Key points

Introduction

Acute lymphoblastic leukemia (ALL) is the most common cancer diagnosed in children. It has an overall survival of approximately 80%, with certain subsets experiencing greater than 98% cure rate.

Incremental advances in therapy have led to marked improvements in survival since it was first treated, with these advances highlighting the importance of clinical trials through cooperative multicenter groups ( Table 1 ). Childhood ALL also often serves as the paradigm for risk-based therapy, whereby stratification of treatment intensity is based on risk of treatment failure ( Fig. 1 ).

Improved overall survival in childhood acute lymphoblastic leukemia (ALL).

( From Hunger SP, Winick NJ, Sather HN et al. Therapy of low-risk subsets of childhood acute lymphoblastic leukemia: when do we say enough? Pediatr Blood Cancer 2005;45(7):876–80; with permission.)

Introduction

Acute lymphoblastic leukemia (ALL) is the most common cancer diagnosed in children. It has an overall survival of approximately 80%, with certain subsets experiencing greater than 98% cure rate.

Incremental advances in therapy have led to marked improvements in survival since it was first treated, with these advances highlighting the importance of clinical trials through cooperative multicenter groups ( Table 1 ). Childhood ALL also often serves as the paradigm for risk-based therapy, whereby stratification of treatment intensity is based on risk of treatment failure ( Fig. 1 ).

Improved overall survival in childhood acute lymphoblastic leukemia (ALL).

( From Hunger SP, Winick NJ, Sather HN et al. Therapy of low-risk subsets of childhood acute lymphoblastic leukemia: when do we say enough? Pediatr Blood Cancer 2005;45(7):876–80; with permission.)

Risk stratification of newly diagnosed acute lymphoblastic leukemia

One of the hallmarks of the treatment of childhood ALL is the reliance on risk-based stratification. By identifying the features that have been shown to affect prognosis, patients can be classified into groups based on risk of treatment failure. Those with favorable features can be treated with less toxic regimens, whereas more aggressive regimens are reserved for those with more high-risk disease.

It is therefore paramount to identify those features shown to consistently affect prognosis and, thus, influence treatment. Several clinical characteristics have been shown to aid in this classification, including age and white blood cell count (WBC) at presentation, together referred to as the National Cancer Institute (NCI) criteria. Age between 1 and 10 years is a standard risk feature, with more aggressive disease seen in infants and those older than 10 years. In part this is due to the higher rate of favorable cytogenetics in those aged 1 to 10 years. The initial WBC at presentation also has been directly associated with increased risk, with high risk noted with WBC greater than 50,000/μL. Of importance is that this is a continuous function, but for practical purposes this threshold has been chosen as a useful categorical cutoff. The application of the NCI criteria results in those aged 1 to 10 years with initial WBC less than 50,000/μL classified as standard risk, with those not meeting those parameters classified as high risk.

Sanctuary sites are extramedullary anatomic locations that have historically been difficult to penetrate with systemic chemotherapy, and involvement of these sites at initial diagnosis has thus also been considered a high-risk feature. Approximately 3% of patients will demonstrate overt central nervous system (CNS) disease at diagnosis, defined either as a diagnostic lumbar puncture with the presence of leukemic blasts on cytospin and greater than 5 leukocytes/μL, or clinical evidence of CNS involvement (such as a cranial nerve palsy). Approximately 2% of boys with newly diagnosed ALL will present with testicular involvement, usually presenting with an enlarged, nonpainful testis. Leukemic involvement of the CNS or testis precludes a classification of standard risk in most treatment schemas.

Patients treated with corticosteroids before their complete diagnostic workup are also considered as high risk, as the tremendous efficacy of steroids to treat ALL may underestimate initial WBC and involvement of sanctuary sites, and limit confidence in staging.

The characteristics of the leukemia cells themselves can also be used to determine which patients are at higher risk. The immunophenotype describes the leukemic cells in terms of the proteins that are expressed, and whether these are more similar to cells that would eventually become B lymphocytes or T lymphocytes. This determination has also been shown to affect prognosis. At approximately 80%, most pediatric patients with ALL have B-precursor immunophenotype (Bp ALL), which encompasses a broad range of patients, including many of the lowest-risk patients with childhood ALL. Conversely, those with T-cell immunophenotype comprise approximately 10% to 15 % of pediatric ALL, and have historically been associated with a lower cure rate. However, identification of these patients and treatment with more aggressive regiments has led to survival rates that approach that of Bp ALL, with the possible exception of early T-precursor (ETP) ALL, a particular subset of T-cell ALL that has been associated with a poor prognosis in some studies. Additional rare immunophenotypic groups include those acute leukemias of mixed lineage, which occur in less than 5% of pediatric acute leukemias. These groups include undifferentiated acute leukemias that cannot be sufficiently characterized as either lymphoid or myeloid in origin, as well as those biphenotypic lineages that include markers of both myeloid and lymphoid origins and/or both B-cell and T-cell origins. These ambiguous immunophenotypes are often inconsistently defined but, regardless of exact classification, are associated with a poorer prognosis.

Recurrent cytogenetic abnormalities in the leukemic blasts allow a molecular classification of risk, with certain markers shown to be associated with favorable or unfavorable outcomes. The 2 most well-established favorable cytogenetic aberrations include high hyperdiploidy and the ETV6/RUNX1 translocation. High hyperdiploidy is seen in 20% to 25% of cases of Bp ALL. It is defined as 51 to 65 chromosomes per cell or a DNA index of greater than 1.16, and is particularly favorable when associated with simultaneous trisomy 4 and 10. The ETV6/RUNX1 translocation (due to t[12;21], formerly TEL/AML1) is also seen in approximately 20% to 25% of cases of Bp ALL, and is associated with improved survival, including improved survival even after relapse. Both favorable subgroups occur in lower frequency in African Americans (sub-Saharan Africans), and in part accounts for the lower overall outcome in this population. Several unfavorable cytogenetic changes have also been identified. One feature strongly associated with poor outcome is hypodiploidy, defined as fewer than 44 chromosomes or a DNA index of less than 0.81. Additional cytogenetic changes associated with higher-risk ALL include BCR-ABL fusion of t(9;22), known as the Philadelphia chromosome (seen in 3% of pediatric ALL), MLL rearrangements involving 11q23 (seen in 5% of pediatric ALL, often infants and adolescents), and, most recently identified, intrachromosomal amplification of chromosome 21 (iAMP21, seen in 1%–2% of Bp ALL).

In addition to these features that are used to inform prognosis, the response to the initial therapy has emerged as a particularly powerful independent predictor. Traditionally a complete remission has been defined as less than 5% detectable blasts on microscopic morphology at the end of induction. Induction failure is seen in approximately 3% to 5% of children with newly diagnosed ALL and portends a very poor prognosis, with an overall survival of approximately 33%. It is most closely associated with patients with T-cell immunophenotype, Bp immunophenotype with a high presenting leukocyte count, MLL rearrangement, Philadelphia chromosome, or older age.

Evaluation of the bone marrow by microscopy is often relatively insensitive, and has been shown to be complemented and, in part, displaced, by evaluation of minimum residual disease (MRD). This technique uses flow cytometry or the polymerase chain reaction (PCR) to assess for disease at a significantly lower limit of detection (1 leukemic blast in 10,000–100,000 cells). Evaluation of bone marrow MRD at the end of induction has proved to be an independent factor predicting outcome, and has also been shown to be useful in the peripheral blood as early as day 8 of therapy. End-induction MRD has been established in the risk stratification of Bp ALL patients while studies using MRD to adjust the treatment of T-cell ALL are currently ongoing but also promising.

The application of these risk factors is operationalized in various methods by the different cooperative groups of pediatric oncology specialists. One group, the Children’s Oncology Group (COG), utilizes a combination of the NCI criteria in addition to cytogenetics and response to therapy. Other groups, such as the Berlin-Franklin-Münster (BFM) Group, rely almost solely on response to initial therapy using MRD thresholds, although with certain cytogenetic changes treated as high risk regardless of response to therapy ( Box 1 ).

Treatment of newly diagnosed acute lymphoblastic leukemia

There are 4 major components of treatment of newly diagnosed ALL, reflecting a reliance on multidrug regimens to avoid development of resistance. Different blocks of chemotherapy have varying intensity depending on the group of patients at risk, with increasingly intensive regimens corresponding to more aggressive disease categories.

Remission induction is the first block of chemotherapy, lasting 4 to 6 weeks. Patients are usually admitted to the hospital for their initial treatment and workup, but once any complications have stabilized the patient may be discharged before the completion of this phase with close outpatient follow-up. The goal of this block of therapy is to induce a complete remission by its completion, with approximately 95% of all patients achieving this benchmark. Of those that do not achieve completion remission by the end of induction, half suffer induction failure and the remainder succumb to treatment-related mortality. For those with induction failure, an allogeneic bone marrow transplant is usually pursued, although there is no consensus standard of care regarding the chemotherapy used to achieve remission before transplant.

The agents used during induction include vincristine, corticosteroids, and asparaginase, with most regimens adding an anthracycline (usually doxorubicin or daunorubicin). Both anthracyclines have been shown to have similar efficacy and toxicity in randomized trials. Certain groups spare the addition of anthracyclines to those lower-risk groups in an effort to decrease toxicity. The corticosteroid used is usually prednisone or dexamethasone, with dexamethasone demonstrating improved CNS penetration and decreased risk of relapse, but with increased incidence of toxicities, including avascular necrosis, infection, and reduction in linear growth. Several different agents for asparagine depletion exist as well, including PEG asparaginase and Erwinia asparaginase. PEG asparaginase has been modified by covalently attaching polyethylene glycol, which has been demonstrated to result in a longer half-life and decreased immunogenicity in comparison with native Escherichia coli l -asparaginase. Randomized trials have also shown superior efficacy of the pegylated formulation. Erwinia asparaginase is usually given to those patients who have experienced an allergic reaction to PEG asparaginase, and requires a more frequent administration schedule.

Remission induction is followed by consolidation, which aims to eradicate the submicroscopic residual disease that remains after a complete remission is obtained. Lasting approximately 6 to 9 months, it varies in length and intensity among different protocols, with those patients with higher-risk disease receiving longer and more intensive consolidation regimens. Consolidation is usually administered on an outpatient basis, although there are protocols with more aggressive regimens that require inpatient care. This phase of chemotherapy involves combinations of different chemotherapeutic agents to maximize synergy and minimize drug resistance, often including agents not used in the initial remission induction, such as mercaptopurine, thioguanine, methotrexate, cyclophosphamide, etoposide, and cytarabine.

Maintenance chemotherapy is the final, and longest, stage of treatment in childhood ALL. A much less intensive regimen than the prior chemotherapy, the prolonged maintenance phase has been demonstrated to lower the risk of relapse once remission has been established. It usually lasts at least 2 years (extended to 3 years for boys in some protocols), is administered on an outpatient basis, and typically is associated with less disruptive toxicity. The cornerstone of maintenance therapy is antimetabolite therapy with methotrexate and mercaptopurine, both available in oral formulations, making strict adherence crucial. Furthermore, emerging evidence regarding the pharmacogenomics of these drugs underscores the importance of interindividual differences in metabolism. For example, genotypic polymorphisms in the enzyme thiopurine methyltransferase are associated with increased myelosuppression and other toxicities, whereas other polymorphisms confer a “hypermetabolizer” state, with decreased levels of the active metabolite. Understanding these differences in metabolism is particularly important because studies have shown that the degree of myelosuppression correlates with relapse risk. Accordingly, many protocols include guidelines for dose adjustments to assist in achieving the goal of balancing the risks of inadequate myelosuppression with the risks of severe pancytopenia (infection, bleeding, and so forth). Some regimens also include monthly vincristine and steroids, although the evidence for additional benefit is unclear.

The fourth component of the treatment of ALL is therapy directed against the CNS. This approach includes both treatment of patients with clinical CNS disease at diagnosis and prophylaxis for patients with subclinical disease. The importance of this component was clearly demonstrated before the 1970s, when treatment lacked this component. Although bone marrow remission could be achieved using systemic chemotherapy, most children eventually developed CNS relapse in the absence of specific therapy directed toward this sanctuary site.

There are several methods of achieving the goal of eradication of disease from the CNS, including direct intrathecal administration of chemotherapy, systemic administration of chemotherapy able to penetrate the blood-brain barrier, and cranial radiation. All treatment plans include intrathecal administration of chemotherapy beginning during remission induction. Some protocols include intrathecal treatment throughout therapy, whereas others do not include it in maintenance. Options for intrathecal chemotherapy include including intrathecal methotrexate or a combination of intrathecal methotrexate, cytarabine, and hydrocortisone (known as triple intrathecal). Studies have shown no definitive difference in overall or event-free survival between the two, although some evidence points to decreased frequency of CNS relapse with the use of triple intrathecal therapy. Systemically administered chemotherapy with CNS effects includes dexamethasone, high-dose methotrexate, cytarabine, and asparaginase.

Given the risk of toxicity of cranial radiation, manifesting primarily as intellectual disability (particularly with younger patients) and as second malignant neoplasms, its utilization has been progressively declining. Many protocols reserve its use for only those at highest risk of CNS relapse while some institutions defer its use altogether. For those patients with overt CNS disease at presentation, several small studies have shown that by increasing the intensity of the intrathecal and systemic chemotherapy, cranial irradiation can also be deferred for these patients. However, larger studies are needed to confirm this strategy.

The role of allogeneic hematopoietic stem cell transplant (HSCT) in first remission of ALL is not yet well defined, and is a controversial topic. Broadly speaking, HSCT is considered for those patients with the very highest risk of relapse and/or treatment failure, which has been most closely associated with those patients demonstrating hypodiploidy or induction failure. General tenets of HSCT for ALL include the use of total body irradiation (TBI) in the preparative regimen, and improved outcomes for patients who undergo transplant after achieving MRD-negative disease status. The optimal donor has historically been a matched sibling, although advances with alternative donor sources are now also showing promise.