Epidemiology

ALL accounts for 75% of childhood leukemia and is diagnosed in 2500 to 3000 children in the United States annually^1,2 with a peak age incidence between 2 and 5 years. The incidence of ALL is higher among boys than girls. This difference is more pronounced in pubertal children³ and particularly evident in cases of T-cell ALL. In the United States, ALL occurs more commonly among white children than among African-American children.⁴

Etiology

A single etiology of ALL has not been identified; however, several predisposing risk factors include genetics, environmental factors, viral infection, and immunodeficiency. Genetics appears to play a role in the cause of ALL in some children as evidenced by the association between specific chromosomal abnormalities and childhood ALL. The most common constitutional chromosomal abnormality associated with childhood leukemia is trisomy 21 or Down syndrome⁵; less common chromosomal abnormalities include Kleinfelter’s syndrome and trisomy G syndrome.⁶ The majority of cases of leukemia in children cannot be attributed to any known factor.

ALL is believed to develop from the malignant transformation of a single abnormal progenitor cell that then expands indefinitely. In pediatric ALL, evidence suggests that it occurs in committed lymphoid precursors. In contrast, in Philadelphia chromosome–positive ALL and in pediatric acute myelogenous leukemia (AML) this event may occur earlier, because mutations are observed in multiple cell lineages.^7,8 The sequence of events resulting in malignant transformation is likely to be multifactorial. Leukemia can develop along any point in the normal stages of lymphoid development, which reflects the fact that ALL is a biologically heterogeneous disorder.

Molecular Biology and Classification

ALL has been classified according to morphologic, immunologic, cytogenetic, biochemical, and molecular genetic characteristics. The most widely accepted morphologic classification of ALL is that proposed by the French-American-British (FAB) Cooperative Working Group.^9,10 The FAB system defines three categories of lymphoblasts:

The morphologic classification, thought useful in the past, is not currently a prognostic factor nor used in determining treatment, with the exception of the L3 morphology, which is almost always associated with mature B-cell phenotype and therefore requires therapy as for Burkitt’s lymphoma.

Immunophenotyping

The immunobiologic studies of ALL cells indicate that leukemic transformation and clonal expansion occurs at different stages of lymphocyte maturation. The use of monoclonal antibodies has classified leukemias; however, these monoclonal antibodies are not purely lineage specific and so the term lineage associated is used. In 80% to 85% of cases lymphoblasts are found to have surface markers consistent with B-cell lineage, although very few have surface immunoglobulin (found on mature B cells). Most cases of B-cell lineage ALL in children and adolescents are found to have the common ALL antigen (CALLA) CD10 on their cell surfaces.11,12 Patients with B-cell precursor ALL whose lymphoblasts express the common ALL antigen (CALLA) or CD10 on their cell surfaces have a more favorable prognosis than those who do not, primarily owing to the strong association of CD10 negativity with rearrangements of the MLL gene on chromosome 11.^13,14 Children with mature B-cell ALL (characterized by the presence of mature B-cell antigens, including surface immunoglobulin) have a poorer prognosis than those of earlier B-cell lineage if treated with standard ALL therapies; such patients are more appropriately treated on regimens for patients with advanced-stage Burkitt’s lymphoma. Ten to 15 percent of children will have ALL that is of T-cell origin. In contrast to B-cell precursor ALL, T-cell ALL is more frequently associated with older age at diagnosis, a higher presenting leukocyte count, and bulky extramedullary disease. Extramedullary involvement can include an anterior mediastinal mass, lymphadenopathy, hepatosplenomegaly, and overt CNS involvement.¹⁵

Cytogenetics

Advances in cytogenetics have led to an increasing understanding of the biology of ALL.¹⁶ Cytogenetic abnormalities identified in ALL include abnormalities in chromosomal number (ploidy) as well as those in chromosomal structure. Ploidy is determined either by counting the modal number of chromosomes in a preparation of a metaphase karyotype or by measuring DNA content by flow cytometry. Hyperdiploidy exists when there are more than 46 chromosomes. Most cases of ALL are diploid or hyperdiploid. Ploidy in childhood B-lineage leukemia has prognostic significance.^17,18,19 Children with higher ploidy, greater than 50 chromosomes, have a better prognosis than other children with ALL, including those with 47 to 50 chromosomes; those with 51 to 65 chromosomes (higher hyperdiploid) with specific trisomies (including chromosomes 4, 10, and 17) appear to have an especially favorable prognosis.^20,21,22 Those patients who have hyperdiploid leukemia with favorable trisomies also tend to have other favorable prognostic features (see later). Hypodiploidy (modal chromosomal number less than 44) is associated with inferior outcomes, especially among those with near-haploid ALL, with 24 to 28 chromosomes.^23,24,25

Structural chromosomal abnormalities occur in ALL leukemia cells. The most common of these abnormalities are translocations. Translocations lead to alterations in the regulation of oncogenes. The activation of oncogenes and the loss of tumor suppressor genes are examples of the altered regulation believed to be involved in the evolution of leukemia. The most frequently identified translocation in ALL is t(12;21) and leads to the TEL-AML1 fusion protein. This occurs in approximately 20% of cases (nearly exclusively in those with B-lineage phenotype) and is associated with a favorable prognosis.26,27 In contrast, t(9;22) and t(4;11) translocations are associated with early treatment failure.^19,28

The t(9;22)(q34;q11) translocation results in the formation of the Philadelphia (Ph) chromosome and is found in 5% of childhood ALL. Children with Ph chromosome–positive ALL have a poor response to standard therapy but may fare better with intensive regimens that include the use of a tyrosine kinase inhibitor such as imatinib.29,30,31,32,33

Other unique chromosomal translocations have been observed in childhood ALL. The role of these and all translocations in the development of leukemia remains speculative. Some translocations that result in altered gene expression appear to confer a proliferative advantage for the leukemic cells. When patients with ALL are in remission, the cytogenetic abnormalities just described are not detectable. When relapse occurs, reappearance of the original cytogenetic abnormality confirms that there is a recurrence of the original leukemia.

Presentation and Evaluation

The presenting signs and symptoms of ALL are related to the degree of bone marrow infiltration as well as the extent of extramedullary involvement. These can include fever, petechiae, lymphadenopathy, and bone pain. Hepatosplenomegaly is found in two thirds of patients. Children with T-cell ALL have some distinctive features. T-cell ALL more frequently occurs in older boys, presents as high initial white blood cell (WBC) counts, and often has a mediastinal mass at diagnosis. The definitive diagnosis is made by bone marrow aspiration, although a bone marrow biopsy may be required in some cases and allows for morphologic assessment, immunophenotyping, karyotyping, and molecular evaluations such as reverse transcriptase/polymerase chain reaction and fluorescent in-situ hybridization for chromosomal characterization.

At the time of initial diagnosis the evaluation of patients includes cerebrospinal fluid examination, evaluation of uric acid levels, liver function tests, renal function tests, coagulation screening, chest radiography, evaluation for infection, and echocardiography (Table 71-1). Manifestations of CNS involvement such as headache, lethargy, papilledema, and cranial neuropathies (most often the third, fourth, sixth, and seventh) are infrequently present at diagnosis. CNS involvement is divided into three categories, based on the number of WBCs and the presence of lymphoblasts on the Cytospin. CNS-1 is defined as no evidence of lymphoblasts in the cerebrospinal fluid. CNS-2 is defined as fewer than 5 WBCs per microliter with blasts present in the Cytospin, and CNS-3 is defined as 5 or more WBCs per microliter with blasts in the Cytospin or the presence of cranial nerve palsy. CNS-3 involvement of the CNS at diagnosis is found in less than 5% of children with ALL.³⁴

TABLE 71-1 Initial Evaluation

* Bone marrow studies: morphology, cytogenetics, immunophenotype, cytochemistry.

† Therapeutic lumbar puncture performed for diagnosis concurrent with administration of intrathecal chemotherapy.

Prognostic Factors

Several clinical and biologic factors determined at diagnosis have been identified and shown to have prognostic significance (Table 71-2). These prognostic factors are used to assign patients to risk strata that tailor the treatment. These factors have been integral in the development of all modern therapeutic trials. Although many of these factors have shown prognostic importance, not all are currently used to determine risk stratification.

TABLE 71-2 Prognostic Factors in Acute Lymphoblastic Leukemia

The presenting leukocyte counts and age at diagnosis are uniformly accepted as prognostic features. Children with the highest initial WBC count have a poorer prognosis.^35,36 Infants younger than 1 year of age have the worst prognosis, and adolescents also have a poorer event-free survival than younger children, except those younger than age 1 year.^37,38 Children older than 10 often have other poor risk features, including T-cell phenotype.³⁸ Infants with ALL have shown the worst outcome, with an event-free survival of 10% to 20%.³⁷ Infants also have other poor-risk features, including very high presenting WBC count, presence of CNS leukemia, massive organomegaly, and a poor day 14 response to induction chemotherapy.³⁷ Infant ALL has unique biologic features, including chromosomal abnormalities that are associated with a worse prognosis. Structural abnormalities of chromosome 11, such as rearrangement of band q23, within the MLL/ALL1 gene are often present.^39,40 The t(4;11) translocation is commonly seen, and many have leukemia cells that coexpress myeloid markers (CD15).^37,39,41 This finding suggests that infant ALL arises in a multipotent precursor cell. Chromosomal abnormalities in the leukemia cells also have prognostic significance. The presence of the Ph chromosome and of MLL rearrangements (even in noninfant patients) is associated with a higher risk of relapse. T-cell immunophenotype and the presence of CNS disease at diagnosis are also considered high-risk features. Early responses to initial therapy are also prognostic. These include the time to achieve remission and the decrease in peripheral blast count after administration of corticosteroids, when longer time to remission or poor response to initial corticosteroid therapy is associated with worse outcomes. Rapid morphologic clearance of blasts in the marrow 7 to 14 days after the initiation of multiple-agent chemotherapy has been associated with more favorable outcomes.

Nearly all current leukemia protocols include measurements of minimal residual disease at the end of induction chemotherapy. Multiple-parameter flow cytometry and polymerase chain reaction assays of leukemia-specific gene arrangements have been used to assess submicroscopic levels of minimal residual disease soon after the initiation of therapy.42,43,44,45 High levels of minimal residual disease at the end of the first month of therapy and other early time points has been associated with a very high risk of subsequent relapse.^{46,47,48,49,50}

On most protocols, children with high levels of minimal residual disease at the end of remission induction are now considered to be high or very high risk (regardless of other presenting features) and receive intensified therapy.

As therapy for ALL has become more intensive and the outcomes have improved, many of these factors have lost statistical significance as prognostic factors. In addition, the definitions of high leukocyte counts and age varied among cooperative groups. In order to allow comparisons of results, the National Cancer Institute held a workshop to develop a uniform set of prognostic factors.⁵¹ The age and WBC count criteria for childhood ALL agreed on at the Cancer Treatment and Evaluation Program/National Cancer Institute Workshop are shown in Table 71-3.

Treatment

The foregoing prognostic indicators are used to stratify children with ALL into risk groups. The stratification is dependent on prompt, accurate evaluation and sophisticated techniques. To accomplish this, as well as to treat patients with the increased intensity of current therapy, evaluation and treatment is most appropriately carried out in recognized pediatric oncology centers. The treatments for the different risk groups vary in intensity and by institution. However, the framework for all risk groups includes four key elements: remission induction, CNS preventative therapy, consolidation, and maintenance (Fig. 71-1).

Figure 71-1 Treatment algorithm for acute lymphoblastic leukemia. BMT, bone marrow transplant; CNS, central nervous system; 6-MP, 6-mercaptopurine; MTX, methotrexate. GAP-POMP (GAP refers to the schedule of mercaptopurine administration given on 14 days of each 21-day cycle, thus a 7-day GAP) repeats a 3-week cycle until 2 years of continuous complete remission: vincristine, prednisone, mercaptopurine, methotrexate; high-dose cycle; vincristine, mercaptopurine, high-dose intravenous and intrathecal methotrexate.

Induction treatment is designed to achieve complete remission, defined as no evidence of leukemia. Peripheral blood cell counts must be within the normal range, and the bone marrow must be of normal cellularity and contain less than 5% lymphoblasts. There should be no evidence of CNS disease on examination of the cerebrospinal fluid, and there must be no extramedullary disease. Remission can be achieved in 85% of children with ALL using vincristine and a glucocorticoid; however, the addition of L-asparaginase and/or the addition of an anthracycline will improve the remission induction rate to 95%.⁵² In general, most current regimens utilize a three- or four-drug induction phase, including vincristine, a corticosteroid, and L-asparaginase plus or minus an anthracycline.

The prevention of CNS disease is one of the major elements of successful ALL treatment. The approach to CNS prophylaxis has evolved over the decades since it was first incorporated into ALL therapy in the 1970s. Craniospinal irradiation and cranial irradiation plus intrathecal methotrexate reduced the relapse rate in the CNS from 50% to less than 10%.⁵³ To avoid the myelosuppression and the effects on spinal growth, intrathecal methotrexate was substituted for spinal irradiation as part of the standard CNS preventative therapy. Intrathecal therapy has been shown to successfully prevent CNS relapse in standard-risk patients and avoids the neurocognitive sequelae and secondary tumors that can result from cranial irradiation. Intrathecal chemotherapy in conjunction with intensive systemic therapy has also provided adequate CNS prevention for some children with high-risk features.

Children with an increased risk of CNS relapse include those with other high-risk features such as high initial WBC count, T-cell phenotype, very young age, t(4;11), the presence of the Ph chromosome, and lymphomatous presentation. For patients with these features, cranial irradiation (see later) is generally incorporated into the treatment protocol, with the exception of those with infant ALL, who are treated with very intensive chemotherapy to avoid the late sequelae of cranial irradiation in this age group. The CNS-directed therapy for CNS-3 patients is not preventative but therapeutic.

Once complete remission has been achieved, consolidation and maintenance therapy are used to provide continued cytoreduction of the leukemic cell burden without permitting the emergence of drug-resistant clones. Components of the consolidation phase vary according to risk group, with more intensive treatments used for patients classified as high or very-high risk. The continuation phase consists of less-intensive, low-dose chemotherapy. Nearly all regimens include weekly methotrexate and daily 6-mercaptopurine during this phase. Some protocols also use intermittent pulses of vincristine and corticosteroid in addition to methotrexate and 6-mercaptopurine.⁵⁴ Total duration of therapy is typically 2 to 3 years. On some regimens, boys receive a longer maintenance phase than girls.

Indications for Radiation Therapy

The role of radiation therapy (RT) in the treatment of childhood ALL has diminished in recent years. Improvements in the efficacy of chemotherapy administration and a clearer understanding of the prognostic factors have been responsible for this changing role. However, RT remains an important modality in ALL in specific situations. These include prophylactic cranial irradiation in select groups of high-risk patients, therapeutic CNS treatment for patients with meningeal leukemia, treatment of testicular disease, and total-body irradiation for some patients for whom stem cell transplant is recommended.

In most regimens, CNS prophylaxis for patients at lower risk is achieved with systemic and intrathecal chemotherapy without cranial irradiation. Children with high-risk features are at an increased risk of CNS relapse and, historically, have received prophylactic cranial irradiation. These features include a presenting WBC count of 50,000/µL or greater; those with WBC counts over 100,000/µL are at particularly high risk of CNS relapse. Additional high-risk features that are indications on some treatment protocols for cranial irradiation are T-cell phenotype, Ph chromosome–positive ALL, and the presence of t(4;11). Infants younger than age 12 months with 11q23 abnormalities are at very high risk of CNS relapse but because of their young age are usually treated without cranial irradiation, using intensified systemic and intrathecal chemotherapy to treat the CNS. Historically, the standard dose for high-risk ALL cranial prophylaxis had been 1800 cGy; however, published trials from the Berlin-Frankfurt-Munster group used 1200 cGy in patients with CNS-1 disease with very good results.¹² The standard dose for prophylactic cranial irradiation in those high-risk patients still treated with irradiation is now 1200 cGy.

Currently, fewer than 20% of children with ALL are treated with prophylactic cranial irradiation. Investigators are continuing to evaluate whether cranial irradiation can be eliminated in a greater proportion of patients, especially in very young children in whom the toxicities from irradiation are the most significant. Some protocols have omitted cranial irradiation entirely, using strategies that intensify the use of intrathecal chemotherapy and CNS-penetrant systemic chemotherapy, such as high-dose methotrexate.55,56 In one such study from St. Jude Children’s Research Hospital, the 5-year cumulative risk of isolated CNS relapse was 2.7% and that of any CNS relapse (isolated or combined with bone marrow) was 3.9%, although the risk of relapse was higher in some high-risk patient subsets.⁵⁶

Children who present with CNS-3 disease at diagnosis, regardless of the other features of their disease, are deemed at high risk and are considered to have meningeal leukemia. The treatment of these patients varies, and some protocols continue to employ craniospinal irradiation or cranial irradiation in addition to intrathecal chemotherapy. Irradiation is not administered until the patient is in remission, including clearance of the cerebrospinal fluid. After achievement of complete remission, 1800-cGy cranial irradiation has been administered early in the therapy.¹²

Isolated CNS relapse is rare with the current approach to ALL treatment. The treatment depends on the time from first remission as well as the extent of prior CNS therapy. Patients with an isolated CNS relapse occurring 18 months or more after initial diagnosis have an event-free survival of approximately 80% when treated with intensive systemic chemotherapy and cranial or craniospinal irradiation. A dose of 1800 cGy to the cranial field has been shown to be effective in such patients.⁵⁷