several major recent advances in the understanding of disease biology and host determinants of response, toxicity profiles, and outcomes. Breakthroughs in the genomic characterization of disease, coupled with the development of new targeted medications, have fostered the development of new precision medicine approaches for several ALL subtypes. Refinements in risk classification and MRD assessment have now identified subgroups of favorable risk patients who have outstanding chances for cure with less toxic regimens. While relapsed leukemia remains a significant challenge, several new immunotherapeutic treatment approaches have been developed, and active investigation to understand the longer-term impact of these therapies is ongoing. Finally, as survival rates now exceed 90% for children with ALL, dedicated efforts are in progress to reduce the risk for treatment-related long-term effects and to improve the quality of ALL survivorship.

SUGGESTED READINGS

INTRODUCTION

Acute myeloid leukemia (AML) accounts for approximately 20% of cases of acute leukemia in children versus 80% of cases of acute leukemia among adults. Over the past 20 years, there has been little improvement in rates of cure for AML despite maximally intensive therapy; with intensive therapy, 50% to 60% of children with AML will survive. This is in contrast to childhood acute lymphoblastic leukemia (ALL), where more than 85% of children will survive. In fact, survival rates in ALL continue to increase even with deintensification of treatment to reduce long-term side effects, and with improvement in quality of life in clearly defined subsets of children with ALL. AML therapy remains maximally toxic and requires multiple, near myeloablative courses of treatment with chemotherapeutic agents and often hematopoietic stem cell transplantation (HSCT). Recent progress in AML has largely been through the identification of genomically defined risk groups. Genetic events predictive of higher and lower cure rates permit the allocation of patients to conventional cytotoxic therapy or HSCT, respectively, while other genetic events provide promise for more targeted therapies.

The chronic myeloid forms of leukemia are extremely rare in children. These myeloproliferative syndromes (MPS) most commonly include the adult type of Philadelphia chromosome–positive (Ph+) chronic myelogenous leukemia (CML), and juvenile myelomonocytic leukemia (JMML). The clinical course, biologic characteristics, and molecular pathogenesis of CML and JMML are quite different. Until recently, allogeneic HSCT from either a related or an unrelated donor was the management of choice for children with Ph+ CML but the kinase inhibitor imatinib has changed the treatment paradigm for that disease in both children and adults, although HSCT is still the only known curative therapy for CML. Similarly, allogeneic HSCT remains the only known curative option for children with JMML.

EPIDEMIOLOGY

The annual incidence of AML throughout childhood remains constant, except for slight peaks during infancy and adolescence. In fact, in infants, AML is more common than ALL. There are approximately 7 cases per million children per year, or approximately 600 new cases per year in the United States. After age 20 years, the incidence of AML slowly increases with age (Fig. 446-1). There is a slightly higher incidence in Hispanic children with 9 cases per million per year. The incidence also appears slightly higher in Japan, Australia, and Zimbabwe.

The cause of AML is unknown, and most children have no known predisposing factors. Known risk factors include exposure to high-dose ionizing radiation, previous chemotherapy (especially with alkylating agents and epipodophyllotoxins), Down syndrome, congenital bone marrow failure syndromes (Diamond-Blackfan anemia and Kostmann agranulocytosis; see Chapter 437), chromosome fragility and impaired DNA repair mechanisms (such as Fanconi anemia), and inherited disorders, such as neurofibromatosis type 1 (NF1). Neurofibromatosis type 1 is caused by mutations in neurofibromin, a RAS-directed GTPase, leading to constitutively active signaling through the RAS pathway and resulting in an increased risk of malignant disease, including MPS and MDS.

The risk of secondary AML (or therapy-related AML) among children and adults previously treated with alkylating agents and topoisomerase-II inhibitors, especially epipodophyllotoxins, is well established. Leukemia associated with use of an alkylating agent occurs within 4 to 10 years after initial therapy, is usually associated with abnormalities of chromosomes 5 and 7, and carries a grave prognosis. Leukemia associated with use of an epipodophyllotoxin (eg, etoposide) has a shorter latency (2–4 years) and usually is of the myelomonocytic or monocytic subtype; characteristically, translocations involving chromosome 11q23 with rearrangement of the KMT2A (previously called MLL) gene are noted. Children with secondary leukemia often achieve complete remission with chemotherapy but invariably relapse and die; some may be successfully treated with HSCT.

CLINICAL AND LABORATORY FEATURES

The initial signs and symptoms for most children with AML result from anemia, thrombocytopenia, and neutropenia caused by bone marrow infiltration with leukemic blasts and decreased production of normal cells. Patients commonly present with pallor, fatigue, epistaxis, gum bleeding, petechiae, or purpura, as well as fever or infection that does not respond to antibiotic therapy. Children with AML may have bone or joint pain, but these symptoms occur more often in children with ALL. Bulky peripheral or mediastinal lymphadenopathy is not a common finding, and massive hepatosplenomegaly is rare with AML except among infants. Extramedullary leukemia can present as gingival hyperplasia, central nervous system (CNS) leukemia (headache, cranial nerve palsy), and skin nodules. Neonates and infants with AML frequently have leukemia cutis characterized by a papular or nodular rash that is salmon or bluish to slate gray in color. Clinical findings of CNS leukemia at diagnosis are rare. They include signs of increased intracranial pressure or cranial nerve palsy, seventh nerve palsy being the most common. Fewer than 5% of patients with AML have chloromas (also known as granulocytic sarcoma) at diagnosis or during the course of the illness. These are solid tumors of blasts and immature myeloid cells that typically occur in the bones and soft tissues of the head and neck (often involving the orbits), or intracranial or epidural sites.

Peripheral blood counts at diagnosis in children with AML can be quite varied. The leukocyte count ranges from less than 1000/μL to more than 500,000/μL. Approximately 15% to 20% of children have an initial leukocyte count greater than 100,000/μL. Higher leukocyte counts are associated with the French-American-British (FAB) M4 and M5 subtypes, whereas lower leukocyte counts (< 5000/μL) are commonly seen in acute promyelogonous or M3 leukemia (APL). Most patients have a normocytic anemia (median hemoglobin concentration of 7 g/dL), and approximately 50% of patients have platelet counts less than 50,000/μL. Disseminated intravascular coagulation (DIC) is extremely common among almost all patients with APL and some infants with monocytic leukemia.

The characteristic bone marrow findings include hypercellularity with more than 20% blasts (usually 70–90% blasts). A bone marrow biopsy infrequently shows myelofibrosis (except for megakaryoblastic) and occasional multilineage dysplasias.

In most cases of AML, the diagnosis is straightforward after examination of the peripheral blood sample and a bone marrow aspirate. Other conditions that can cause diagnostic difficulty include the myeloproliferative disorders such as juvenile myelomonocytic leukemia, myelodysplastic syndromes (MDS), sepsis that causes a leukemoid reaction, or neutropenia caused by maturation arrest in granulocytic-monocytic precursors. In the presence of sepsis, the bone marrow findings may suggest APL because of a promyelocyte arrest with toxic granulation. However, normal granulocytic maturation ensues within a few days of the resolution of the infection. AML among neonates with Down syndrome is difficult to differentiate from transient myeloproliferative disorder (TMD), which can progress to life-threatening hyperleukocytosis with massive hepatosplenomegaly, but will nonetheless spontaneously resolve with supportive care.

LEUKEMIA SUBTYPES

For the purposes of this chapter and the discussion of molecular pathogenesis, AML will be divided into 3 subgroups: acute promyelocytic leukemia (APL), myeloid leukemia of Down syndrome (ML-DS), and all other forms of AML. APL and ML-DS are examples of well-characterized molecular subtypes of AML driven by specific oncoproteins. Current clinical research strategies in ML-DS and APL aim to reduce exposure to cytotoxic therapies by taking advantage of the leukemogenic driver events. AML in all other patients tends to be more heterogeneous with no “smoking gun” mutation amenable to targeting with currently available therapies. With an increasing focus on precision medicine, the development of specific targeting strategies in all AML subsets is a high research priority.

ACUTE PROMYELOCYTIC LEUKEMIA

GENETICS

Fusion of the N-terminus of the promyelocytic leukemia (PML) protein on chromosome 15 with the C-terminus of the retinoic acid receptor-α (RARA) transcription factor located on chromosome 17 can be found in most cases of childhood APL. The translocation produced is a functional fusion protein (PML-RARA) that causes uncontrolled self-renewal and maturation arrest in the promyelocyte phase of myeloid differentiation. Nucleophosmin 1 (NPM1), nuclear mitotic apparatus (NUMA), and promyelocytic leukemia zinc finger (PLZF) are other less common N-terminal fusion partners for RARA. PML-PLZF predicts a very poor response to treatment and a poor prognosis. In normal promyelocytes, retinoic acid (RA) serves as a hematopoietic transcription factor inducing myeloid maturation. PML-RARA does not function effectively as transcriptional activator at physiologic levels of RA. Each of the RARA translocations leads to myeloid maturation arrest at the promyelocyte stage.

TREATMENT

Acute promyelocytic leukemia is relatively unresponsive to conventional chemotherapy with historical complete remission rates of 75% to 80%, and cure rates of only 35% to 40%. Rapid lysis and apoptosis of malignant cells with conventional chemotherapy leads to a release of the cellular contents, including procoagulants like tissue factor (TF) and annexin A2. Tissue factor activates the coagulation cascade causing a rapid and inappropriate accumulation of thrombin. As many as 17% of patients with APL treated with traditional AML therapy died from bleeding complications associated with DIC triggered by the APL blasts during induction therapy. With the introduction of more targeted therapy with all-trans retinoic acid (ATRA) and arsenic trioxide (ATO), cure rates for APL now exceed 90% with greatly reduced early toxic mortality.

It is important to approach a child with APL as an oncologic emergency. ATRA should be considered as soon as APL is suspected. While physiologic levels of RA do not affect the PML-RARA oncoprotein, supraphysiologic levels of ATRA can drive PML-RARA to activate the RA receptor elements leading to cellular differentiation and subsequent polyubiquitination with proteolysis of the PML-RARA oncoprotein. ATRA alone is capable of inducing promyelocyte differentiation, decreasing TF release and up-regulating thrombomodulin expression. The net result is a greatly reduced risk for early coagulopathy-associated mortality.

While ATRA acts on the RARA moiety, ATO acts on the PML moiety. ATO induces oxidation of 2 cysteine residues, ultimately leading to polyubiquitination and proteolysis. Independent targeting of both the PML and RARA moieties results in critical degradation of the PML-RARA protein and remarkable clinical efficacy. Adults treated with just ATO and ATRA now have survival rates that exceed 90% with rare treatment-related mortality. A clinical trial of ATO and ATRA in children is underway. Only those children with a white blood cell count greater than 10,000/μL at diagnosis will receive conventional chemotherapy, and even then only in the first week of therapy.

MYELOID LEUKEMIA OF DOWN SYNDROME

GENETICS

AML in children with Down syndrome (DS) diagnosed prior to 4 years of age is a unique subset of leukemia with several distinguishing characteristics. First, children with DS have a 10- to 20-fold higher incidence of leukemia, and of those younger than 4 years of age have a 500-fold increased incidence of AML. While acute megakaryoblastic leukemia (AMKL) is a rare AML phenotype in non-DS children, nearly half of all ML-DS are of the AMKL subtype. AMKL in non-DS AML is characteristically defined by the t(1;22) translocation. This chromosomal translocation is not seen in ML-DS, even among those with the AMKL subtype. Nearly all patients with ML-DS diagnosed before 4 years of age share a common mutation in GATA1, an essential transcription factor in hematopoiesis. The fact that ML-DS is so unique in both presentation and pathogenesis suggest that GATA1 mutations are a driver event in ML-DS. Each of the identified mutations introduces a stop codon in exon 2. An alternative exon 3 translation start site is employed, resulting in a truncated but functional GATA1 protein with a deficient or absent amino-terminal activation domain. Mutations in GATA1 are found in nearly all patients with ML-DS diagnosed prior to age 4 years, but are rare in non-DS AML and in ML-DS in children older than 4 years of age. Additionally, ML-DS often is preceded by a TMD in neonates with DS. While TMD may be so rapidly progressive that some children will succumb to side effects of leukostasis and massive hepatosplenomegaly with cardiopulmonary compromise, most will experience a spontaneous resolution. GATA1 mutations can be identified in all patients with both TMD and ML-DS, but additional mutations are common only in ML-DS. These findings suggest that sequential acquisition of mutations leads to the self-sustaining malignant transformation of ML-DS.

Directly targeting the dysfunctional GATA1 oncoprotein is not currently a feasible treatment strategy. However, the unique genomic landscape of ML-DS offers other opportunities for disease and host-specific therapies. The extra copy of chromosome 21 results in increased expression of cystathionine β-synthase (CBS) and zinc-copper superoxide dismutase (SOD1), genes involved in oxidative metabolism. The current model for leukemogenesis in DS suggests that the small GATA1 exon 2 insertions, deletions, duplications, and base substitutions in ML-DS are the result of increased constitutive oxidative stress and disrupted folate metabolism resulting from SOD1 and CBS over expression, respectively.

Oxidative stress and altered folate metabolism both contribute to a nucleus deficient in DNA repair and hypersensitive to DNA damage from chemotherapeutic agents, including daunorubicin and cytarabine (ara-C) that are commonly used in the treatment of ML-DS. Essentially, ML-DS cells are more susceptible to chemotherapy than those of non-DS AML. The disrupted folate metabolism mediated by overexpression of CBS also favors more ara-C conversion to the active ara-CTP form in ML-DS than in non-DS AML. Mutated GATA1 contributes to ara-C–mediated cell death through the transcriptional repression of cytidine deaminase (CDA), the enzyme responsible for ara-C degradation. Trisomy 21 essentially creates a nuclear environment that is both mutagenic, commonly resulting in a truncated but functional GATA1 protein, and permissive of cytarabine toxicity, while deficient in repair of chemotherapy-induced DNA damage.

TREATMENT

When treated with conventional doses of chemotherapy, patients with ML-DS experience very high rates of treatment-related mortality. Enhanced event-free and overall survival in ML-DS is reported when reduced doses of cytarabine and daunorubicin are used. Event-free survival rates in the range of 83% to 93% among children younger than 4 years of age (presumed to have the GATA1 mutation) are reported using reduced-intensity therapy. Recent results of the Children’s Oncology Group (COG) AAML0431 trial demonstrate that a 3-year event-free survival rate of 90% can be maintained even with a decrease in the total anthracycline dose, and a reduction in intrathecal treatments from a total of 7 to 2. Future trials in ML-DS will confirm whether the unique pharmacogenomic effect of trisomy 21 and GATA1 mutations will permit targeted dose reductions in conventional chemotherapy to improve survival and reduce the side effects of treatment.

ALL OTHER AML

GENETICS

Acute myeloid leukemia is defined by its genomic complexity since virtually all children with AML harbor at least 1 cytogenetic or molecular abnormality or a combination of alterations that lead to disease evolution or progression. Overall, karyotypic (cytogenetic) alterations are identified in nearly 80% of childhood AML with significant age-associated variation, as discussed below. Nearly 300 recurrent cytogenetic abnormalities, including translocations, deletions, or duplications, have been identified, which culminate in similar disease phenotypes. The most common chromosomal abnormalities in children and young adults with AML include inv16, t(15;17), t(8;21), and chromosome 11q23 abnormalities (Fig. 446-2).

Cumulatively, these karyotypic alterations account for nearly half of the AML cases and occur at a higher rate in pediatric populations compared to adults. Other cytogenetic abnormalities, including monosomy 5 or deletion 5q, monosomy 7, and trisomy 8, are less common in children and have a higher prevalence in adults. A clear age-associated variation in genomic alterations has been established. As an example, there is a significantly higher prevalence of 11q23 (KMT2A) translocations in infants (more than 50% of infant AML), decreasing to less than 10% in those over 15 years of age (Fig. 446-3A). Similar age-based variation is seen with inv(16) and t(8;21) mutation, collectively referred to as core binding factor AML (CBF-AML). Core binding factor AML is less common in children less than 10 years of age but increases in prevalence over the subsequent 2 decades. Those without karyotypic alterations (normal karyotype; NK) constitute a lower proportion of younger patients, increasing in incidence to 40% to 50% in adults. In addition to karyotypic chromosomal translocations mentioned above, cryptic translocations that are not easily detected by conventional methodologies are implicated in pediatric AML. Recent data demonstrates presence of cryptic translocations such as CBFA2T3/GLIS2, NUP98/KDM5A, and NUP98/NSD1 in younger pediatric AML patients, accounting for nearly 10% of cases (Fig. 446-3B).

In addition to chromosomal alterations, somatic mutations in a number of genes have been implicated in AML pathogenesis. These somatic mutations have been used as biomarkers for prognostic determination as well as targets for directed therapy. The most common AML-associated mutations in pediatrics include FLT3 internal tandem duplication (FLT3/ITD), NPM1, CEBPA, and WT1. KIT mutations are prevalent in patients with CBF-AML. Mutations in DNMT3A, IDH1/2, and RUNX1 are highly prevalent in adult AML but rare or absent in childhood AML.

CLASSIFICATION

Precise diagnosis and classification are essential to successful management and biologic investigation of childhood leukemia. In 1976, the FAB Cooperative Group proposed a classification system based primarily on morphology and cytochemical features of the blasts and required at least 30% bone marrow blasts for the diagnosis of AML. Although FAB classification provided a valuable tool for general classification of AML, it lacked the ability to predict accurately cytogenetic subclasses and, in general, did not provide reliable prognostic information.

In 2002, the World Health Organization (WHO) proposed a new classification system that incorporated diagnostic cytogenetic information, which more reliably correlated with outcome in AML. The WHO has continued to expand the genomic basis for AML classification and more recently, expanded the number of cytogenetic abnormalities linked to AML classification, included specific gene mutations (CEBPA, NPM1, and RUNX1) in its classification system, and most notably included several pediatric-specific karyotypic and cryptic translocations in its 2016 AML classification (Table 446-1). Such a genetically based classification system links AML class with outcome and provides significant biologic and prognostic information. With new, emerging technologies aimed at genetic, epigenetic, proteomic, and immunophenotypic classification, AML classification will continue to evolve and provide informative, prognostic, and biologic guidelines to clinicians and researchers.

TREATMENT

A substantial improvement in the survival rate from less than 10% to approximately 50% of children with AML has occurred during the past 40 years. The improvement is the result of a higher percentage of children entering complete remission; a decrease in the relapse rate because of more effective postremission strategies, including allogeneic HSCT; and improvements in supportive care. All children with AML should be referred to pediatric oncology centers and treated on clinical trials, if available.

Induction of Remission

The most widely used remission-induction regimen includes treatment with an anthracycline (usually daunorubicin) and cytarabine with or without etoposide. With these regimens, 85% to 95% of children with AML enter complete remission after receiving 1 to 2 cycles of induction chemotherapy. Because the remission-induction phase of therapy is associated with prolonged cytopenias (3–5 weeks), 2% to 5% of patients may die of infectious or hemorrhagic complications before completing the induction phase. Deaths during the first several days after diagnosis are rare and often are caused by leukostasis or DIC. Leukostasis, or the plugging of blasts in vessels, is associated with elevated peripheral blast counts (more than 100,000/μL) and can cause hemorrhagic infarction of the brain or other organs. A greatly elevated leukocyte count is a medical emergency, and measures should immediately be taken to decrease the leukocyte count with chemotherapy (eg, hydroxyurea), exchange transfusion, or leukopheresis. Symptoms of severe leukostasis include hypoxemia and mental status changes.

Supportive care measures during all phases of therapy for AML are critical. They include inpatient admission for observation during periods of neutropenia, indwelling central venous access, antiemetic agents, psychosocial support for the child and family, monitoring for the metabolic consequences of leukemic cell lysis (tumor lysis syndrome), empiric therapy for fever and neutropenia with broad-spectrum antibiotics, prophylactic platelet transfusion, administration of allopurinol or rasburicase in situations of very high leukemic counts or hyperuricemia, and prophylaxis of Pneumocystis jirovecii pneumonia and fungal infections. Use of hematopoietic growth factors has not increased remission rates or overall survival rates among children with AML. In some studies, use of these factors has been associated with slightly less infectious morbidity during periods of neutropenia.

Central Nervous System Therapy

In contrast with ALL, management of occult CNS leukemia with intrathecal chemotherapy alone or combined with cranial irradiation has not been shown to improve overall survival among children with AML. Most AML protocols, however, include intrathecal chemotherapy because isolated CNS disease has been found in approximately 20% of children who receive no CNS-directed therapy. Between 3% and 20% of children with newly diagnosed AML have been reported to have leukemic blasts in the cerebrospinal fluid; they are treated with weekly intrathecal chemotherapy until the cerebrospinal fluid is cleared of leukemic blasts. Cranial radiation is not required.

Treatment in Remission

Even with maximally intensive therapy, the 5-year leukemia-free survival rates is only 50% overall, but up to 75% in some groups with low-risk AML. The optimal intensity and duration of postremission chemotherapy remain under active investigation. Most effective consolidation regimens employ high-dose cytarabine, often in combination with an anthracycline or etoposide. Cumulative doses of anthracyclines will exceed 400 mg/m² for most children with AML. Such high cumulative doses of this class of drug are necessary for cure but place the survivor at risk for early and late cardiac complications.

Bone Marrow Transplantation

The use of marrow ablative doses of chemotherapy with or without total-body irradiation, followed by HSCT from a histocompatible family donor, in the care of children with AML was first attempted in the mid-1970s. Although there is a statistically significant disease-free survival advantage for allogeneic HSCT compared with chemotherapy, HSCT-related mortality remains a significant challenge. Currently, allogeneic transplant is offered only to those patients in first remission with high-risk AML (discussed below).

RISK STRATIFICATION AND PROGNOSTIC MARKERS

Risk for relapse in childhood AML is predicted by 2 main factors: AML genetics and response to induction therapy. Patients at low risk for relapse include those with CBF-AML [t(8;21) and inv(16)] as well as those with NPM1 and CEBPA mutations. Disease-free survival in these patients is approximately 75%. High-risk AML can be defined as AML with a monosomic karyotype (monosomy 5 or 7) or AML with FLT3/ITD (other less common molecular and karyotypic markers that predict a high risk for relapse are discussed below). Additionally, any patient with residual leukemia detected at the end of a single course of induction chemotherapy (minimal residual disease, or MRD) is considered high risk unless they have a low-risk genetic marker. Disease-free survival in high-risk patients is less than 25%. Standard-risk patients include those with a genetic mutation not considered high risk or low risk and no detectable MRD at the end of a single cycle of induction therapy. Standard-risk patients have a disease-free survival rate of approximately 50%.

Disease characteristics inherent to AML include factors such as diagnostic white blood cell (WBC) count, morphologic classification (FAB subtype), and biological characteristics, such as cytogenetics or gene mutations. Although the WBC count has been demonstrated to be a prognostic factor in AML, where those with a high WBC count have a worse outcome, it has been determined that WBC count reflects the underlying biology of the disease, and in the context of cytogenetics and other molecular alteration, it is not an independent predictor of relapse.

AML Karyotypes

Diagnostic cytogenetics is widely recognized as 1 of the most significant prognostic factors in AML, distinguishing favorable risk patients from those at high risk of relapse. Two of the most commonly identified translocations in pediatric AML are the CBF mutations, t(8;21) and inv(16), accounting for 15% to 20% of childhood AML. Numerous adult and pediatric studies have demonstrated that patients with t(8;21) or inv(16) have superior outcome compared to other AML patients.

Karyotypes associated with poor outcome have been identified in a smaller proportion of pediatric patients with AML. Monosomy 7 (–7), monosomy 5 (–5), deletion of the q arm of chromosome 5 (del5q), as well as t(6;9), which collectively account for approximately 6% to 8% of the cases of childhood AML, are strongly associated with poor remission induction and high relapse risk. Recently, COG studies demonstrated that deletions and translocations involving ETV6 (chromosome 12p) predict adverse outcomes. These studies also demonstrated that the specific translocation fusion partner predicts the prognosis in patients with KMT2A mutations. Only 11q23 alterations involving MLLT1, MLLT4, MLLT10, and ABI1 have an adverse outcome.

Somatic Mutations

Until recently, cytogenetic alterations were the only means of risk identification in AML. However, the majority of children with AML lack clinically informative karyotypes and are not amenable to risk-based therapy allocation based solely on cytogenetic subgroups. More recently, somatic mutations in FLT3, NPM1, and CEBPA have been shown to correlate with outcome and were incorporated into clinical practice for treatment allocation. Cumulatively, clinically significant cytogenetic alterations and somatic mutations account for nearly 35% of AML in children; however, the majority of children with AML still lack a prognostic biomarker. Although other clinically significant mutations have been demonstrated in adult AML, such mutations are either not seen in childhood AML (eg, IDH1 or DNMT3A mutations) or are not independently prognostic (WT1 mutations).

Minimal Residual Disease

In patients without an informative genomic biomarker, response to therapy as defined by the presence or absence of MRD based on detection by multidimensional flow cytometry (MDF) has been utilized to inform relapse risk. The combination of diagnostic molecular risk factors with the postinduction response assessment by MRD detection has allowed risk assessment in all patients with AML. This strategy allows for patients with informative molecular markers to be assigned to the appropriate risk class and those lacking any prognostic biomarkers to be risk-stratified based on the presence or absence of MRD at the end of induction therapy. Such an approach allows appropriate risk-based therapy allocation for nearly all patients by creating a 2-tier risk allocation system where patients are assigned to either a high-risk (cytogenetic/molecular high risk or standard risk with MRD) or a low-risk (cytogenetic/molecular low risk or standard risk without MRD) arm based on the most comprehensive set of prognostic data available. The utility of such an approach is in the allocation of patients for HSCT. Allogeneic HSCT is currently recommended for all high-risk patients. It is important to note that low-risk AML is still associated with a 30% to 40% risk for relapse with conventional chemotherapy alone. Emergence of novel biomarkers will further enhance the existing risk-based therapy schema, where those with either high risk of relapse or those with appropriate targets for directed therapy can be allocated to an alternate therapy to optimize outcome and minimize undue toxicity.

Recurrent or Refractory AML

The prognosis is poor for children who do not enter remission with an anthracycline-cytarabine regimen or who have a relapse. Allogeneic HSCT offers these patients the best chance for long-term survival.

MYELODYSPLASTIC SYNDROME AND MYELOPROLIFERATIVE SYNDROMES

MYELODYSPLASTIC SYNDROME

MDS and MPS represent heterogeneous groups of clonal hematopoietic disorders. Combined, they represent less than 10% of myeloid malignancies in children. MDS is characterized by ineffective hematopoiesis, dysplastic maturation of bone marrow progenitors, and increased apoptosis. Mild to severe cytopenias result from these changes that in turn lead to the need for red blood cell and platelet transfusions. These are accompanied by increased susceptibility of infection due to neutropenia. The bone marrow of patients with MDS is usually hypercellular and displays dysplastic changes of myeloid precursors. Hypercellularity and common chromosomal abnormalities, such as monosomy 7, deletion of chromosome 5q, or trisomy 8, distinguish it from severe aplastic anemia. MDS is a preleukemic syndrome in that a high proportion of patients progress to AML over months to years.

The outcome for patients with MDS is largely based on the disease characteristics, including the degree and extent of dysplasia, the requirement for transfusions, the blast count, and cytogenetics. The DNA methyltransferase inhibitor, 5-azacytidine, has been approved by the US Food and Drug Administration (FDA) for the treatment of adults with MDS. This agent significantly improves hematopoiesis and reduces the need for transfusions and the rate of conversion to AML. There are no strong data that this type of treatment can cure patients with MDS.

JUVENILE MYELOMONOCYTIC LEUKEMIA

JMML is a disorder unique to children with some features of MDS and MPS. It is characterized by leukocytosis, an absolute monocytosis, a high fetal hemoglobin level, marrow hypercellularity and dysplasia, and myeloid progenitors that display granulocyte-macrophage colony-stimulating factor (GM-CSF) hypersensitivity. JMML most commonly occurs in children under age 2 years and is frequently associated clinically with lymphadenopathy, hepatosplenomegaly, and skin changes that include xanthoma, café au lait spots, and eczematous rashes. Usually, the karyotype is normal in JMML. However, somatic activating mutations in NF1, PTEN, NRAS, KRAS, or CBL can be identified in more than 90% of cases of JMML. Often, JMML occurs in the setting of a germline mutation in 1 of these genes. Approximately 50% of children with JMML have long-term survival when treated with HSCT.

MYELOPROLIFERATIVE SYNDROMES

The MPS are characterized by increased proliferation and survival of hematopoietic progenitors, resulting in very high peripheral blood counts depending on the type of disorder. For example, in essential thrombocythemia, the platelet count is extremely high, while in polycythemia vera, the red blood cell mass is expanded. These 2 disorders are characterized molecularly by mutations in signal transduction gene products, particularly JAK2. CML, characterized molecularly by the chromosomal translocation t(9;22) and fusion of the genes for BCR-ABL, is by far the most frequently observed MPS in children and is characterized by chronic, accelerated, and blast crisis phases. A chronic phase, which usually lasts 3 to 5 years, typically presents with a profound leukocytosis with all stages of granulocytic differentiation present, frequently with thrombocytosis and hypercellular bone marrow demonstrating normal granulocytic maturation. The accelerated phase is characterized by increasing percentages of leukemic blasts or immature progenitors in the peripheral blood and bone marrow, basophilia, and increased chromosomal abnormalities. Splenomegaly and thrombocytopenia usually develop or worsen. Blast crisis is when CML converts to a picture that is indistinguishable from acute leukemia.

Tyrosine kinase inhibitors (TKIs), like imatinib, have made CML a chronic disease. The optimal treatment duration has not yet been studied in children, and many children may be on a TKI for life. Some pediatric oncologists still favor the use of HSCT, especially in patients with a well-matched human leukocyte antigen (HLA) donor, and a history of CML refractory to 1 or more TKIs.

SUGGESTED READINGS

INTRODUCTION

Lymphoma represents the third most common pediatric cancer in children between the ages of 0 and 19 years. Sixty percent of all childhood and adolescent lymphomas are classified as non-Hodgkin lymphoma (NHL). NHL represents approximately 10% of all malignancies that occur in children and adolescents. Between 1975 and 2010, the 5-year survival rate for NHL increased from 45% to 87% in children younger than 15 years and from 48% to 82% for adolescents aged 15 to 19 years.

PATHOGENESIS AND EPIDEMIOLOGY

NHL occurs most commonly in the second decade of life, and occurs infrequently in children younger than 3 years. The incidence of childhood NHL is more common in males than in females and is higher in whites than in nonwhites. Risk factors for the development of NHL in children and adolescents are largely unknown, except that children with either inherited and/or acquired immunodeficiencies have a significantly increased risk for developing NHL.

As opposed to NHL seen in adults, where indolent and intermediate subtypes predominate, over 95% of childhood and adolescent NHL are intermediate or aggressive subtypes. Types of childhood NHL according to World Health Organization (WHO) classification are (1) mature B-cell lymphoma, including Burkitt lymphoma (BL; 40% of pediatric NHL) and diffuse large B-cell lymphoma (DLBCL) and variants (15%), and (2) mature T- and natural killer (NK)-cell lymphomas, including anaplastic large-cell lymphoma (ALCL; 15%). Lymphoblastic lymphoma (LL; 20%) is classified with its precursor acute lymphoblastic leukemia (ALL) in the WHO classification. The other 10% of cases of NHL observed in childhood include lymphoproliferative disease in immunodeficient patients, indolent B-cell lymphoma, and other mature T-cell lymphomas.

CLINICAL MANIFESTATIONS

Less than 10% of childhood and adolescent cases of NHL presents in peripheral nodal tissue. The majority will present with extranodal disease, including lymphoid tissue of the head and neck or gastrointestinal tract. Other presenting sites include the liver, spleen, bone marrow, cerebrospinal fluid, skin, bone, gonads, or central nervous system (CNS). The most common sites of presentation include the head and neck (30%), abdomen (30%), and mediastinal and/or hilar lymph nodes (30%). Each pathologic subtype has distinct primary organ sites of presentation as outlined in Table 447-1.

DIAGNOSIS

Most patients present with nonspecific signs and symptoms, such as fatigue, fever, or pain; therefore, the differential diagnosis includes infections as well as other malignancies, such as leukemia. In addition to a tissue biopsy, the diagnostic workup of children or adolescents suspected of having NHL includes studies and tests to determine stage of disease (eg, lumbar puncture, bone marrow aspirate and biopsy, and radiographic imaging), as well as complete blood count and serum electrolytes, including uric acid, liver, and renal function tests.

The Ann Arbor staging system for lymphoma, which is widely used in adult lymphomas, is less useful in predicting outcome in pediatric NHL, due to the propensity for extranodal involvement, noncontiguous lymphatic spread, and a higher propensity for metastatic spread to the bone marrow and CNS. Therefore, the St. Jude staging system is generally used for pediatric NHL (Table 447-2). In general, limited disease includes stages I and II, and advanced disease includes stage III and stage IV. Approximately 70% of children and adolescents with NHL have stage III or IV at the time of diagnosis. For treatment assignment, the major distinctions are (1) resected localized disease (good risk), (2) CNS or high burden (≥ 25% blasts) of marrow disease (advanced disease), and (3) all other disease, which may include stages 1 to 4 (intermediate risk). For LL, all stages are treated with ALL regimens (see Table 447-1).

TREATMENT

The primary therapeutic modality for all pathologic subtypes of childhood and adolescent NHL is multiagent chemotherapy. The role of surgery is quite limited in the treatment of childhood and adolescent NHL; however, it is critically important in establishing the diagnosis and for disease staging. Currently, radiotherapy is not used in the treatment of pediatric NHL except when required for emergency treatment in the management of large mediastinal masses precipitating acute respiratory distress and spinal cord compression, and for the treatment of CNS disease in children with LL.

Due to the rapid growth and lysis of tumor cells with therapy, tumor lysis syndrome (TLS) is a common complication of NHL, particularly in patients with BL or LL. TLS may result in marked hyperkalemia, hyperphosphatemia, hypocalcemia, and hyperuricemia, which can lead to renal failure, cardiac arrhythmias and dysfunction, and/or severe neurologic complications. Although TLS usually occurs following initiation of therapy, it may develop before therapy is started. The mainstay of prophylaxis and treatment of TLS includes aggressive hydration, frequent monitoring of electrolytes, diuresis, and use of an agent to prevent or reduce hyperuricemia, either allopurinol or recombinant urate oxidase.

SPECIFIC TREATMENT AND PROGNOSIS FOR PEDIATRIC NON-HODGKIN LYMPHOMA

Mature B-Cell Non-Hodgkin Lymphoma

In contrast to adult patients, the treatment for DLBCL and BL in childhood and adolescence is the same and is described below. Recent molecular genetic studies have demonstrated substantial overlap between DLBCL and BL in children than that observed in adults with these disease entities.

The prognosis of childhood and adolescent mature B-cell NHL is excellent. Patients with completely resected limited-stage disease have a 99% survival rate and require only 3 weeks of chemotherapy. Those with intermediate-risk disease have over 90% survival rate with approximately 12 weeks of chemotherapy. Patients with advanced disease, that is, extensive bone marrow or CNS involvement, have had an 85% overall survival rate with 4 to 6 months of therapy (see Table 447-1). Recent studies suggest that the addition of the monoclonal B-cell antibody rituximab improves the outcome for intermediate and advanced disease.

Though results are excellent for first-line therapy, it is very difficult to control recurrent or refractory B-cell NHL. Survival rates for patients with recurrent or refractory disease is less than 25%, even with aggressive therapy, including hematopoietic stem cell transplantation (HSCT).

Primary mediastinal B-cell lymphoma (PMBL) accounts for 1% to 2% of childhood NHL, is predominantly seen in older adolescents, and is more common in females. PMBL was previously considered a subtype of DLBCL, but is now a separate entity in the recent WHO classification. Differentiating PMBL from DLBCL and Hodgkin lymphoma can be very challenging even for the most experience pathologists. As opposed to DLBCL, survival following treatment of PMBL is 65% to 85%.

Anaplastic Large-Cell Lymphoma

Diagnosis of ALCL can be difficult due to waxing and waning symptoms, and it can be difficult to differentiate from Hodgkin lymphoma, DLBCL, and some infectious diseases. All cases of ALCL are CD30-positive and more than 90% of pediatric cases have a chromosomal rearrangement involving the anaplastic lymphoma kinase (ALK) gene. The WHO classification recognizes 2 subtypes of ALCL: a primary cutaneous subtype and a systemic nodal and extranodal subtype. Primary cutaneous and ALK-negative ALCL have a very good prognosis requiring little therapy beyond resection. However, these tumors are very rare in children and adolescents. Children and adolescents with localized systemic ALCL have a 90% survival rate compared to patients with advanced disease, who have a 60% to 75% survival rate (see Table 447-1).

For patients who have refractory or progressive disease, the prognosis is better than with mature B-cell lymphomas, with a survival rate of about 50%. Treatments for refractory or relapsed disease range from single-agent chemotherapy, biologic agents such as conjugated monoclonal antibodies, and ALK inhibitors to aggressive chemotherapy and HSCT.

Lymphoblastic Lymphoma

LLs can be of either T-cell or B-cell lineage, but more than 75% have T-cell markers. Approximately 75% of children with LL have an anterior mediastinal mass at diagnosis, which may lead to a medical emergency at presentation. Depending on the age of the patient and the compressibility of the trachea and bronchi, a mediastinal mass may cause rapid onset of dyspnea, wheezing, stridor, or dysphagia. Swelling of the head and neck due to compression of the superior vena cava by the mass are also common manifestations of mediastinal involvement. Care must be taken to keep the child as calm as possible and to avoid supine positioning, which can result in respiratory collapse that cannot be relieved by tracheal intubation.

In contrast to other forms of NHL in children, in which intensive chemotherapy given over a period several months yields excellent results, the best results in LL include treatment with a relatively intense induction and consolidation therapy phase, followed by 2 years of much less intense maintenance therapy. The treatment strategy for patients with LL is the same for those with localized versus more advanced disease. Survival for LL in pediatrics is 80% to 90%. However, outcome for patients with refractory or recurrent disease remains poor, similar to mature B-cell lymphoma, with an overall survival rate of less than 25%.

NON-HODGKIN LYMPHOMA OF IMMUNODEFICIENCY

The incidence of lymphoproliferative disease or lymphoma is 100-fold higher in immunocompromised children than in the general population. It is estimated that as many as 100 to 150 new cases of NHL in immunodeficient children and adolescents are observed in the United States each year, making this group of disorders as common as DLBCL or ALCL. The immune deficiencies may be the result of a genetically inherited defect or human immunodeficiency virus (HIV) infection, or it may be iatrogenic following transplantation (solid organ transplantation or allogeneic HSCT). Epstein-Barr virus (EBV) is associated with most of these tumors, although some tumors are not associated with any infectious agent.

NHL in immunocompromised patients commonly occurs in extralymphatic sites, such as in primary CNS lymphoma, which rarely occurs in children who are immunocompetent. Regardless of the specifics of the underlying immune defect, immunodeficient children with NHL have a worse prognosis than those who are not immunocompromised. Immunodeficient children with NHL usually tolerate cytotoxic therapy poorly and have increased morbidity and mortality due to an increased rate of infectious complications and end-organ toxicities.

Recent studies demonstrate that children with HIV and NHL should be treated with standard chemotherapy regimens for NHL along with highly active antiretroviral therapy (HAART), but careful attention to prophylaxis against and early detection of infection are warranted. Patients with primary immunodeficiency can achieve complete and durable remissions, though correction of the underlying immunologic defect through allogeneic HSCT is often required to prevent recurrence.

Post-transplant lymphoproliferative disease (PTLD), a well-recognized complication after solid organ or hematopoietic stem cell transplant, is often associated with EBV reactivation or primary infection. PTLD is characterized by heterogeneous lymphoid proliferations, ranging from B-cell hyperplasia to polymorphic or monomorphic disease that histologically resembles NHL. The first-line therapy for PTLD is reduction of immunosuppressive agents to doses as low as possible. Treatment with anti–B-cell monoclonal antibodies has been used with some success. PTLD following organ transplant can be cured with a low-dose chemotherapy and anti–B-cell monoclonal antibodies.

SUMMARY

The cure rates for NHL in children and adolescents approach 70% to 90%, even for those with the most advanced disease. Treatments have been developed over the years to reduce the risk of long-term complications, though short-term toxicities remain problematic for these patients. The challenges and goals for the future are (1) to identify those patients who can maintain excellent outcomes with less therapy and/or the incorporation of novel biologic, more targeted agents, and (2) to identify patients at high risk for relapse and intervene prior to relapse in order to improve the currently dismal outcome following disease relapse.

SUGGESTED READINGS

INTRODUCTION

Hodgkin lymphoma (HL) is 1 of the most common cancers arising in adolescents and young adults, and 1 of the most curable. This was not the case when Thomas Hodgkin first described the “morbid appearances of the absorbent glands and spleen” in 1832; all 7 of the cases he described died of their disease, as no effective therapies were available. Today, the vast majority of adolescents and young adults diagnosed with HL will be long-term survivors, as a result of highly effective treatment regimens developed through cooperative group trials conducted during the last 5 decades.

The challenges of ongoing clinical research today are to advance the competing objectives of (1) further increasing the probability of long-term cures for all patients with HL while (2) eliminating the long-term effects of cytotoxic chemotherapy and radiotherapy, including organ dysfunction and risk of second malignancies. Highly active targeted agents, which are actively being developed as further insights into the abnormal biology of malignant lymphoma cells come to light, are moving forward to the clinic, with the goal of minimizing off-target organ toxicity. Risk stratification and response-based treatment approaches identify those who can be cured with less toxic regimens. As the bicentennial of Thomas Hodgkin’s initial description approaches, clinicians caring for patients with his eponymous disease are closer to being able to expect that all who are diagnosed will be cured without fear of residual disease- or severe treatment-related sequelae.

PATHOGENESIS AND EPIDEMIOLOGY

The predominant clinical presentation of classical HL (cHL) is notably different in developed versus developing countries, as well as between age groups. These and other epidemiologic observations suggest that the etiology of cHL is dependent on a complex interaction among environmental, infectious, and host factors.

Classical HL has a bimodal incidence, with a first peak among adolescents and young adults and another among adults in the seventh to eighth decade of life. The young adult form (ages 15–34 years) shows a predominance of the nodular sclerosing histologic subtype in white adolescents and young adults in developed countries. HL is uncommon among preadolescent children, where it is associated with poorer socioeconomic environments, male sex, Epstein-Barr virus (EBV) infection, and the mixed cellularity histologic subtype.

HL is more common in males than in females in all parts of the world. The gender ratio varies from 2:1 in Europe and America to over 3.5:1 in Asia. The male predominance is most marked in patients younger than age 10 years. In adolescents, the gender difference in incidence is less conspicuous, particularly for the nodular sclerosing histologic subtype.

A genetic predisposition to HL is suggested by the variation in incidence among racial and ethnic groups, familial aggregation of the disease, and association with specific human leukocyte antigens. Many investigators have observed concordance of HL in first-degree relatives, including sibling and parent–child pairs. Standardized incidence ratios range from 3-fold for parent-child pairs to over 50-fold for monozygotic twins. HL also develops more frequently in individuals with congenital or acquired immunodeficiency, leading to the speculation that an inherited subtle immune abnormality may predispose to the development of HL by increasing the risk of malignant transformation induced by environmental factors. Genome-wide association studies have identified HLA and non-HLA susceptibility loci.

EBV was indirectly implicated in epidemiologic studies demonstrating an excess risk of HL following infectious mononucleosis and elevated anti-EBV titers in patients with HL compared to healthy controls. Supporting evidence of EBV association with HL includes the localization of viral genomes in the Reed-Sternberg cells, expression of EBV-latent gene products in the tumor specimens, and the demonstration of clonal infection. EBV infection increases expression of PD-1 ligand, suppressing immune clearance of developing malignant cells. Clear differences exist with the expression of EBV-associated antigens in cases of HL with respect to histologic subtype, geographic region, and age. EBV–associated HL occurs more frequently in developing countries in children who present with mixed cellularity histology, as well as in patients with primary or acquired immunodeficiency disorders. Taken as a group, these data suggest that EBV may be a cofactor in the pathogenesis of some cases of HL.

CLINICAL MANIFESTATIONS

The most common clinical presentation of HL is painless cervical or supraclavicular lymphadenopathy. Involved nodes are typically very firm, matted, and nontender. Mediastinal involvement (present in up to two-thirds of adolescent cases) may cause a positional cough due to airway compression. Axillary or inguinal lymphadenopathy is less common. Bone or bone marrow involvement is infrequent, and is most often asymptomatic.

Approximately one-quarter of adolescents and young adults with HL will experience constitutional symptoms, a result of cytokine release by the malignant cells. These “B” symptoms can include fevers, drenching night sweats, and unexplained weight loss. The classic fever pattern, recurring cycles of high fevers for 5 to 7 days alternating with an equal period without fevers, has carried the eponym Pel-Ebstein fever since its nineteenth century descriptions. However, this classic fever pattern is rarely, if ever, observed in clinical practice. B symptoms are associated with more extensive disease and an inferior prognosis. Intense pruritus in the absence of a rash is not an uncommon presenting symptom, but it has no prognostic impact so is no longer considered a B symptom.

Laboratory abnormalities in HL are nonspecific, usually indicating an inflammatory state, with activation of the reticuloendothelial system and subsequent acute phase reactants. Frequent findings include elevation of the erythrocyte sedimentation rate, ferritin, and C-reactive protein, as well as a normocytic anemia of chronic inflammation. Elevations of serum lactate dehydrogenase or alkaline phosphatase are less common.

Active autoimmune conditions are occasionally coexistent at the time of diagnosis with HL. The most common among these are autoimmune cytopenias: Coombs-positive hemolytic anemia and immune thrombocytopenic purpura (ITP). Both tend to resolve with the initiation of HL-directed therapy. Other abnormalities of the immune system can either predate or present concurrently with a diagnosis of HL, including enhanced sensitivity to suppressor T-lymphocytes and reduction of natural killer cell cytotoxicity, leading to deficient cell-mediated immunity.

The differential diagnosis of persistent lymphadenopathy in children and adolescents is lengthy, dominated by nonmalignant infectious or inflammatory conditions that in aggregate are much more common than lymphoma. Specific infectious etiologies (eg, atypical mycobacteria, toxoplasma, cat scratch disease, EBV, human immunodeficiency virus) may be suspected on the basis of their characteristic clinical presentations or specific laboratory testing. Non-Hodgkin lymphoma may have similar presenting signs and symptoms, but a more rapidly progressive course and more frequent elevations of serum uric acid and lactate dehydrogenase levels. Any supraclavicular nodes as well as any abnormally large nodes in other regions that persist beyond 3 or 4 weeks warrant further investigation, and consideration should be given to an excisional biopsy.

DIAGNOSIS AND STAGING

A definitive diagnosis of HL can be made only by histologic confirmation. Most frequently this is achieved with an excisional lymph node biopsy. Noninvasive needle biopsies are typically insufficient, as they do not provide material adequate to view the malignant cells, the Hodgkin and Reed-Sternberg (HRS) cells, within the context of surrounding stromal cells and nodal architecture.

HL includes at least 2 distinct disease entities that differ in histologic appearance, immunophenotype, epidemiology, etiologic risk factors, and treatment approaches: nodular lymphocyte–predominant HL (NLPHL) and cHL. This chapter will focus primarily on the latter, as NLPHL is a relatively rare condition, accounting for approximately 5% of all cases of HL. NLPHL tends to present in an early stage of disease, with few systemic symptoms and a strong male predominance (approximately 3:1). In contrast with cHL, the malignant cells of NLPHL typically express the B-lymphocyte marker CD20, which confers sensitivity to agents targeting this surface protein, such as rituximab.

A unique feature of cHL is the relative rarity (0.1–5%) of the malignant HRS cells within the tumor mass. The HRS cells are seen surrounded by infiltrating immune cells and eosinophils, all within the context of abnormal nodal architecture.

HRS cells are thought to derive from germinal center B cells that have lost their mature B-cell features, including expression of the mature B-cell marker CD20. Instead, CD30, a tumor necrosis factor family member and transmembrane receptor, is present. One of the more frequent genetic changes in cHL is amplification of the 9p24.1 locus, leading to overexpression of the tyrosine kinase JAK2, as well as ligands for the receptor PD-1.

This latter change, overexpression of PD-1 ligands, appears to underlie an intriguing biologic feature of cHL—the ineffective nature of the robust immune infiltrate within involved nodes. When activated by its ligand, the PD-1 receptor on T cells inhibits activation and proliferation. Consequently, when PD-1 ligand expression by HRS cells is driven by JAK2 overexpression and/or by EBV infection, HRS cells appear capable of inactivating an effective antitumor immune response, thus evading immune surveillance.

Three main histologic subtypes of cHL exist: nodular sclerosis, mixed cellularity, and lymphocyte depleted. Nodular sclerosis, which has a slight female predominance, is the most frequent subtype, especially in adolescents and young adults. A syncytial variant of nodular sclerosis cHL, characterized by sheets of lacunar cells, has been associated with an inferior prognosis in some reports, although not in children. The mixed cellularity histology is more common among younger children (younger than age 14 years), and is associated with lower socioeconomic status, EBV infection, and immunodeficiency. It is the most common form of HL in the developing world.

Accurate clinical staging is a process of identifying all involved sites of disease for the purposes of (1) risk assessment, allowing stratification of planned treatment intensity; (2) establishing a baseline for response evaluation, necessary for refinement of subsequent treatment; and (3) planning radiation fields, for those patients whose treatment requires the use of this modality. Clinical staging typically involves the following: a thorough physical examination with documentation of all abnormal lymphadenopathy; posteroanterior (PA) and lateral chest radiographs; computed tomography (CT) scans of the neck and chest; CT or magnetic resonance imaging (MRI) scans of the abdomen and pelvis; and a fluorodeoxyglucose positron emission tomography (FDG-PET) scan. Bilateral bone marrow biopsies may be done if marrow involvement is suspected. Some studies suggest that MRI and/or functional imaging studies have higher sensitivity and specificity for bone or marrow disease. These noninvasive studies may one day supplant marrow biopsies in staging algorithms. Surgical staging, including laparotomy and splenectomy, is not recommended.

Table 448-1 outlines the Ann Arbor staging classification, which has been the internationally accepted staging system for HL since 1971. The pediatric stage distribution is I, 19%; II, 49%; III, 19%; and IV, 13%, similar to young adults.

TREATMENT APPROACHES

Initial Diagnosis

The major focus in pediatric HL is the development of various strategies aimed at identifying the optimal balance between maintaining high rates of overall survival and avoiding the long-term morbidity and mortality associated with HL-directed therapy, with treatment protocols that are often quite different from those used for adults with HL.

At diagnosis, pediatric patients are classified into risk groups based on stage and the presence of clinical, biologic, and serologic risk factors. The details of risk group allocations differ substantially across cooperative research groups (Fig. 448-1). Nevertheless, the standard across groups is to direct those patients with the lowest predicted risk of relapse toward the least intensive therapy, with the objective of minimizing both acute and persistent treatment-related organ dysfunction. Ongoing trials for patients with favorable disease presentations are evaluating the effectiveness of treatment with fewer cycles of combination chemotherapy that limit doses of anthracyclines and alkylating agents.

Several trials have established the efficacy of treatment with non–cross-resistant chemotherapy, although no single combination has emerged as a standard used throughout the pediatric oncology world. In North America, the Children’s Oncology Group is primarily utilizing ABVE-PC (doxorubicin [Adriamycin], bleomycin, vincristine, etoposide, prednisone, and cyclophosphamide) and its derivatives across the risk groups. Advantages of this regimen include a resultant lower cumulative dose of anthracyclines, alkylating agents, and bleomycin compared with the MOPP (mechlorethamine, vincristine [Oncovin], procarbazine, and predisone) or ABVD (doxorubicin [Adriamycin], bleomycin, vincristine, and dacarbazine) regimens and their variations used in prior decades. This approach is expected to translate into a long-term reduction in second malignant neoplasms, as well as cardiovascular, pulmonary, and fertility complications.

Early response to chemotherapy is currently being studied as a means of further refining therapy planning: directing those with an inadequate response toward more intensive therapy and reducing therapy for those with a rapid early response. Clinical trials by the Children’s Oncology Group, for example, have demonstrated that radiotherapy can be safely omitted for those with an adequate response to chemotherapy. Fluorodeoxyglucose PET has become an established technique for assessing response to treatment by detecting metabolic activity and distinguishing between residual disease and necrosis or fibrosis. It is important to emphasize that interim PET/CT has not been validated as a predictive endpoint and its use in this setting remains investigational. Reduction in tumor mass as measured by 2-dimensional area or volume on CT, or in total metabolic volume on PET may add predictive value over PET alone.

Radiotherapy has historically been an essential component of HL therapy. Whereas radiotherapy is routinely incorporated into the treatment plan for adults with early-stage disease who are treated with ABVD chemotherapy, pediatric regimens omit radiotherapy for rapid or complete responders or utilize reduced radiotherapy dose and or volume strategies. The risk of radiation-induced secondary cancers, cardiovascular disease, and thyroid dysfunction throughout life is the primary driver of the different approach for pediatric HL.

Relapsed or Refractory Disease

Most recurrences of HL occur within 3 years after initial diagnosis, although some patients may relapse as long as 10 years after initial diagnosis. Treatment and ultimate prognosis depend on initial staging and treatment, time to relapse, extent of disease at relapse, and presence of B symptoms at relapse.

Patients with lower-risk relapses (eg, those with late relapse of low-stage disease) may be cured with salvage chemotherapy, with or without consolidative radiotherapy. Those with primary refractory disease or very early progression (< 3 months from the end of primary therapy) tend not to respond to conventional salvage therapy and have a poor prognosis for long-term disease-free survival (30–55%). For those with higher-risk relapses or primary refractory disease who can achieve a complete remission with chemotherapy, a consolidation with myeloablative therapy followed by autologous stem cell rescue improves disease-free survival over chemotherapy alone.

Multiple salvage regimens have been used in this context, including ICE (ifosfamide, carboplatin, and etoposide), IV (ifosfamide and vinorelbine), MIED (high-dose methotrexate, ifosfamide, etoposide, and dexamethasone), and GV (gemcitabine and vinorelbine). Although these regimens have never been compared head to head, all have similar overall response rates. As with initial therapy, clinicians must weigh treatment intensity against toxicity, often choosing to start with the least toxic regimen and advancing to more aggressive regimens if the initial response is inadequate.

Reduced-intensity or nonmyeloablative allogeneic hematopoietic stem cell transplantation (HSCT) is under evaluation as a retrieval therapy for children with recurrent/refractory disease after autologous HSCT. Nonmyeloablative conditioning regimens most often use a nontoxic immunosuppression with the goal of establishing a graft-versus-lymphoma effect that provides a platform for adoptive cellular immunotherapy.

Introduction of Novel Agents

With recent insights into the biology of HL, a large array of novel, targeted treatment strategies have been developed (Fig. 448-2). Two of these have received US Food and Drug Administration (FDA) approval for use in adults with relapsed disease, based on single-agent overall response rates in excess of 70%. Both are currently being investigated in combination with conventional chemotherapy for adults and/or pediatric patients with HL in the newly diagnosed and relapsed/refractory settings.

Brentuximab vedotin is an antibody-drug conjugate, combining a chimeric anti-CD30 targeting antibody with monomethyl auristatin E, a tubulin inhibitor. After binding CD30, on the HRS cell surface, brentuximab vedotin is internalized, the peptide linker is selectively cleaved, and the drug is released into the cytoplasm where it induces apoptosis. The efficacy of brentuximab vedotin in conjunction with conventional chemotherapy is being investigated for newly diagnosed and relapsed pediatric HL in clinical trials in North America.

Nivolumab is a human monoclonal antibody directed against PD-1. It therefore acts as a “checkpoint inhibitor,” releasing the inhibition of immune surveillance produced by PD-1 ligand-expressing HRS cells. Among adults with relapsed or refractory disease, it was well tolerated with an overall response rate of 87%. The efficacy of nivolumab may be further increased by an “abscopal effect,” in which radiotherapy or conventional chemotherapy causes the malignant cells to reveal additional or novel immune-stimulating antigens. The safety of checkpoint inhibitors such as nivolumab is currently being studied in pediatric patients with refractory or relapsed disease, and combination approaches are under development.

COMPLICATIONS AND LATE EFFECTS

Five-year survival rates for pediatric patients with cHL exceed 90%. However, long-term survivors experience treatment-related morbidity leading to reduced fertility and impaired neurocognitive, thyroid, cardiac, and pulmonary function. In addition, both chemotherapy and radiotherapy may induce secondary malignancies among HL survivors, at rates exceeding those seen among patients treated for other cancer types. The result of treatment-induced organ dysfunction, secondary malignancies, and late relapses of HL is a substantial increase in the absolute age-adjusted risk of mortality in the decades following curative treatment for HL. Thus, new approaches to current primary treatment have been undertaken to reduce these risks, especially second malignancies, cardiovascular events, and infertility. It is especially critical for survivors of childhood and adolescent HL to undergo lifelong monitoring for treatment-related organ toxicities, as, to date, there does not appear to be a plateau in the risk for their development.

CONCLUSIONS AND FUTURE DIRECTIONS

With improvements in response-based imaging techniques and identification of biologically based prognostic factors, titration of chemotherapy and RT to specific subgroups of patients at risk for tumor progression and/or specific long-term adverse effects of therapy will be more fully realized. Targeted agents without the same spectrum of late toxicities as conventional chemotherapy or RT may allow reduction or elimination of these conventional therapies, without a decrement in HL control rates. Although short-term survival rates remain among the highest of the childhood cancers, the ongoing elevated risk for late treatment-related chronic health conditions and premature death demonstrate the need for continued research in this field.

SUGGESTED READINGS