Rudolph's Pediatrics, 22nd Ed.

CHAPTER 449. Acute Lymphoblastic Leukemia

James B. Nachman

Acute lymphoblastic leukemia (ALL) is the most common childhood malignancy, accounting for almost 25% of cancer diagnoses seen in children younger than age 15 years. Patients with more than 25% lymphoid blasts in the bone marrow are considered to have leukemia regardless of the presence of extramedullary disease. Patients who have an extramedullary lymphoid mass and 5% to 25% lymphoid blasts in the marrow are considered to have lymphoblastic lymphoma with marrow involvement. There are approximately 2400 new cases of ALL diagnosed in children and adolescents less than 20 years of age each year in the United States.1 The peak incidence of ALL is in the 2- to 6-year-old age group. The incidence of ALL appears to be highest in Hispanic children and lowest in African American children.2

Progress made in treating ALL has been remarkable. As recently as 50 years ago, cure rates for patients with ALL were less than 10%. In 2008, we expected to cure approximately 80% to 85% of patients with ALL.3-6 Patients are generally considered cured if they remain in remission more than 5 years from discontinuation of therapy. Chemotherapeutic principles and supportive care guidelines derived from the experience in treating ALL have had a major impact on the treatment of other types of cancer.

BIOLOGY

In the developmental pathway of lymphocytes (see Fig. 449-1), maturation to the mature B-cell and T-cell state occurs without the necessity of antigenic stimulation—so-called antigen independent differentiation. When a B or T cell combines with its specific antigen, proliferation and differentiation occur, resulting in long-lived memory cells for antigen recognition and plasma cells which secrete antibody or T cells with various effector functions, such as cytotoxicity, B-cell help, cytokine secretion, and so on. In B-lineage acute lymphoblastic leukemia (ALL), approximately 98% of patients show a B-precursor immunophenotype. The rare patient with mature B-cell leukemia has a form of Burkitt lymphoma and requires different treatment. The various stages of B-cell development cannot be judged by cell morphology. Surface antigen expression is used to determine the stage of B-cell development. The earliest stage in B-cell development is the pro-B cell, which is characterized by surface expression of histocompatibility complex (HLA-DR) and the lack of expression of CD10 (common ALL antigen). The common precursor B cell expresses HLA-DR and CD10 but does not express cytoplasmic immunoglobulin. The pre-B cell expresses cytoplasmic immunoglobulin, and may or may not express CD10. The incidence of the various subtypes of B-precursor ALL is approximately 75% common B, 20% pre-B, and 5% pro-B.4 The pro-B immunophenotype is most often seen in infants. In T-ALL, cases are evenly distributed between the early, mid-, and late thymocyte state determined by the presence or absence of various T-cell antigens such as CD 2, 4, 5, 7, 8, and cytoplasmic CD3. Rarely, ALL cases may show a biphenotypic, bilineage, or natural killer cell immunophenotype.

FIGURE 449-1. Development states of B- and T-cell leukemia in childhood.

To determine cytogenetic abnormalities in childhood ALL, metaphase preparations of leukemic cells are prepared and abnormalities in chromosome number and structure are determined. It is now well known that certain genetic abnormalities are not detected by routine chromosome analysis. Therefore, florescent in situ hybridization (FISH) is also performed on leukemic cell spreads. Most cases of B-lineage ALL show chromosomal abnormalities detected by either standard cytogenetic or molecular techniques. These abnormalities may include gains or losses of whole chromosomes, structural abnormalities involving individual chromosomes, and translocations involving 2 or more chromosomes. The two most common genetic aberrations in childhood B-lineage ALL are hyperdiploidy (usually > 51 chromosomes) and a translocation between chromosomes 12 and 21, often referred to as TEL – AML1 because of the genes involved in the breakpoints on chromosomes 12 and 21. Both of these genetic abnormalities are most common in children 1 to 9 years of age, show a common B immunophenotype, and are associated with a good prognosis.7 Hypodiploidy (< 44 chromosomes) and a translocation between chromosomes 9 and 22, the so-called Philadelphia chromosome, are rare genetic abnormalities associated with a poor prognosis.8,9 The incidence of Philadelphia-chromosome-positive ALL increases with age and may account for 5% of cases in patients 16 to 21 years of age. Other genetic abnormalities are not associated with either favorable or unfavorable prognosis. Approximately 25% of cases of pre-B ALL show a translocation between chromosomes 1 and 19.10 In T-cell ALL, it is uncommon to find cytogenetic abnormalities based on examination of chromosome spreads. However, molecular analysis often reveals recurring abnormalities. For example, mutations in the Notch gene occurs in approximately 50% of cases of childhood T ALL and appear to be associated with a good prognosis.11

CLINICAL PRESENTATION

The demographic features of childhood acute lymphoblastic leukemia (ALL) are shown in Table 449-1.

Table 449-1. Demographics of Children and Adolescents with ALL

In many cases of childhood ALL, production of the various types of blood cells other than lymphocytes is seriously impaired. A low red cell blood count (RBC) count leads to pallor; decreased platelets lead to bruising, and a low neutrophil count may lead to fever as a sign of infection. The triad of fever, pallor, and bruising is a common presentation of ALL, particularly in younger children. In some younger patients, severe bone pain in the back and lower extremities with mild or no abnormalities in the blood count may be presenting symptoms of ALL. A number of patients with ALL present with lymphadenopathy and hepatospenomegaly, due to infiltration by malignant lymphocytes. Rarely, patients with ALL may present with respiratory compromise due to the presence of a large mediastinal mass (particularly more common in T-cell ALL). This is a medical emergency and requires immediate intervention to prevent respiratory arrest. Another rare presentation of ALL is that of a lytic bone lesion. Prior to performing a biopsy of any lytic bone lesion, a complete blood count should be obtained and a careful review of the peripheral smear should be carried out.

DIAGNOSIS

Most patients with acute lymphoblastic leukemia (ALL) will show at least some blast cells on a peripheral blood smear. ALL blasts are characterized by a high nuclear cytoplasmic ratio, clumpy nuclear chromatin, and few small nucleoli. Even if blast cells are present in the peripheral blood, performing a bone marrow aspiration is highly recommended to obtain specimens for determination of immunophenotype by flow cytometric techniques and cyto-genetic analysis. Both immunophenotype and cytogenetics are necessary for treatment stratification.7 Most cases of ALL are either of clear Bor T-lineage origin. Either B- or T-lineage ALL cases may show aberrant expression of myeloid antigens.12 If myeloid antigen(s) are present, a scoring system can be used to determine whether or not the leukemia has biphenotypic features, which is generally associated with a poor prognosis. This does not appear to be the case in acute myeloid leukemia, which expresses lymphoid antigens. Approximately 80% of cases of childhood ALL are B lineage, whereas 15% are T lineage. T-lineage cases account for approximately 10% of cases in children ages 1 to 9 years and approximately 20% of cases in adolescents and young adults. In any patient with ALL, a lumbar puncture should be performed to assess whether blasts are present in the cerebrospinal fluid. The initial puncture should be done by an experienced physician and younger patients should be under sedation to minimize the likelihood of a traumatic tap. More than 10 RBC/μL in cerebral spinal fluid (CSF) is an adverse prognostic factor. In a non-traumatic tap, the presence of 5 or more white blood cells and the presence of blasts in a cytocentrifuged specimen are needed to make a diagnosis of central nervous system involvement (CNS-3). Patients with less than 5 white blood cells in the cerebrospinal fluid but with leukemic blasts on the cytocentrifuge specimen are given the designation CNS-2. Patients with symptoms referable to the CNS should always have either a head computerized tomography (CT) scan or magnetic resonance imagine (MRI) prior to having a lumbar puncture.

PROGNOSTIC FACTORS

A large number of prognostic factors have been identified that are utilized to determine treatment. Prognostic factors can generally be divided into three groups: disease-related factors, host-related factors, and treatment response factors.

Age is an important prognostic factor in acute lymphoblastic leukemia (ALL).13 Generally, patients age 1 to 9 years have a better prognosis than those under 1 year of age or those older than 10 years. This is in part related to the significantly higher incidence of favorable cytogenetic subtypes of ALL (hyperdiploidy and the 12/21 translocation) in the 1-to-9-year age group. Between 50% and 60% of children in the 1-to-9-year age range will have either hyperdiploidy or translocation.12,21 Infants less than 1 year of age have a high incidence of abnormalities associated with the 11q23 chromosome locus.14 The most common abnormality is translocation.4,11 However, even if patients are matched for cytogenetic abnormalities, patients 1 to 9 years fare better than older patients.

White blood cell (WBC) count is another important prognostic factor in ALL. It is more important in B-lineage as opposed to T-lineage ALL. Many studies utilize an arbitrary cutoff of 50,000 WBC/μL at diagnosis in determining risk group assignment. Extreme hyperleukocytosis (> 200,000) is often associated with unfavorable cytogenetics and generally carries a poor prognosis.

The combination of age and white blood cell (WBC) count was one of the initial prognostic factors used to determine outcome. Patients 1 to 9 years of age and WBC less than 50,000/μL had a better prognosis than patients 1 to 9 years and more than 50,000 WBC or than patients more than 10 years of age at the time of diagnosis.

Race has been evaluated as a prognostic factor in many studies, but with modern treatment, race tends not to be an important prognostic factor.15 However, it is clear that racial background may play a role in determining clearance and toxicity associated with various chemotherapeutic agents.

The presence of central nervous system leukemia at the time of diagnosis (> 5 WBC/μL and the presence of blast cells in the cerebrospinal fluid, or presence of cranial nerve palsies) is a poor prognostic factor. Initial testicular involvement in males is also thought to be a high risk factor.16 Hepatosplenomegaly and other extramedullary involvement at diagnosis has no independent prognostic significance.

Pretreatment factors were initially used to determine risk group and treatment assignment. However, rapidity of response to induction chemotherapy is currently recognized as the most important factor determining outcome.17Within any subgroup defined by age, WBC, and/or cytogenetics, patients with no or minimal residual disease (MRD) at the end of induction have a significantly better outcome compared to patients with higher levels of MRD. Therefore, children with ALL are assigned to an induction treatment based on age, white blood cell count, and immunophenotype and then treatment is modified at the end of induction based on rapidity of response and cytogenetics (See Fig. 449-2 and Table 449-2). Initially, morphologic response in peripheral blood following a 7-day treatment prophase with prednisone alone was shown to have prognostic significance.17 Patients with less 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%. Patients with more than 25% blasts in the bone marrow on day 8 of induction therapy had a significantly worse prognosis compared to patients showing less than 25% blasts on the day-8 marrow examination. Patients showing slow early response were assigned to more aggressive treatment strategies and in some studies this led to an improvement in outcome.18

FIGURE 449-2. Current algorithm for treatment of acute lymphoblastic leukemia, 1 to 21 years of age. CNS, central nervous system; WBC, white blood cell.

Table 449-2. Definitions in Childhood Acute Lymphoblastic Leukemia

Morphologic assessment of response is imprecise and only allows detection of relatively high levels of disease. We consider the presence of greater than 5% blasts in bone marrow to represent residual leukemia. This is a sensitivity of 1 in 20 cells. Recently, two new techniques for measuring so-called minimal residual disease have been developed. Because leukemia is a clonal disorder, each cell in the clone carries a unique immunoglobulin gene rearrangement. This rearrangement can be sequenced and one can then determine the percentage of cells in peripheral blood or marrow that carries the unique rearrangement. This technique has a sensitivity of 1 in 10,000 cells. In most cases, the leukemic clone expresses an aberrant immunophenotype, which can be detected by flow cytometric techniques.

Cytometric determination of minimal residual disease is between 1 in 1000 and 1 in 10,000 cells. Patients with high levels of minimal residual disease at the end of induction therapy have a worse outcome than those patients with lower or undetectable levels.19,20 Gene profiling is currently being evaluated as a prognostic factor in childhood ALL.

INITIAL MANAGEMENT

It is important to recognize that, with rare exception, immediate treatment with chemotherapy is not necessary and may be contraindicated. Patients often present with metabolic abnormalities, hematologic abnormalities, or infectious complications which must be managed before treatment can be commenced in a safe manner.

Initial assessment of the newly diagnosed patient should begin with a careful determination of the vital signs. A patient with in acute lymphoblastic leukemia (ALL) and a large mediastinal mass may present with respiratory distress due to tracheal compression; if untreated, this can progress to respiratory arrest. The treatment of choice is low-dose radiation to a small field around the airway. Following the initiation of radiation, significant improvement in respiratory status is usually seen within 12 hours. If radiation therapy is not available, steroid therapy will generally shrink the mass but may put the patient at risk for tumor lysis syndrome (see below).

Many patients with ALL will present with fever. If the fever is associated with neutropenia (absolute neutrophil count < 500/μL), there is a high risk for bacterial sepsis. In patients with fever and neutropenia, broad-spectrum antibiotic coverage should be initiated. A third-generation cephalosporin, such as ceftazidime or cefipime, is often used as initial therapy. If hypotension is present, an aminoglycoside is often added to the antibiotic regimen. Cultures of blood and urine and cultures of additional sites as clinically indicated should be obtained prior to beginning antibiotic therapy.

Tumor lysis syndrome refers to the release of waste products from leukemic cells.21 The two most important products are uric acid and phosphorus. Patients with ALL and high white blood cell (WBC) counts or large extramedullary masses often present with an elevated blood uric acid level. In rare cases, the patients may present with significant renal compromise due to uric acid nephropathy. In any patient with ALL and initial high uric acid level, renal function should be carefully assessed. If there is no renal compromise, measures may be instituted to lower the uric acid level. In patients with renal compromise who do not show rapid improvement in renal function with uric-acid-lowering measures, dialysis may be required until the uric acid nephropathy resolves.

Even if uric acid levels are initially normal, they may rise as therapy is initiated. Thus, all patients should receive therapy to prevent the development of uric acid nephropathy. In patients in whom the tumor burden is low and uric acid concentration is not elevated at the time of diagnosis, the combination of allopurinol therapy, which blocks uric acid production, and hyperhydration is often used. Use of sodium bicarbonate is no longer recommended. For patients with high tumor burdens and/or elevated uric acid levels, urate oxidase is recommended. Urate oxidase degrades uric acid to harmless metabolites. Blood levels of uric acid fall precipitously following a single dose of urate oxidase. Serum phosphate levels, although rarely significantly elevated at diagnosis, may rise as treatment is initiated and lead to hypocalcemia and the precipitation of calcium phosphate in the renal tubules. Measures to reduce phosphate are not as effective as those designed to reduce uric acid.

Patients with ALL are often anemic at diagnosis. Anemia in ALL is a chronic process and is generally associated with an increased blood volume. Initially, small, packed red cell transfusions can 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 the risk for thrombosis.

Extreme leukocytosis (WBC > 200,000/μL) is usually asymptomatic but may be associated with central nervous system (CNS) symptoms and/or thrombosis. Leukopharesis is often carried out to reduce the WBC count but whether this has any benefit is unproven.

GENERAL TREATMENT STRATEGY

Three general treatment principles for treating cancer were derived from the early treatment experience with childhood acute lymphoblastic leukemia (ALL). When patients with ALL were given prednisone, approximately 25% showed no evidence of disease after 4 weeks. However, when prednisone was discontinued, leukemia recurred in all patients. It became apparent that clinical resolution of disease was not synonymous with cure. The term remissionwas coined to define a state in which there was no clinical or laboratory evidence of residual disease. Postremission treatment is necessary for ALL if remission is to be converted to cure.

Vincristine given weekly for 4 weeks induces a 25% remission rate. However, when Vincristine and prednisone are given together, approximately 75% of patients attain a remission; this provided one, if not the first, example of the efficacy of combination chemotherapy for cancer treatment.

When patients with ALL were given combination induction and postremission chemotherapy for 6 months, many patients developed severe headache, nausea and emesis, and papilledema. Examination of the cerebral spinal fluid revealed a large number of leukemic blast cells, whereas examination of the bone marrow often showed no residual leukemia. Thus, it became apparent that the central nervous system was a “sanctuary” site and that chemotherapy administered orally, intravenously, or intramuscularly did not produce adequate drug levels in cerebrospinal fluid. Administration of chemotherapy directly into the cerebrospinal fluid in conjunction with the cranial radiotherapy was employed as pre-symptomatic central nervous system therapy.

These principles led to the first treatment regimen which produced cures for ALL. The regimen consisted of a 3-drug induction phase with concomitant intrathecal chemotherapy, CNS intensification, and maintenance (see Table 449-3). The third drug, L-asparaginase, depletes the serum of asparagines, which seems to be an essential amino acid for ALL cells but not for normal cells. The induction phase included no myelosuppressive drugs. Maintenance therapy consisted of vincristine once per month, prednisone 5 days per month, daily oral 6-mercaptopurine, and weekly oral methotrexate.

Because a large majority of the patients relapsed during the maintenance phase, attempts were made to intensify maintenance therapy but these interventions produced no improvement in outcome. Goldie and Coldman observed that many patients who relapsed were able to achieve a second remission but that the second remission was shorter than the first, and eventually it became impossible to achieve remission. This implied that at diagnosis, the majority of leukemic cells were drug sensitive. However, in patients who failed therapy, end-stage disease was refractory to all chemotherapy. They postulated that drug resistance arose as a result of mutational events in leukemic cells.22 If the mutational rate for a given cell type is constant, tumor burden should be an indicator of the number of mutational events occurring prior to diagnosis. They predicted that patients with high initial tumor burdens would be more likely to harbor small numbers of resistant subclones and thus be more likely to fail therapy. This fit with clinical data showing that patients with high initial white blood cell counts or bulky extramedullary disease had a very high rate of treatment failure. They suggested that intensifying maintenance therapy failed because it was applied too late and that resistant cell clones had already become established. They argued that early treatment intensification might result in the eradication of the resistant subclones and increase the cure rate.

Norton and Simon stated that after initial intensive therapy, any residual disease was probably relatively drug resistant. They suggested that a delayed intensification incorporating new drugs might improve cure rate.23

These two theories, arguing for intensification of early treatment and a delayed intensification, were put into practice by Dr Riehm in Germany, who developed a treatment program referred to as BFM (Berlin, Frankfort, Munster) in the late 1970s (see Table 449-3).24

The development of the BFM regimen was a major advance in the treatment of childhood ALL. This regimen featured the addition of daunomycin to the prior 3-drug induction; the addition of cytosine arabinoside and cyclophosphamide to the CNS-intensification phase; and the incorporation of a delayed reinduction/reintensification phase in which dexamethasone replaced prednisone, doxorubicin replaced daunomycin, and 6-thioguanine replaced 6-mercaptopurine. Initially, the two intensive phases of therapy were separated by a period of interim maintenance consisting of daily oral 6-mercaptopurine and weekly oral methotrexate. In the second and subsequent iterations of the protocol, 4 courses of high-dose methotrexate with leucovorin rescue at 2-week intervals replaced the weekly oral methotrexate. This regimen led to an approximate 60% cure rate for children with ALL.

Table 449-3. Comparison of Standard Therapy, BFM Therapy, and Augmented BFM Therapy for Acute Lymphoblastic Leukemia

In the United States, the Children’s Cancer Group (CCG) showed that the BFM regimen was superior to standard therapy for patients with white blood cell (WBC) counts greater than 50,000/μL and for those with bulky extramedullary disease.25 BFM therapy has two major components—intensive induction consolidation and reinduction/reintensification. The CCG studied the relative contributions of each of these intensive phases of therapy in a group of patients with white blood counts less than 50,000/μL. The CCG also studied whether intrathecal therapy alone could provide adequate CNS prophylaxis versus the combination of intrathecal therapy plus cranial radiation.26For patients less than 10 years of age, only the reinduction/reintensification phase was important for improving outcome and intrathecal therapy alone provided adequate CNS protection. However, for patients over 10 years of age, only the full BFM protocol and cranial radiation produced a good outcome. These observations led to the formulation of the National Cancer Institute (NCI) criteria for initial stratification of ALL. Patients ages 1 to 9 years and WBC less than 50,000/μL were considered standard risk and all other patients were considered high risk.

The Children’s Cancer Group utilized a day-7 bone marrow evaluation following 7 days of 4-drug induction to measure early response. Patients with more than 25% blasts on day 7 were considered slow responders and had an event-free survival of less than 50%. CCG investigators took note of the failure to improve outcome for slow responders in German studies utilizing intensive pulses of myelosuppressive therapy. Therefore, they developed a regimen in which the dose intensity of nonmyelosuppressive drugs (vincristine, L-Asparaginase) was increased and intravenous methotrexate without leucovorin rescue was utilized. To further intensify therapy, a second interim maintenance and delayed intensification phase was added prior to maintenance. This regimen was referred to as “Augmented” Berlin, Frankfort, Munster program (ABFM) and is shown in Table 449-3.18 Compared to standard BFM, ABFM featured twice as much vincristine and three times as much asparaginase during the first year of therapy. In a head-to-head trial of standard BFM and ABFM, ABFM produced a significant improvement in outcome for NCI high-risk patients with greater than 25% blasts in the bone marrow on day 7. In a recently completed CCG trial, intensification of early therapy with nonmyelosuppressive drugs and the incorporation of intravenous MTX without rescue improved outcome for NCI high-risk rapid-responder patients.27 However, the addition of a second interim maintenance and delayed intensification phase produced no benefit in this patient population. Thus, the current standard therapy for NCI high risk patients and all T-cell patients with a rapid morphologic response is Augmented BFM with only one interim maintenance and delayed intensification phase while full ABFM is the standard treatment for high-risk slow-responder patients.

For NCI standard risk patients, the CCG showed that utilizing dexamethasone during all phases of therapy produced a significant improvement in outcome compared to prednisone during all phases except delayed intensification.28The CCG also demonstrated that patients with greater than 5% blasts in a day-14 marrow specimen had a significantly better outcome when therapy was switched to ABFM.

Infants with ALL pose a unique challenge. The majority of the patients less than 6 months of age show myeloid/lymphoid leukemia (MLL) gene rearrangements, most commonly a translocation between chromosomes 4 and 11 and have a poor outcome. Infants without MLL gene rearrangements have a significantly better outcome.

The current paradigm for treating patients ages 1 to 21 years with B-precursor ALL begins by assigning patients to either a standard or a high-risk group based on age and WBC. Patients ages 1 to 9 years who have less than 50 K WBC are assigned to a standard risk group and receive a 3-drug induction that includes dexamethasone as the steroid. All other patients are considered high risk and receive a 4-drug induction. High-risk patients ages 1 to 9 years receive either prednisone or dexamethasone, whereas older patients receive prednisone.

Patients are then reclassified at the end of induction based on the initial risk group, cytogenetics, and early response measured by both marrow morphology and minimal residual disease determination by flow cytometry. Patients with the Philadelphia chromosome or extreme hypodiploidy are considered very high risk regardless of initial risk group or early response. These patients receive very intensive chemotherapy and are eligible for bone marrow transplant in first remission. Patients with Philadelphia-chromosome-positive ALL receive Gleevac, a tyrosine kinase inhibitor, which inhibits the BCR-ABL protein, in conjunction with aggressive chemotherapy. Patients with more than 5% marrow blasts on day 14 and/or more than 0.1% blasts by flow cytometry on day 28 (slow responders) are considered high risk and receive intensive treatment. Rapid-responder patients are assigned to treatment regimens based on initial risk group and cytogenetics. Standard risk rapid responders with either the TEL-AML1 translocation or hyperdiploidy with extra copies of chromosomes 4 and 10 have a projected event-free survival of greater than 90%.

RECURRENT DISEASE

In 2008, long-term event-free survival for children and young adults with acute lymphoblastic leukemia (ALL) approaches 80%. Approximately 2% to 3% will die of infectious or other toxic complications and 1% develop a second malignant neoplasm. Therefore about 15% of patients will relapse. Relapse can occur in the bone marrow, extramedullary sites, with central nervous system (CNS) and testes being most common, or in both locations. 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 does late relapse. The general strategy for treating patients with early isolated or combined bone marrow and concurrent extramedullary relapse is to reinduce the patient and then perform an allogeneic bone marrow transplant utilizing either a matched sibling donor or a matched unrelated donor. More recently, cord blood has been utilized as a source of stem cells. Late isolated extramedullary relapse carries a favorable prognosis and is treated with chemotherapy and local therapy directed toward the extramedullary site. Currently, there is controversy regarding optimal therapy for early extramedullary relapse and late bone marrow relapse with or without a concurrent extramedullary relapse. Some physicians favor an allogeneic stem cell transplant and others favor chemotherapy alone. In the case of late marrow relapse the use of minimal residual disease (MRD) determination at the end of reinduction may aid in determining optimal therapy.

COMPLICATIONS AND LATE EFFECTS

Complications occurring during induction have already been discussed above. During postinduction treatment, patients with acute lymphoblastic leukemia (ALL) tend to have extremely low lymphocyte counts, particularly early in maintenance, and are at risk for infection with Pneumocystis carinii, which can be fatal. Thus, all patients with ALL receive prophylactic trimethoprim-sulfamethoxazole therapy 2 to 3 days per week. Vincristine can cause peripheral neuropathy, which may be severe. If vincristine neurotoxicity is noted, the dose of vincristine is reduced or withheld and symptoms generally improve. Steroids, particularly dexamethasone, can cause significant mood swings in children. Some patients exhibit aggressive behavior during steroid administration. Steroids may also be associated with the development of osteonecrosis,29 particularly in older patients. Weight-bearing joints are usually involved and symptoms may be severe. In certain cases, hip-replacement surgery has been necessary. Discontinuous use of dexamethasone during delayed intensification decreases the incidence of osteonecrosis. Approximately 5% of patients with ALL may experience a seizure associated with the intrathecal administration of methotrexate. Arachnoiditis may also occur.

Thus far, late adverse effects of ALL therapy have been relatively infrequent. A major concern is cognitive impairment following intrathecal chemotherapy with or without cranial radiation. In 2008, approximately 10% to 15 % of ALL patients received cranial radiation and the dose had been decreased from 24 Gy to 12 Gy. Recent studies suggest that cognitive impairment following intrathecal chemotherapy is relatively mild or nonexistent.30,31 Obesity and growth impairment are infrequent in the chemotherapy-alone era. Approximately 1% of patients treated with chemotherapy alone may develop a second malignancy.32 Almost all patients with ALL maintain fertility. It must be pointed out that there are few long-term survivors more than 40 years of age, so the issue of very late toxicity remains to be resolved.

FUTURE DIRECTIONS

Identification of subgroups of acute lymphoblastic leukemia (ALL) patients who can be cured with minimal therapy is one priority. Currently, National Cancer Institute standard-risk patients with favorable cytogenetics and undetectable early minimal residual disease have a projected 5-year event free survival of 98%. In the genomic age, identifying patients who do not metabolize certain drugs effectively or rapidly clear other drugs may lead to individualizing treatment for each patient. Identifying patients who will benefit from hematopoietic stem cell transplantion in first remission is also a priority.