Chapter 91: Acute Lymphoblastic Leukemia

Acute lymphoblastic leukemia (ALL) is a neoplastic disease that results from multistep somatic mutations in a single lymphoid progenitor cell at one of several discrete stages of development. The immunophenotype of leukemic cells at diagnosis reflects the level of differentiation achieved by the dominant clone. The clonal origin of ALL has been established by cytogenetic analysis, by analysis of restriction fragments in female patients who are heterozygous for polymorphic X chromosome-linked genes, and by analysis of rearrangements of T-cell receptor or immunoglobulin genes. Leukemic cells divide more slowly and require more time to synthesize DNA than do normal hematopoietic counterparts. However, leukemic cells accumulate relentlessly because of their altered response to growth and death signals.^1,2 They compete successfully with normal hematopoietic cells, resulting in anemia, thrombocytopenia, and neutropenia. At diagnosis, leukemic cells not only have replaced normal marrow cells but also have disseminated to various extramedullary sites.

Velpeau³ is generally credited with the earliest report of leukemia in 1827. Virchow,⁴ Bennett,⁵ and Craigie⁶ recognized the condition as a distinct entity by 1845. In 1847, Virchow coined the term “weisses blut” and, later, “leucaemie,” applying it to two distinct types of the disease—splenic and lymphatic—that could be distinguished from each other based on splenomegaly and enlarged lymph nodes and on the morphologic similarities of the leukemic cells to those normally found in these organs.⁷ Ehrlich’s introduction of staining methods in 1891 allowed further distinction of leukemia subtypes.⁸ By 1913, leukemia could be classified as acute or chronic, and as lymphatic or myelogenous.⁹ The greater prevalence of ALL in children, especially those ages 1 to 5 years, was recognized in 1917.¹⁰

Shortly after leukemia was recognized as a discrete disease entity, physicians began using chemicals as palliative therapy. The first advance was the use of a 4-amino analogue of folic acid (aminopterin), prompted by Farber’s observation that folic acid appeared to accelerate the proliferation of leukemic cells. Strikingly, for the first time, complete clinical and hematologic remissions that lasted for several months were seen in children.¹¹ A year after the report of aminopterin-induced clinical remissions, a newly isolated adrenocorticotrophic hormone was reported to induce prompt, though brief, remissions in patients with leukemia.¹² Almost concurrently, Elion and colleagues synthesized antimetabolites that interfere with synthesis of purines and pyrimidines.¹³ Their findings led to the introduction of mercaptopurine, 6-thioguanine, and allopurinol into clinical use. From 1950 to 1960, many new antileukemic agents were introduced, and occasional cures were seen. Pinkel and colleagues at St. Jude Children’s Research Hospital, in 1962, devised a “total therapy” approach, consisting of four treatment phases: remission induction; intensification or consolidation; therapy for subclinical CNS leukemia (or preventive meningeal treatment); and prolonged continuation therapy.¹⁴ By the early 1970s, as many as 50 percent of children achieved long-term event-free survival (EFS) using this innovative strategy. During the same period, a better understanding of the genetics of human histocompatibility and wider use of human leukocyte antigen (HLA) typing culminated in the successful use of hematopoietic stem cell transplantation for treatment of patients in whom leukemia relapsed. In the early 1980s, Riehm and coworkers introduced a so-called reinduction or delayed intensification treatment during early continuation therapy, consisting mainly of repetition of the initial remission induction and early intensification phases, and further improved the EFS to approximately 70 percent.¹⁵ Parallel to advances in treatment has been the improved understanding of the biology of ALL. The recognition of ALL as a heterogeneous group of diseases—clinically, immunologically, and genetically¹⁶—set the stage for risk-directed therapy.^16a

Treatment of ALL has progressed incrementally, beginning with the development of effective therapy for CNS disease, followed by intensification of early treatment, especially for patients at high risk of relapse. The current cure rates of nearly 90 percent for children (Fig. 91–1) and 40 percent for adults attest to the steady progress made in treating this disease.^17,18 Rapid evolution and convergence of multiple genome-wide platforms to identify the total complement of genetic and epigenetic alterations almost certainly will lead to the identification of new targets for specific treatment.^19,20 A clear advance was the development of imatinib mesylate and dasatinib, which target leukemias with the BCR-ABL1 fusion.²¹

Initiation and progression of ALL are driven by successive mutations that alter cellular functions, including an enhanced ability of self-renewal, a subversion of control of normal proliferation, a block in differentiation, and an increased resistance to death signals (apoptosis).^1,2 Familial disorders of DNA repair may play a role. Environmental agents, such as ionizing radiation and chemical mutagens, have been implicated in the induction of ALL in some patients. However, in most cases, no etiologic factors are discernible. In the favored theory, leukemogenesis reflects the interaction between host pharmacogenetics (susceptibility) and environmental factors, a model that requires confirmation in well-designed population and molecular epidemiologic studies.

INCIDENCE

The American Cancer Society has estimated that in the United States there will be approximately 6020 new cases of ALL in 2014 (3140 in males and 2880 in females) and approximately 1440 deaths from ALL (810 in males and 630 in females).²² Most cases of ALL occur in children, but most deaths from ALL (approximately four of five) occur in adults.

The age-adjusted incidence rate of ALL was 1.6 per 100,000 males and 1.2 for females per year in the United States, based on cases diagnosed in 1975 to 2010 from 17 Surveillance, Epidemiology, and End Results (SEER) geographic areas.²³ The risk for developing ALL is highest in children younger than 5 years of age. The risk then declines slowly until the mid-20s, and begins to rise again slowly after age 50. The incidence is 7.9 per 100,000 children 1 to 4 years old and 1.2 for those older than age 60 years. Only 20 percent of adult acute leukemias are ALL, but about one-third of ALL cases are in adults. The average person’s lifetime risk of developing ALL is less than one in 750. The risk is slightly higher in males than in females, and higher in whites than in African Americans (Fig. 91–2).²³ The median age at diagnosis for ALL is 13 years and approximately 61 percent are diagnosed before the age of 20 years, however, because of the bimodal peak in incidence, the age of 13 years is mathematically correct but medically nearly useless. ALL is the most common malignancy diagnosed in patients younger than age 15 years, accounting for 23 percent of all cancers and 76 percent of all leukemias in this age group.

The sharp incidence peak of ALL during early childhood has been observed only since the 1930s in the United Kingdom and the United States.²⁴ In the United States, the peak first appeared in children of European descent, and subsequently was seen in children of African descent in the 1960s. The age peak is absent in many developing or underdeveloped countries, suggesting a leukemogenic contribution from factors associated with industrialization. Except for a slight predominance for females in infancy, ALL affects males of European descent more often than females in all age groups (see Fig. 91–2). The frequency distribution is similar among those of African descent. In most age groups, the incidence of ALL is higher in those of European descent than in those of African descent, especially among children ages 2 to 3 years.

The incidence of ALL differs substantially in different geographic areas. Rates are higher among populations in northern and western Europe, North America, and Oceania, with lower rates in Asian and African populations.²⁵ In Europe, the highest rates of ALL among males are found in Spain and the highest rates among females in Denmark. In the United States, the highest rates for both sexes are among Latinos in Los Angeles.

RISK FACTORS

Genetic Syndromes

The precise pathogenetic events leading to the development of ALL are unknown. Only a minority (5 percent) of cases are associated with inherited, predisposing genetic syndromes. Children with Down syndrome have a 10 to 30 times greater risk of leukemia; acute megakaryoblastic leukemia predominates in those patients younger than age 3 years, and ALL is predominant in older age groups. ALL in patients with Down syndrome is a heterogeneous disorder, comprising subtypes with the same well-recognized genetic abnormalities found in the general population, such as hyperdiploidy greater than 50 and t(12;21)[ETV6-RUNX1], plus those more commonly associated with Down syndrome such as +X, del(9), and CEBPD rearrangement.^26,27 Studies show that P2RY8-CRLF2 fusion and activating JAK mutations together contribute to leukemogenesis in approximately half of the cases of Down syndrome patients with ALL.^28,29 Almost all ALL patients with Down syndrome have a deletion of IKZF1.³⁰ Autosomal recessive genetic diseases associated with increased chromosomal fragility and a predisposition to ALL include ataxia-telangiectasia, Nijmegen breakage syndrome, and Bloom syndrome.³¹ Patients with ataxia-telangiectasia have a 70 times greater risk of leukemia and a 250 times greater risk of lymphoma, particularly of the T-cell phenotype.³² The causative gene, termed ATM (ataxia-telangiectasia mutated), encodes a protein involved in DNA repair, regulation of cell proliferation, and apoptosis. Laboratory studies supporting the diagnosis of ataxia-telangiectasia include an elevated serum concentration of α-fetoprotein, presence of characteristic chromosomal aberrations, absent or reduced intranuclear serine protein kinase ATM, and increased in vitro radiosensitivity. A high prevalence of germline truncating and missense ATM gene alterations in children with sporadic T-cell ALL suggests a pathogenetic role of ATM in lymphoid malignancies. Although impaired immune surveillance contributes to the increased risk of Epstein-Barr virus-related malignancies in patients with acquired immunodeficiencies, no compelling evidence indicates defective immunity contributes to the predisposition to ALL in patients with ataxia-telangiectasia or other congenital immunodeficiency syndromes. Genome-wide association studies have identified common allelic variants in four genes (IKZF1, ARID5B, CEBPE, and CDKN2a) that are consistently associated with childhood ALL.^33,34,35 These genes are key regulators of blood cell development, and acquired mutations of each are also detected in ALL cases. Thus, the risk of childhood ALL may be influenced by coinheritance of multiple low-risk variants. Inherited allelic variation may also affect response to treatment.³⁶

Environmental Factors

In utero (but not postnatal) exposure to diagnostic x-rays confers a slightly increased risk of ALL, which correlates positively with the number of exposures.³⁷ The evidence is weak for an association between the development of ALL and nuclear fallout; exposure to occupational, natural terrestrial, or cosmic ionizing radiation exposure; or paternal radiation exposure prior to conception. There has been concern that exposure to low-energy electromagnetic fields produced by a residential power supply may be associated with the development of childhood ALL. Case-control studies suggested a slightly increased risk of leukemia at very high levels of exposure; assuming the association is real, only approximately 1 percent of leukemias could be attributed to the exposure.^38,39 Pesticide exposure (occupational or home use) and parental cigarette smoking before or during pregnancy, administration of vitamin K to neonates, maternal alcohol consumption during pregnancy, and increased consumption of dietary nitrites have each been suggested causes. However, each of these associations is controversial, and most have been refuted after careful, controlled investigation. High birth weight is associated with an increased risk of leukemia before the age of 5 years with fair consistency,⁴⁰ and the birth weight is likely a marker for an endogenous factor, such as insulin-like growth factor.

Host Pharmacogenetics

Subtle genetic polymorphisms of xenobiotic-metabolizing enzymes, DNA repair pathways, and cell-cycle checkpoint functions might interact with environmental, dietary, maternal, and other external factors to affect the development of ALL.^2,41 Although the number of investigations and sample sizes are limited, data exist to support a causal role for polymorphisms in genes encoding detoxifying enzymes (e.g., glutathione S-transferase, nicotinamide adenine dinucleotide phosphate [NAD(P)H]:quinone oxidoreductase), folate-metabolizing enzymes (serine hydroxymethyltransferase and thymidylate synthase), cytochrome P450, methylenetetrahydrofolate reductase, and cell-cycle inhibitors in the development of adult and childhood ALL.^42,43 However, all these associations must be confirmed by larger studies with careful attention to ethnic and geographic diversity in the frequency of polymorphisms. Using genome-wide analysis, germline single-nucleotide polymorphisms (SNPs) of AT-rich interactive domain 5b (ARID5B) gene have been associated with childhood hyperdiploid B-cell precursor ALL,⁴⁴ a clear example of host genetic variations affecting the susceptibility to the development of childhood ALL.

Development of Acute Lymphoblastic Leukemia in Utero

Retrospective identification of leukemia-specific fusion genes (e.g., KMT2A/AFF1 [also known as MLL-AF4] and ETV6-RUNX1 [also known as TEL-AML1]), hyperdiploidy, or clonotypic rearrangements of immunoglobulin or T-cell receptor loci in archived neonatal blood spots (Guthrie cards), and development of concordant leukemia in identical twins clearly indicate some leukemias have a prenatal origin.^45,46 In identical twins with the t(4;11)/KMT2A/AFF1, the concordance rate is nearly 100 percent, and the latency in the time of occurrence in the two twins is short (a few weeks to a few months). These findings suggest this fusion gene alone either is leukemogenic or requires only a small number of cooperative mutations to cause leukemia. By contrast, the lower concordance rate in twins with the ETV6-RUNX1 fusion or T-cell phenotype and the longer postnatal latency period suggest additional postnatal events are required for leukemic transformation in these subtypes.⁴⁵ This theory is supported by the identification of rare cells expressing ETV6-RUNX1 fusion transcripts in approximately 1 percent of cord blood samples from newborns, a frequency 100 times higher than the incidence of ALL defined by this fusion transcript.⁴⁵ The presence of a preleukemic clone with the ETV6-RUNX1 has been established.⁴⁷ Hyperdiploid ALL, another common subtype of childhood ALL, also appears to arise before birth but requires postnatal events for full malignant transformation.⁴⁶ The observations of a peak age of development of childhood ALL of 2 to 5 years, an association of industrialization and modern or affluent societies with increased prevalence of ALL, and the occasional clustering of childhood leukemia cases have fueled two parallel infection-based hypotheses to account for postnatal events. The “delayed-infection” hypothesis suggests that some susceptible individuals with a prenatally acquired preleukemic clone had low or no exposure to common infections early in life because they lived in an affluent hygienic environments.⁴⁵ Such infectious insulation predisposes the immune system of these individuals to aberrant or pathologic responses after subsequent or delayed exposure to common infections at an age commensurate with increased lymphoid cell proliferation. The “population-mixing” hypothesis predicts that clusters of childhood ALL result from exposure of susceptible (nonimmune) individuals to common but fairly nonpathologic infections after population mixing with carriers.⁴⁸ However, clearly not all childhood cases develop in utero. For example, t(1;19)/TCF3-PBX1 (also known as E2A-PBX1) ALL appears to have a postnatal origin in most cases.⁴⁹ Cases of adult ALL most certainly arise over a protracted time.

ACQUIRED GENETIC CHANGES

Acquired genetic abnormalities are a hallmark of ALL; 80 percent of all cases have recurring cytogenetic or molecular lesions with prognostic and therapeutic relevance (Table 91–1).^2,19,41 Chromosomal changes include abnormalities in the number (ploidy) and structure of chromosomes.^50,51,52 The latter comprise translocations (the most frequent abnormality), inversions, deletions, point mutations, and amplifications. Although the frequency of particular genetic subtypes differs between childhood and adult cases, the general mechanisms underlying the induction are similar. Mechanisms include aberrant expression of oncoproteins, loss of tumor-suppressor genes, and chromosomal translocations that generate fusion genes encoding transcription factors or active kinases.

Primary genetic rearrangement by itself is insufficient to induce overt leukemia. Cooperative mutations are necessary for leukemic transformation and include genetic and epigenetic changes in key growth regulatory pathways.^19,20 The candidate gene approach has identified deletion of the CDKN2A/CDKN2B tumor-suppressor locus⁵³ and mutations of NOTCH1 in T-cell ALL.⁵⁴ Current searches applying genome-wide microarray and high-throughput sequencing methodologies have identified a high frequency of common genetic alterations in both B-cell precursor ALL and T-cell ALL. Using SNP microarray, a mean of 6.46 DNA copy number abnormalities (CNAs) per case was identified, suggesting that gross genomic instability is not a feature for most ALL cases.⁵⁵ There was a wide variation in the number of CNAs across leukemic subtypes. Interestingly, infant ALL cases with MLL rearrangement had less than one CNA per case, suggesting that few additional genetic lesions are required for leukemogenesis in these cases. By contrast, ETV6-RUNX1 and BCR-ABL1 cases had more than six CNAs per case, with some having more than 20 lesions, a finding consistent with the concept that, although the initiating events may occur early in childhood, additional lesions are required for subsequent development of ALL. More than 40 percent of B-cell precursor ALL cases had mutations in genes encoding regulators of normal lymphoid development.⁵⁵ The most frequent target was the lymphoid transcription factor PAX5 (mutated in approximately 32 percent of cases), which encode a paired-domain protein required for the pro–B-cell to pre–B-cell transition and B-lineage fidelity. The second most frequently involved gene was IKZF1 (mutated in almost 28 percent of the cases), encoding the IKAROS zinc finger DNA-binding protein that is required for the earliest lymphoid differentiation. IKZF1 was deleted in the vast majority of cases of BCR-ABL1 ALL cases as well as chronic myeloid leukemia in lymphoid blast crisis (but not chronic phase).⁵⁶ Approximately half of BCR-ABL1 ALL cases also had deletions of CDKN2A/B and PAX5. This finding further supports the concept that multiple signaling pathways need to be disrupted to induce leukemia. A subgroup of ALL with very poor outcome was strongly associated with the presence of IKZF1 deletions.^57,58 Together, these findings suggest that IKZF1 directly contributes to treatment resistance in ALL.

BCR-ABL1–like B-cell ALL lacks the BCR-ABL1 fusion or t(9;22) by cytogenetic, fluorescence in situ hybridization (FISH), or molecular analyses, but it shares the same gene-expression profile with typical BCR-ABL1–positive ALL.⁵⁹ In half of these cases, the CRLF2 gene is involved in a cryptic translocation with the IGH gene or is fused to the P2RY8 gene; both rearrangements lead to overexpression of CRLF2.^29,60 Mutations in JAK2 or JAK1 are detected in 30 to 40 percent of these cases, and many of the remaining have activating mutations in cytokine receptor and kinase signaling pathways.³⁰ Microarrays and genomic DNA sequencing identified monoallelic deletion of the PAX5 gene at chromosome band 9p13.2 in 28 percent of ALL patients with cryptic or larger deletions on 9p.⁵⁵

Gene-expression profiling with DNA microarrays allows nearly all T-cell cases to be grouped according to multistep oncogenic pathways.⁶¹ Gene-expression studies also show that overexpression of FLT3, a receptor tyrosine kinase important for development of hematopoietic stem cells, is a secondary event in almost all cases with either MLL rearrangements or hyperdiploidy.⁶² The finding has provided an impetus for clinical testing of FLT3 inhibitors in ALL. Other genome-wide interrogations of both leukemia cells and germline tissues have identified other genetic variations with prognostic or therapeutic relevance and may lead to the development of specific treatment.^17,36

Epigenetic changes, including hypermethylation and silencing of tumor-suppressor genes and hypomethylation of oncogenes and abnormalities in posttranscriptional control mechanisms, such as those involving microRNA, are common findings in cancer. These changes are reversible and do not alter the DNA sequence, yet they can alter gene expression in subtle ways that encourage malignant transformation and progression. The analysis of epigenetic alterations has begun to apply to the development of new biomarkers for risk assignment or disease monitoring, and to the design of alternative treatment in ALL.⁶³ Evidence indicates that the methylation of multiple genes in ALL is associated with a worse outcome. Surprisingly, methylation of genes was as prominent in childhood as in adult ALL. The differences in the response of children and adults appear not to be related to quantitative methylation but to the specific genes and the specific pathways deactivated. Preliminary studies of hypomethylating agents (e.g., azacitidine and decitabine) are being tested in patients refractory or resistant to current drug programs.⁶⁴

SIGNS AND SYMPTOMS

The clinical presentation of ALL is highly variable. Symptoms may appear insidiously or acutely. The presenting features generally reflect the degree of marrow failure and the extent of extramedullary spread (Table 91–2).^18,65,66,67 Approximately half of patients present with fever, which can be caused by either neutropenia-induced infection or leukemia-released cytokines (e.g., interleukin-1, interleukin-6, and tumor necrosis factor) released from leukemia cells. In these patients, fever resolves within 72 hours after the start of antileukemia therapy.

Fatigue and lethargy are common manifestations of anemia in patients with ALL. In older patients, anemia-related dyspnea and lightheadedness may be the dominant presenting features. More than 25 percent of patients, especially young children, may have a limp from bone pain or arthralgia, or an unwillingness to walk because of leukemic infiltration of the periosteum, bone, or joint or because of expansion of the marrow cavity by leukemia cells. Children with prominent bone pain often have nearly normal blood counts, which can contribute to delayed diagnosis. In a small proportion of patients, marrow necrosis can result in severe bone pain and tenderness, fever, and a very high level of serum lactate dehydrogenase.^68,69 Arthralgia and bone pain are less severe in adults. Less common signs and symptoms include headache, vomiting, altered mental function, oliguria, and anuria. Occasionally, patients present with a life-threatening infection or bleeding (e.g., intracranial hematoma). Intracranial hemorrhage occurs mainly in patients with an initial leukocyte count greater than 400 × 10⁹/L.⁷⁰ Very rarely, ALL produces no signs or symptoms and is detected during routine examination.

PHYSICAL FINDINGS

Among frequent findings are pallor, petechiae, and ecchymosis in the skin and mucous membranes, and bone tenderness as a result of leukemic infiltration or hemorrhage that stretches the periosteum. Liver, spleen, and lymph nodes are the most common sites of extramedullary involvement, and the degree of organomegaly is more pronounced in children than in adults. An anterior mediastinal (thymic) mass is present in 8 to 10 percent of childhood cases and in 15 percent of adult cases (Fig. 91–3). A bulky, anterior mediastinal mass can compress the great vessels and trachea and possibly lead to the superior vena cava syndrome. Patients with this syndrome present with cough, dyspnea, orthopnea, stridor, cyanosis, dysphagia, facial edema, increased intracranial pressure, and sometimes syncope. Painless enlargement of the scrotum can be a sign of testicular leukemia cell infiltration or hydrocele, the latter resulting from lymphatic obstruction. Both conditions can be readily diagnosed by ultrasonography. Overt testicular disease is relatively rare, is generally seen in infants or adolescents with T-cell leukemia and/or hyperleukocytosis, and does not require radiation therapy.⁷¹ Other uncommon presenting features include ocular involvement (leukemic infiltration of the orbit, optic nerve, retina, iris, cornea, or conjunctiva), subcutaneous nodules (leukemia cutis), enlarged salivary glands (Mikulicz syndrome), cranial nerve palsy, and priapism (resulting from leukostasis of the corpora cavernosa and dorsal veins or sacral nerve involvement). Epidural spinal cord compression at presentation is a rare but serious finding that requires immediate treatment to prevent permanent paraparesis or paraplegia. In some pediatric patients, infiltration of tonsils, adenoids, appendix, or mesenteric lymph nodes leads to surgical intervention before leukemia is diagnosed.

Anemia, neutropenia, and thrombocytopenia are common in patients with newly diagnosed ALL. The severity reflects the degree of marrow replacement by leukemic lymphoblasts (Table 91–3).^18,65,66,67 Presenting leukocyte counts range widely, from 0.1 to 1500 × 10⁹/L (median: 10–12 × 10⁹/L). Hyperleukocytosis (>100 × 10⁹ white cells/L) occurs in 11 to 13 percent of white children, but occurs more often in black children (23 percent) and in adults (16 percent) because they are more likely to have T-cell ALL. Profound neutropenia (<0.5 × 10⁹/L) is found in 20 to 40 percent of patients, elevating their risk for infection. Approximately 90 percent of patients have circulating leukemic blast cells at diagnosis. Hypereosinophilia, generally reactive, may precede the diagnosis of ALL by several months. Some patients, mostly male, have ALL with the t(5;14)(q31;q32) chromosomal abnormality and a hypereosinophilic syndrome (pulmonary infiltration, cardiomegaly, and congestive heart failure). These patients often do not have circulating leukemic blasts or other cytopenias and have a relatively low percentage of blasts in the marrow.⁷² Activation of the interleukin-3 gene on chromosome 5 by the enhancer element of the immunoglobulin heavy-chain gene on chromosome 14 is thought to contribute to leukemogenesis and the associated eosinophilia in these cases.⁷² In patients with anemia, hemoglobin levels are lower in younger children. Occasionally, a child with ALL has a hemoglobin level as low as 1 g/dL.

Decreased platelet counts are common at diagnosis (median: 48–52 × 10⁹/L). Severe bleeding is uncommon, even when platelet counts are as low as 20 × 10⁹/L, provided infection and fever are absent.^73,74 Occasional patients, principally male, present with thrombocytosis (>400 × 10⁹/L). Pancytopenia followed by a period of spontaneous hematopoietic recovery may precede the diagnosis of ALL in rare cases.⁷⁵ Coagulopathy, usually mild, can be seen in 3 to 5 percent of patients, most of whom have T-cell ALL, and is only rarely associated with clinical bleeding.⁷⁶ The level of serum lactate dehydrogenase is increased in most patients with ALL and is well correlated with tumor burden. Increased levels of serum uric acid are common in patients with a large leukemia cell burden, reflecting an increased rate of purine catabolism. Leukemic infiltration of the kidneys can lead to increased levels of creatinine, urea nitrogen, uric acid, and phosphorus. Rarely, patients present with hypercalcemia resulting from release of parathyroid hormone-like protein from lymphoblasts and leukemic infiltration of bone. The t(17;19)(q22;p13.3) with E2A-HLF fusion, found in 0.5 percent of B-cell precursor ALL, is associated with adolescent age group, disseminated coagulopathy, hypercalcemia, and dismal prognosis.⁷⁷ Liver dysfunction as a result of leukemic infiltration occurs in 10 to 20 percent of patients and is usually mild. However, recognition of carriers of hepatitis B virus is important because prompt lamivudine therapy can prevent serious complications from virus reactivation during immunosuppressive treatment.⁷⁸

Chest radiography is needed to detect enlargement of the thymus or mediastinal nodes and pleural effusions (see Fig. 91–3). Although bony abnormalities, such as metaphyseal banding, periosteal reactions, osteolysis, osteosclerosis, and osteopenia, can be found in 50 percent of patients, especially children with low leukocyte counts at presentation, skeletal films are not necessary for management. Magnetic resonance imaging (MRI) is useful in patients with suspected vertebral collapse or meningeal or nerve root involvement.

Examination of the cerebrospinal fluid (CSF) is an essential diagnostic procedure. Leukemic blasts can be identified in as many as one-third of pediatric patients and approximately 5 percent of adult patients at diagnosis of ALL; most of these patients lack neurologic symptoms.^79,80 Traditionally, CNS leukemia is defined by the presence of at least 5 leukocytes/μL of CSF (with leukemic blast cells apparent in a cytocentrifuged sample) or by the presence of cranial nerve palsies. However, with the omission of prophylactic cranial irradiation in contemporary clinical trials, the presence of any leukemic blast cells in the CSF is associated with increased risk of CNS relapse and is an indication to intensify intrathecal therapy.⁸⁰ Different opinions exist regarding when the first lumbar puncture should be performed. Most protocols now require the procedure at diagnosis and instill the first dose of chemotherapeutic agents intrathecally. Contamination of the CSF by leukemia cells as a result of traumatic lumbar puncture at diagnosis is associated with an inferior treatment outcome in children with ALL.⁸¹ The risk of traumatic lumbar puncture can be decreased by administering platelet transfusions to thrombocytopenic patients and by having the most experienced clinician perform the procedure.

DIAGNOSIS AND CELL CLASSIFICATION

Examination of a marrow aspirate is recommended for diagnosis of ALL because as many as 10 percent of patients lack circulating blasts at the time of diagnosis and because high concentrations of marrow cells are better than blood cells for genetic studies. Fibrosis or tightly packed marrow can lead to difficulties with marrow aspiration that necessitate biopsy. In patients with marrow necrosis, multiple marrow aspirations are sometimes needed to obtain diagnostic tissue.

Morphologic and Cytochemical Analysis

Diagnosis of ALL begins with morphologic analysis of Romanowsky-stained (Wright-Giemsa or May-Grünwald-Giemsa) marrow films. Lymphoblasts tend to be relatively small (ranging from the same size to twice the size of small lymphocytes) with scanty, often light-blue cytoplasm; a round or slightly indented nucleus; fine to slightly coarse and clumped chromatin; and inconspicuous nucleoli (Fig. 91–4A). In some cases, the lymphoblasts are large, with prominent nucleoli, moderate amounts of cytoplasm, admixed with smaller blasts (Fig. 91–4B). Cytoplasmic granules are found in the lymphoblasts of some patients with ALL (Fig. 91–4C). The granules usually are amphophilic (and stain fuchsia), readily distinguishable from primary myeloid granules (which stain deep purple), and demonstrated to be mitochondria by electron microscopy. B-cell blasts in Burkitt-type ALL are characterized by intensely basophilic cytoplasm, prominent nucleoli, and cytoplasmic vacuolation (Fig. 91–4D).

Cytochemical stains (Sudan black stain and the stains for myeloperoxidase and the nonspecific esterases) distinguish ALL from acute myeloid leukemia but are now less commonly used for diagnosis than immunophenotyping.

Immunologic Classification

Immunophenotyping is an essential part of the diagnostic evaluation. Antibodies distinguish clusters of differentiation (CD) groups, but most leukocyte antigens lack specificity. Hence, a panel of antibodies is needed to establish the diagnosis and to distinguish among the different immunologic subclasses of leukemic cells. Typical panels include antibodies to at least one highly sensitive marker (CD19 for B-cell lineage, CD7 for T-cell lineage, and CD13 or CD33 for myeloid cells) and antibodies to a highly specific marker (cytoplasmic CD79a and CD22 for B-cell lineage, cytoplasmic CD3 for T-cell lineage, and cytoplasmic myeloperoxidase for myeloid cells).¹⁷ Although ALL can be further subclassified according to the recognized steps of normal maturation within the B-cell lineage (pro-B, early pre-B, pre-B, transitional pre-B, and mature B cells) or T-cell lineage (pre-T, mid-, and late thymocyte) pathways, the only distinctions of therapeutic importance at present are those between T-cell, mature B, and other B-cell lineage (B-cell precursor type) immunophenotypes (Table 91–4).¹⁷ A distinct subset of T-cell ALL that retain stem cell–like features, termed early T-cell precursor ALL, has been identified and associated with a dire prognosis with conventional chemotherapy.⁸² Knowing the pattern of antigen expression at diagnosis is critically important for the detection of minimal residual disease by flow cytometry after treatment.⁸³

Myeloid-associated antigens may be aberrantly expressed on otherwise typical lymphoblasts. Because of differences in monoclonal antibodies and immunophenotyping techniques, the frequencies of myeloid-associated antigen expression range from 5 to 30 percent in childhood cases and from 10 to 50 percent in adult cases.^67,84 The pattern of myeloid-associated antigen expression is correlated with certain genetic features of blast cells. CD15, CD33, and CD65 are expressed in ALL cases with a rearranged MLL gene, and CD13 and CD33 are expressed in cases with the ETV6-RUNX1.⁸⁴ There is a subset of cases that coexpress both lymphoid and myeloid markers but do not cluster with T-cell, B-cell precursor, or acute myeloid leukemia in gene-expression profiling. These cases may not respond to myeloid-directed therapy but attain remission with ALL-directed induction treatment.^85,86 The presence of myeloid-associated antigens lacks prognostic significance in contemporary treatment programs but can be useful in immunologic monitoring of patients for minimal residual leukemia.^67,83

Genetic Classification

ALL arises from a lymphoid progenitor cell that has sustained multiple specific genetic alterations that lead to malignant transformation and proliferation. Thus, genetic classification of ALL yields more relevant biologic information than any other means. Approximately 75 percent of adult and childhood cases can be readily classified into prognostically or therapeutically relevant subgroups based on the modal chromosome number (or DNA content estimated by flow cytometry), specific chromosomal rearrangements, and molecular genetic changes.^{1,2,17,18,19,20,41,54,55,56,87,88,89} Table 91–5 summarizes the prominent clinical and biologic features of cases with the most common genetic abnormalities.

Two ploidy groups (hyperdiploidy >50 chromosomes and hypodiploidy <44 chromosomes) have clinical relevance. Hyperdiploidy, which is seen in approximately 25 percent of childhood cases and in 6 to 7 percent of adult cases, is associated with a favorable prognosis that may reflect an increased cellular accumulation of methotrexate and its polyglutamates, an increased sensitivity to therapeutic antimetabolites, and a marked propensity of these cells to undergo apoptosis.^90,91 By contrast, hypodiploidy is associated with an exceptionally poor prognosis.^88,89,92 Flow cytometric determination of cellular DNA content is a useful adjunct to cytogenetic analysis because it is automated, rapid, and inexpensive, and its measurements are not affected by the mitotic index of the cell population; results can be obtained in almost all cases. Flow cytometric studies can sometimes identify a small but drug-resistant subpopulation of near-haploid cells that may be missed by standard cytogenetic analysis.

Reciprocal chromosomal translocations in cases of B-cell and T-cell ALL arise from errors in the normal recombination mechanisms that generate antigen receptor genes. Such rearrangements can fuse the promoter/enhancer element of the immunoglobulin heavy- or light-chain gene or the T-cell antigen receptor β/γ or α/δ gene to sites adjacent to a variety of transcription factor genes. More often, genetic rearrangements result from the fusion of two genes encoding different transcription factors. These chimeric genes encode active kinases and altered transcription factors that regulate genes involved in the differentiation, self-renewal, proliferation, and drug resistance of hematopoietic stem cells.

Specific cytogenetic findings are correlated with presenting clinical features, blast-cell phenotypes, and clinical outcome (see Table 91–5). However, there are now compelling reasons to focus on molecular genetic lesions. First, molecular analyses can identify important submicroscopic genetic alterations not visible by standard karyotyping, such as the ETV6-RUNX1 fusion, intrachromosomal amplification of chromosome 21, deletions of tumor-suppressor genes, and mutations of protooncogenes.^{1,2,87,93,94,95} Second, cases with clinically important genetic rearrangements can be missed because of technical errors (e.g., karyotyping residual normal metaphase cells rather than leukemic metaphase cells). Hence, FISH and reverse transcriptase polymerase chain reaction (RT-PCR) assays are increasingly used. The application of microarray-based genome-wide analysis of gene expression and DNA copy number, complemented by transcriptional profiling, resequencing and epigenetic approaches, and next-generation sequencing have identified specific genetic alterations with biologic and therapeutic implications.

The initial manifestations of ALL can mimic a variety of disorders. The acute onset of petechiae, ecchymoses, and bleeding can suggest idiopathic thrombocytopenic purpura. The latter disorder often is associated with a recent viral infection, large platelets in blood films, normal hemoglobin concentration, and absence of leukocyte abnormalities in blood or marrow. Patients with ALL, or promyelocytic leukemia, or aplastic anemia can present with pancytopenia and complications associated with marrow failure. However, in aplastic anemia, hepatosplenomegaly and lymphadenopathy are rare, and the skeletal pain associated with ALL is absent. The results of marrow aspiration or biopsy usually distinguish between these diseases, although the diagnosis can be difficult in a patient who has hypocellular marrow that is later replaced by lymphoblasts. In one study, transient pancytopenia preceded ALL in 2 percent of all pediatric cases.⁷⁵ During the preleukemic phase in these patients, polymerase chain reaction (PCR) analysis demonstrated monoclonality, suggesting inhibition of normal hematopoiesis by leukemia cells.⁹⁶ ALL should be considered in the differential diagnosis of patients with hypereosinophilia, which can be a presenting feature of leukemia or can precede its diagnosis by several months. Occasionally, hematogones in a regenerative marrow may mimic leukemic blast cells; flow cytometry with optimal combinations of antibodies may be required to distinguish them.⁹⁷

Infectious mononucleosis and other viral infections, especially those associated with thrombocytopenia or hemolytic anemia, can be confused with leukemia. Detection of reactive lymphocytes or serologic evidence of Epstein-Barr virus infection helps establish the diagnosis. Patients with acute infectious lymphocytosis, pertussis, or parapertussis can have marked lymphocytosis. However, even when leukocyte counts are as high as 50 × 10⁹/L, the affected cells are mature lymphocytes rather than lymphoblasts. Bone pain, arthralgia, and occasionally arthritis mimic juvenile rheumatoid arthritis, rheumatic fever, other collagen diseases, or osteomyelitis. The marrow should be examined if glucocorticoid treatment is planned for presumed rheumatoid diseases.

In children, ALL should be distinguished from small, round cell tumors involving the marrow, including neuroblastoma, rhabdomyosarcoma, and retinoblastoma. Generally, in patients with solid tumors, a primary lesion will be found by standard diagnostic studies. Disseminated tumor cells often present in characteristic aggregates, and immunophenotypic characteristics of lymphoblasts are absent.

SUPPORTIVE CARE

Optimal management of patients with ALL requires careful attention to supportive care, including immediate treatment or prevention of metabolic and infectious complications (Chap. 24) and rational use of blood products (Chaps. 138 and 139). Other important supportive care measures, such as use of indwelling catheters, amelioration of nausea and vomiting, pain control, and continuous psychosocial support for the patient and family, are essential.

Metabolic Complications

Hyperuricemia and hyperphosphatemia with secondary hypocalcemia are frequently encountered at diagnosis, even before chemotherapy is initiated, especially in patients with B-cell or T-cell ALL or precursor B-cell leukemia with high leukemic cell burden.⁹⁸ Patients should be given intravenous fluids; allopurinol or rasburicase (recombinant urate oxidase) to treat hyperuricemia; and a phosphate binder, such as aluminum hydroxide, calcium carbonate (if the serum calcium concentration is low), lanthanum carbonate, or sevelamer to treat hyperphosphatemia. Allopurinol, a relatively inexpensive drug, is usually used if the uric acid is less than 7 mg/dL. Allergic skin reactions occur in approximately 10 percent, and allopurinol should be stopped as soon as the risk of hyperuricemia from the destruction of a large leukemic cell burden has passed. By inhibiting de novo purine synthesis in leukemic blast cells, allopurinol can reduce the peripheral blast-cell count before chemotherapy.⁹⁹ Allopurinol can decrease both the anabolism and catabolism of mercaptopurine by depleting intracellular phosphoribosyl pyrophosphate and by inhibiting xanthine oxidase. If mercaptopurine and allopurinol are given together orally, the dosage of mercaptopurine must be reduced.

Rasburicase works very rapidly and is extremely effective, especially for very elevated uric acid levels (>7 mg/dL), often with one infusion (a far smaller dose than the manufacturer recommends). Rasburicase breaks down uric acid to allantoin, a readily excreted metabolite that is five to 10 times more soluble than uric acid. Rasburicase is more effective than allopurinol, and it facilitates phosphorus excretion, partly because of rasburicase’s potent uricolytic effect (which obviates the need to alkalinize urine) and partly because of improved renal function with its use.¹⁰⁰ However, rasburicase is contraindicated in patients with glucose-6-dehydrogenase deficiency because hydrogen peroxide, a by-product of uric acid breakdown, can cause methemoglobinemia or hemolytic anemia.

Hyperleukocytosis

For patients with extreme leukocytosis (leukocyte count >400 × 10⁹/L), either leukapheresis or exchange transfusion (in small children) can be used to reduce the burden of leukemic cells. In theory, either treatment should reduce the complications associated with leukostasis, but the short- and long-term benefits of the procedures are questionable.⁷⁰ Emergency cranial irradiation, once advocated by some leukemia therapists, probably has no role in the treatment of these patients. Preinduction therapy with low-dose glucocorticoids, with addition of vincristine and cyclophosphamide in cases of B-cell ALL, is a favored means of ameliorating hyperleukocytosis. This method, when used in conjunction with urate oxidase, has largely eliminated tumor lysis syndrome and the need for hemodialysis in patients with mature B-cell ALL.

Infection Control

Infections are common in febrile patients with newly diagnosed ALL. Therefore, any patient presenting with fever, especially a patient with neutropenia, should be given broad-spectrum antibiotics until infection is excluded. Remission induction therapy can increase susceptibility to infection by exacerbating myelosuppression, immunosuppression, and mucosal breakdown. At least 50 percent of patients undergoing induction therapy experience infections. Special precautions should be taken to reduce the risk of infection during this critical phase of treatment, including protective contact isolation and air filtration; elimination of contact with people with infections; refraining from eating certain food products, such as raw cheese, uncooked vegetables, or unpeeled fruits; and use of antiseptic mouthwash or sitz baths, especially for patients with mucositis. Good hand washing practices and the use of alcohol-based cleansers are important. Administration of granulocyte colony-stimulating factor can hasten recovery from neutropenia and reduce the complications of intensive chemotherapy, but does not improve the EFS rate for children or adults.^101,102 One study suggested growth factor increased the risk of therapy-related acute myeloid leukemia in the context of epipodophyllotoxin-based therapy.¹⁰³ Intensified remission induction regimens, especially in combination with high-dose glucocorticoids, have resulted in an increased risk of disseminated fungal infection and death. Antifungal prophylaxis is commonly given.

All nonallergic patients with ALL are given trimethoprim-sulfamethoxazole, 2 to 3 days per week, as prophylactic therapy for Pneumocystis carinii (Pneumocystis jiroveci) pneumonia. Prophylaxis is started after 2 weeks of remission induction and continues for several months after completion of all chemotherapy. Alternative treatments for patients who cannot tolerate trimethoprim-sulfamethoxazole include aerosolized pentamidine, dapsone, and atovaquone.¹⁰⁴ Live-virus vaccines should not be administered during immunosuppressive therapy. Siblings and other children who have frequent contact with patients can receive routine immunizations, including inactivated poliomyelitis vaccine. Susceptible patients exposed to varicella virus should receive zoster immunoglobulin within 96 hours of exposure together with acyclovir. Such treatment usually prevents or mitigates the clinical manifestations of varicella.

Hematologic Support

ALL and its treatment leads to pancytopenia. Hemorrhagic manifestations are common but usually are limited to the skin and mucous membranes. Although rare, bleeding in the CNS, lungs, or gastrointestinal tract can be life-threatening. Patients with extremely high leukocyte counts (>400 × 10⁹/L) at diagnosis are more likely to develop such complications.⁷⁰ Coagulopathy attributable to disseminated intravascular coagulation, hepatic dysfunction, or chemotherapy is usually mild.⁷⁶ Patients receiving l-asparaginase and a glucocorticoid have a hypercoagulable state. Platelet transfusions should be given therapeutically for overt bleeding and may be used prophylactically when platelet counts are less than 10 × 10⁹/L.¹⁰⁵ Anticoagulants and antiplatelet agents such as aspirin must be avoided. Children generally do not have active bleeding during remission induction therapy with prednisone, vincristine, and l-asparaginase, even when platelet counts are less than 10 × 10⁹/L. A higher threshold for prophylactic platelet transfusions should be considered for active toddlers and patients with fever or infection. Transfusion of packed leukocyte-poor red cells is indicated in patients with anemia and marrow suppression but should be delayed until the leukocyte count is reduced in patients with extreme hyperleukocytosis. Transfusions should be given slowly in patients with profound but chronic anemia to prevent development of congestive heart failure. Granulocyte transfusions are rarely needed, but should be considered for patients with absolute neutropenia and documented Gram-negative septicemia or disseminated fungal infection that is responding poorly to antimicrobial treatment alone. All blood products should be irradiated to prevent transfusion-related graft-versus-host disease.

ANTILEUKEMIC THERAPY

Because ALL is a heterogeneous disease with many distinct subtypes, there is no uniform approach to therapy. Increasingly, treatment is targeted to biologically distinct subgroups. The best results have been reported from experienced treatment centers using well-designed and rigorously applied protocols.^{106,107,108,109} No consensus exists on the risk criteria and the terminology for defining prognostic subgroups. Usually, childhood ALL cases are divided into standard-risk, high- (intermediate- or average-) risk, and very-high-risk groups, although the United States’ Children’s Oncology Group advocates four categories, including low risk, to accommodate patients with a very low risk of relapse. Adult cases are generally divided into two risk groups. Often infant and elderly ALL are considered special subgroups of ALL that require different treatment, primarily related to intolerance. One study showed improved outcome for infant ALL, using a hybrid treatment protocol with elements to treat both ALL and acute myeloid leukemia, and reducing dose intensity in the very young infants.¹¹⁰ Few studies have been performed in those older than 60 years of age and their management remains a therapeutic challenge.^111,112 Some successes have been achieved with the use of dose-reduced regimens and the addition of imatinib or dasatinib for patients with Philadelphia chromosome–positive ALL.^113,114 Because cure is not common in patients older than age 70 years, maintenance of a good quality of life is a major goal for this age group.

Mature B-Cell Acute Lymphoblastic Leukemia (Burkitt-Type)

The most effective contemporary treatment regimens for mature B-cell (Burkitt-type) ALL are drug combinations that include cyclophosphamide and/or ifosfamide given over a relatively short time (3 to 6 months). The first major breakthrough in this disease was reported by French investigators, who achieved a 68 percent EFS rate in their LMB84 study featuring high-dose cyclophosphamide, high-dose methotrexate, vincristine, doxorubicin, and conventional doses of cytarabine. In the LMB89 study, the same group reported a cure rate of 87 percent, which was achieved by using increased doses of methotrexate (to 8 g/m² per dose) and cytarabine (3 g/m² per dose) and by adding etoposide for patients with a large leukemia cell burden.¹¹⁵ This excellent result has been confirmed in a randomized international study.¹¹⁶ Successful treatments also have been developed by the Berlin-Frankfurt-Münster consortium, which uses a multiagent regimen that incorporates cyclophosphamide, high-dose methotrexate (1 g/m² per dose), etoposide, ifosfamide, doxorubicin, dexamethasone, and cytarabine (3 g/m² per dose).¹¹⁷ These intensive pediatric protocols have been translated into adult treatment regimens that are completed in as little as 18 weeks.^118,119,120 Because of its demonstrated efficacy in B-cell lymphoma, rituximab (anti-CD20) has been incorporated in frontline clinical trials for adults with B-cell ALL.^118,120 Maintenance or continuation therapy is not needed. The remission rate is very high, and most remissions are durable. B-cell ALL rarely, if ever, reoccur after the first year.

Effective CNS therapy is an essential component of successful regimens for B-cell ALL and generally consists of methotrexate and cytarabine administered both systematically and intrathecally. Cranial irradiation does not appear to be necessary even for patients presenting with CNS leukemia.^116,120

Precursor B-Cell and T-Cell Acute Lymphoblastic Leukemia

Treatment for leukemias affecting the precursor B-cell and T-cell lineages consists of three standard phases: remission induction, intensification (consolidation), and prolonged continuation (maintenance) therapy.¹²¹ CNS-directed therapy, which overlaps other treatments, is started early and is given for different lengths of time, depending on the patient’s risk of relapse and the intensity of the primary systemic regimen.

Remission Induction The first goal of therapy is inducing a complete remission and restoring normal hematopoiesis. The induction regimen typically includes a glucocorticoid (prednisone, prednisolone, or dexamethasone), vincristine, and l-asparaginase for children or an anthracycline for adults.^{17,67,106,107,108} Children with high- or very-high-risk ALL, and nearly all young adults with ALL, receive four or more drugs (daunorubicin, vincristine, glucocorticoid, and l-asparaginase) during remission induction in contemporary clinical trials. Improvements in chemotherapy and supportive care have resulted in complete remission rates of approximately 98 percent for children and 85 to 90 percent for adults. When a complete clinical remission is induced, patients have various degrees of residual leukemia.¹²² Because the extent of residual disease is well correlated with long-term outcome,^{83,123,124,125,126,127,128,129} the concept of a “molecular” or “immunologic” remission, defined as leukemia less than 0.01 percent of nucleated marrow cells,¹²² is beginning to supplant the traditional perception of remission, which is based solely on microscopic criteria.¹²⁹ Prospective trials are needed to demonstrate that outcomes will improve when interventions to change therapy are based on measurements of residual disease.

Attempts have been made to intensify induction therapy based on the premise that more rapid and complete reduction of the leukemia cell burden forestalls the development of drug resistance. However, several studies have suggested intensive induction therapy is unnecessary for children with standard-risk ALL, provided patients receive postinduction intensification therapy.¹⁰⁷ Intensive induction can lead to increased early morbidity and mortality. More intensive induction regimens with additional cyclophosphamide, high-dose cytarabine, or high-dose anthracycline also have been tested in adults with ALL and have yielded no clear benefit, partly because of the low tolerance of adults to drug toxicity.^130,131 However, in one study, the use of high-dose dexamethasone (10 mg/m² per day) instead of prednisone (60 mg/m² per day) during remission induction significantly improved treatment outcome for children with ALL, especially those with T-cell ALL and good prednisone response, despite a higher induction death rate.^121,132 Conceivably, intensified remission induction with other relatively nonmyelosuppressive drugs can also improve treatment outcome. Dexamethasone provided better control of systemic and CNS disease than did prednisone in two randomized studies of childhood ALL.^133,134

The pharmacodynamics of asparaginase differ by formulation and three forms are available: one derived from Erwinia chrysanthemi, another prepared from Escherichia coli, and a third made of a polyethylene glycol form of the E. coli product (pegaspargase).^135,136 In terms of leukemic control, the dose intensity and duration of asparaginase treatment (i.e., the amount of asparagine depletion) are far more important than the type of asparaginase used. The dosages of the three preparations are based on their half-lives. Pegaspargase, which has the longest half-life, usually is administered at 2500 IU/m² every other week for one to two doses in cases of newly diagnosed ALL. By contrast, the Erwinia preparation, which has the shortest half-life, is administered at 20,000 IU/m² three times per week for 6 to 12 doses. The doses of E. coli l-asparaginase range from 5000 to 10,000 IU/m², administered two to three times per week for 6 to 12 doses. Because of lower immunogenicity, improved efficacy, and less-frequent administration, pegaspargase has replaced the native product as the first-line treatment for children and adults in the United States, and is also increasingly used in other ALL trials around the world.^{137,138,139,140} None of the various anthracyclines (daunorubicin, doxorubicin, idarubicin, and mitoxantrone) given to adults with ALL has proven superior to any other; daunorubicin is used most commonly.

Intensification (Consolidation) Therapy After normal hematopoiesis is restored, patients in remission become candidates for intensification therapy. Such treatment, administered shortly after remission induction, refers to readministration of the induction regimen or to high doses of multiple agents not used during the induction phase. Although there is no dispute on the importance of this treatment in childhood ALL, there is no consensus on the best regimen and duration of treatment. More commonly used regimens for childhood ALL include high-dose methotrexate with or without mercaptopurine, high-dose l-asparaginase given for an extended period, or a combination of dexamethasone, vincristine, l-asparaginase, and doxorubicin, followed by thioguanine, cytarabine, and cyclophosphamide.^{107,108,121,137,141,142} This phase of therapy has improved outcomes, even for patients with low-risk ALL.¹⁴³ Patients with ETV6-RUNX1 have an especially good outcome in clinical trials featuring intensive postremission treatment with glucocorticoids, vincristine, and asparaginase.^144,145 A very high dose of methotrexate (5 g/m²) appears to improve the treatment outcome of patients with T-cell ALL.^141,146 This finding is consistent with data indicating T-cell blasts accumulate methotrexate polyglutamates (active metabolites of the parent compound) less avidly than do B-cell precursors; consequently, higher serum levels of the drug are needed for an adequate therapeutic effect.^147,148 The conventional dose of methotrexate (1 g/m²) may be too low for many patients with B-cell precursor ALL. Among B-lineage ALL, blasts with either ETV6-RUNX1 or TCF3-PBX1 gene fusion accumulate significantly lower methotrexate polyglutamates compared to those with hyperdiploidy or other genetic abnormalities.¹⁴⁹ This finding suggested that patients with ETV6-RUNX1 or TCF3-PBX1 gene fusion would also benefit from a higher dose of methotrexate.

Based on pediatric studies, intensive consolidation therapy has become a standard in the treatment of adult ALL even though early studies failed to show the benefit of this phase of treatment. Various drugs have been used for intensification, including high-dose methotrexate, high-dose cytarabine, cyclophosphamide, and asparaginase.^{67,102,150,151,152,153} Increasingly, intensification treatment is risk-adapted and subtype specific. In the German 06/93 study, high-dose methotrexate was used for patients with standard-risk B-cell precursor ALL, high-dose methotrexate and high-dose cytarabine for high-risk B-cell precursor ALL, and cyclophosphamide for T-cell ALL. The hyper-CVAD (cyclophosphamide, vincristine, Adriamycin, dexamethasone) regimen of the MD Anderson Cancer Center alternates the combination of cyclophosphamide, vincristine, doxorubicin (Adriamycin), and dexamethasone, with high-dose methotrexate and high-dose cytarabine for four courses each. More recently, rituximab has been added for patients with CD20 expression on lymphoblasts.^150,151 In adults, methotrexate dose should probably be limited to 1.5 to 2.0 g/m² because higher doses may lead to excessive toxicities, delayed subsequent treatment, and reduced compliance. In a Cancer and Leukemia Group B study, a five-drug remission induction was followed by early and late intensification courses with eight drugs.⁶⁷ These studies and others suggested the benefit of early intensive consolidation therapy, especially in young adults. In adult T-cell ALL, benefit is derived from cyclophosphamide and cytarabine.^67,102 In other adult cases of standard-risk and high-risk ALL, the benefit is derived from high-dose cytarabine. Two German multicenter trials using high-dose cytarabine, mitoxantrone, and allogeneic hematopoietic stem cell transplantation showed markedly improved results in cases bearing the t(4;11), which generally confers an adverse prognosis.¹⁵⁴ Several ongoing trials are testing the efficacy of asparaginase during intensification in young adult ALL¹³⁷ because this drug clearly improves outcome in childhood ALL and is better tolerated during consolidation treatment than during remission induction.

Patients diagnosed with ALL between the ages of 16 and 39, often considered together as adolescents and young adults, are commonly treated by either adult or pediatric hematologists. Several retrospective comparative analyses have reported that adolescents and young adults with ALL treated on pediatric protocols have had superior event-free and overall survival rates when compared with similar patients enrolled on adult ALL trials. Preliminary data suggest that these patients have better outcomes when treated with pediatric-inspired regimens, and that excess toxicity is not observed.^{153,155,156,157,158,159,160,161,162}

Continuation Therapy Although unnecessary for cure of mature B-cell leukemia, continuation therapy for 2 to 3 years is an integral part of pediatric and adult ALL regimens. Attempts to shorten the duration of treatment have led to inferior outcomes in both childhood and adult ALL,¹⁶³ although as many as two-thirds of childhood cases might be cured with only 12 months of treatment.¹⁶⁴ However, which subgroups of childhood ALL can be cured with abbreviated therapy is unclear. In a meta-analysis of 42 trials, a third year of continuation therapy reduced the likelihood of relapse during the third year, but no advantage to prolonging treatment beyond 3 years was observed.¹⁶⁵ Early studies demonstrated that the third year of continuation therapy benefits boys but not girls.^166,167 Hence, most studies discontinue all therapy for girls after 2 to 2.5 years of treatment. It is uncertain whether with improved contemporary treatment boys still require prolonged continuation treatment. Whether adults with ALL benefit from prolonged continuation therapy is also unclear. In most adult trials, continuation therapy is given for 2 years from diagnosis.

Methotrexate administered weekly and mercaptopurine administered daily constitute the usual continuation regimen for ALL. Accumulation of higher intracellular concentrations of the active metabolites of methotrexate and mercaptopurine and administration of this combination to the limits of tolerance (as indicated by low leukocyte counts) have been associated with improved clinical outcome.^168,169 Many investigators advocate that drug dosage be adjusted to maintain leukocyte counts below 3 × 10⁹/L and neutrophil counts between 0.5 and 1.5 × 10⁹/L to ensure adequate dose intensity during the continuation treatment in childhood ALL.² In one study, the dose intensity of mercaptopurine was the most important pharmacologic factor influencing treatment outcome.¹⁷⁰ Mercaptopurine is most effective when it is given orally on a daily basis. However, overzealous use of mercaptopurine is counterproductive, as such use results in neutropenia and interruption of chemotherapy, reducing overall dose intensity. The effect of mercaptopurine is better when the drug is administered in the evening.¹⁷¹ Mercaptopurine should not be given with milk or milk products containing xanthine oxidase, which can degrade the drug.¹⁷² Antimetabolite treatment should not be withheld because of isolated increases of liver enzymes since such liver function abnormalities are tolerable and reversible.¹⁷³

A few patients (one in 300) have an inherited homozygous deficiency of thiopurine S-methyltransferase, the enzyme that catalyzes the S-methylation (inactivation) of mercaptopurine. In these patients, standard doses of mercaptopurine have potentially fatal hematologic side effects. The drug should be given in much smaller doses (e.g., 10-fold reduction).¹⁷⁴ Approximately 10 percent of patients are heterozygous for the enzyme deficiency and have intermediate levels of thiopurine methyltransferase.¹⁷⁵ This subgroup can be treated safely with only moderate reductions in mercaptopurine dosage and appears to have better clinical outcomes than do patients with the homozygous wild-type phenotype. Importantly, patients with this enzyme deficiency are at risk for therapy-related myeloid leukemia and radiation-related brain tumors.^176,177,178 Identification of this autosomal codominant trait has been enabled by molecular diagnosis and led to increased emphasis on inherited differences in drug metabolism and disposition resulting from genetic polymorphisms in drug-metabolizing enzymes and in drug transporters, receptors, and targets.^1,179,180 Ultimately, therapy can be designed according to the genetic constitution of both the host and the host’s leukemia cells.

Because thioguanine is more potent than mercaptopurine in model systems and leads to higher concentrations of thioguanine nucleotides in cells and cytotoxic concentrations in CSF,¹⁸¹ randomized trials have been performed in children to compare the effectiveness of these two drugs.¹⁸² Thioguanine, given at a daily dose of 40 mg/m² or more, produced superior antileukemic responses to mercaptopurine, but was associated with profound thrombocytopenia, an increased risk of death, and unacceptable rate of hepatic venoocclusive disease.¹⁸² Consequently, mercaptopurine, remains the drug of choice for ALL, although thioguanine is still used in short-term courses during the intensification phase of therapy.

Intermittent pulses of vincristine and a glucocorticoid improved the efficacy of antimetabolite-based continuation regimens and have been widely adopted for both childhood and adult ALL.^165,183 In older children and adults, prolonged glucocorticoid therapy may lead to increased risk of osteonecrosis.¹⁸⁴ Another integral component of many protocols is reinduction therapy introduced relatively soon after the first remission. This treatment, which relies on the same drugs used during the initial phase of induction therapy, has improved outcomes for children and adults with ALL.^{67,106,107,108} A second reinduction phase during continuation treatment may further improve the outcome of patients with standard- or high-risk ALL.^142,185 The benefit of such a double-delayed intensification may result from either the increased dose intensity of other agents such as asparaginase or anthracycline or the timing or scheduling of the intensification regimen.

Therapy of the Central Nervous System The CNS is a common sanctuary for leukemic cells and requires prophylactic or presymptomatic therapy. In the 1970s, the cornerstone of ALL therapy was cranial irradiation (2400 cGy) plus methotrexate administered intrathecally after complete remission was induced. Concerns that cranial irradiation could cause second cancer, late neurocognitive deficits, and endocrinopathy stimulated efforts to replace cranial irradiation with early intensification by intrathecal and systemic chemotherapy. Two early clinical trials tested the feasibility of complete omission of prophylactic cranial irradiation in the treatment of childhood ALL.^186,187,188 Although the cumulative risk of an isolated CNS relapse was relatively low (3 to 4 percent), the EFS rates in the two studies were only 68.4 percent and 60.7 percent.^186,187,188 In another study, prophylactic cranial irradiation appeared to improve outcome in T-cell ALL with leukocyte counts >100 × 10⁹/L.¹⁸⁹ Thus, until recently, virtually all childhood study groups continued to rely on prophylactic cranial irradiation for up to 20 percent of patients.⁸⁰ A radiation dose of 1200 cGy appeared to provide adequate protection against CNS relapse, even in high-risk patients (e.g., those with T-cell ALL and leukocyte counts >100 × 10⁹/L).¹⁴¹ More recently, a study at St. Jude Children’s Research Hospital again tested the feasibility of total omission of prophylactic cranial irradiation in the context of risk-adapted intrathecal and systemic chemotherapy.¹⁹⁰ The 5-year survival rate for the 498 patients enrolled was 93.5 percent and the cumulative risk of an isolated CNS relapse rate was only 2.7 percent, a promising result, suggesting that prophylactic cranial irradiation can be safely omitted in the context of the effective intrathecal and systemic chemotherapy. Another study by the Dutch Childhood Oncology Group also showed that prophylactic cranial irradiation can be safely omitted from children with ALL.¹⁹¹ Preliminary data from additional trials also indicate that prophylactic cranial irradiation is not necessary.

Systemic treatment including high-dose methotrexate, intensive asparaginase, and dexamethasone, as well as optimal intrathecal therapy, is important to control CNS leukemia.^80,192 Triple intrathecal therapy with methotrexate, cytarabine, and hydrocortisone is more effective than intrathecal methotrexate in preventing CNS relapse.¹⁹³ Because the presence of ALL blasts in the CSF, even from traumatic lumbar puncture, is associated with an increased risk of CNS relapse and poor EFS,^80,81 intrathecal therapy should be intensified in patients with this feature. With CNS prophylaxis and high-dose systemic therapy most adults with ALL remain free of CNS disease. CNS disease at the time of leukemia relapse in adults occurs in approximately 10 percent of cases. The frequency of CNS recurrence is about the same whether CNS radiation therapy (12 to 24 Gy) is used or whether only intrathecal cytotoxic therapy is used. Systemic high-dose methotrexate and cytarabine add to the CNS therapy. The outcome after CNS relapse is poor. Survival after CNS relapse is usually less than 1 year in adults. Treatment of CNS relapse requires cranial irradiation, intrathecal chemotherapy, typically via an Ommaya shunt, plus reinduction and reconsolidation systemic therapy.

Stem Cell Transplantation Hematopoietic stem cell transplantation during first remission remains controversial.¹⁹⁴ In adult ALL, long-term disease-free survival rates range from 35 to 50 percent with chemotherapy alone and from 45 to 60 percent with allogeneic transplantation.^195,196 However, interpretation of these results is difficult because of the lack of true randomization. Even so, results from both adult and pediatric studies suggest allogeneic transplantation benefits some high-risk patients.^194,196,197 Because of their unfavorable prognosis, patients with the Philadelphia chromosome–positive ALL and those with a poor initial response to induction therapy have been recommended to undergo allogeneic stem cell transplantation during the first remission.^{194,195,196,198} However, the advent of improved chemotherapy has diminished the survival advantage from transplantation in children with Philadelphia chromosome–positive ALL.¹⁹⁹ The use of a tyrosine kinase inhibitor has further improved the early treatment results,²⁰⁰ casting doubt on the benefit from transplantation in first remission in childhood cases.²⁰¹ Allogeneic transplantation appears to improve the outcome of adults with the t(4;11),¹⁵⁴ but not that of children or infants with the same genotype.²⁰²

Allogeneic transplantation has not been compared to chemotherapy alone in a true randomized trial, and thus the results of comparative studies are biased by availability of appropriate donors and other factors.^{194,203,204,205} The more potent antileukemia activity of allogeneic transplantation is balanced against the considerable nonrelapse mortality and long-term consequences of graft-versus-host disease. A meta-analysis involving 3157 patients supports matched sibling donor allogeneic transplantation as the optimal postremission therapy for adults with ALL with a significant reduction in relapses and a significant increase in treatment related mortality. Results may differ depending upon stem cell source, that is, related or unrelated donors or umbilical cord stem cells. A retrospective study of 421 adults who underwent allogeneic cord blood transplantation reported 2-year leukemia-free survival of 39 percent for patients in first complete remission and 31 percent for second remission. In multivariate analysis, factors associated with poor outcomes were age older than 35 years, myeloablative conditioning, and more advanced disease. Reduced intensity-conditioning allografting has yielded lower nonrelapse mortality and higher relapse rates than myeloablative conditioning with no significant differences in leukemia-free survival.^206,207 The indications for allogeneic transplantation in first remission should be reevaluated as chemotherapy and transplantation continue to improve. Autologous transplantation failed to improve outcome in adult ALL overall, mainly because of a high rate of relapse.¹⁹⁵ Several small studies indicate that autologous stem cell transplantation is feasible and beneficial for adults with Philadelphia chromosome–positive ALL who achieve a molecular remission after combined chemotherapy and imatinib.^208,209

Targeted Therapies The best example of targeted therapy is the use of the tyrosine kinase inhibitors imatinib or dasatinib in Philadelphia chromosome–positive ALL.^{21,113,114,210} Used as single agents or with a glucocorticoid, they can induce complete remission in older patients where this subset of ALL is more common.^114,210,211 In combination with chemotherapy, they not only induce a higher complete remission rate but also a high rate of molecular remission in children and adults.^{200,201,212,213,214,215,216} The duration of these remissions is uncertain, but some have been quite long after additional chemotherapy. Although the need for early transplantation in childhood cases is uncertain, this treatment modality is still a standard for adult cases.^206,217 The use of imatinib or dasatinib yields a higher proportion of adult cases suitable for transplantation. The outcome depends on minimal residual disease before and after transplantation.²¹⁸ In patients with residual disease after transplantation, rapid response to imatinib was associated with a superior survival.²¹⁹ It is uncertain whether and when to discontinue imatinib or dasatinib after the patient is treated with chemotherapy or transplantation.

Surface expression of CD20 by leukemia cells is associated with an inferior outcome in adult,²²⁰ but not childhood, ALL.²²¹ Chemotherapy trials incorporating rituximab, an anti-CD20 antibody, have yielded promising results in adults with CD20-positive B-cell precursor ALL.^150,222 Other monoclonal antibodies that bind CD22 and CD19 are in late-stage clinical development.²²³ Nelarabine is an approved antimetabolite drug that has shown considerable activity in T-cell ALL, both alone and in combination with other chemotherapy.^224,225,226

RELAPSE

Relapse is defined as the reappearance of leukemia cells at any site in the body. Most relapses occur during treatment or within the first 2 years after its completion, although initial relapses have been observed 10 or more years after diagnosis.²²⁷ Molecular studies suggest that in some cases, especially those with the ETV6-RUNX1 fusion, subsequent mutations of the residual preleukemic clone that were not eradicated during initial treatment account for the “late relapse.”²²⁸ The marrow remains the most common site of relapse in ALL. Anemia, leukocytosis or leukopenia, thrombocytopenia, enlargement of the liver or spleen, bone pain, fever, or a sudden decrease in tolerance to continuation chemotherapy may signal the onset of marrow relapse. In contemporary treatment programs for childhood ALL, the rates of CNS and testicular relapse have decreased to 3 percent or less.^190,192 Leukemic relapse occasionally occurs at other extramedullary sites, including the eye, ear, ovary, uterus, bone, muscle, tonsil, kidney, mediastinum, pleura, and paranasal sinus.

For ALL patients who have received a modern, intensive multiagent treatment regimen, with delayed intensification and prolonged continuation therapy, a relapse of disease portends a very poor survival. Although some individuals can be rescued with additional chemotherapy alone, in general, only allogeneic hematopoietic stem cell transplantation offers a reasonable chance for cure and long-term survival. Thus, treatment for relapse is often considered as “a bridge to transplant.” Two new drugs, clofarabine and liposomal vincristine, were approved as single agents in large part based upon their ability to enable patients with relapsed ALL to proceed to a transplant.^229,230 Because the outcome of allogeneic transplantation is poor in the presence of overt ALL, reinduction therapy aims for rapid cytoreduction of lymphoblasts while efforts proceed concurrently to identify an HLA-matched donor. The optimal pretransplant conditioning regimen then depends largely on the age and medical condition of the patient.

Marrow relapse, with or without extramedullary involvement, portends a poor outcome for most patients.^231,232 Factors indicating an especially poor prognosis include relapse while on therapy or after a short initial remission, T-cell immunophenotype, the presence of the Philadelphia chromosome, and an isolated hematologic relapse.^232,233,234 Prolonged second remissions (>3 years) can be achieved with chemotherapy in as many as half of patients with late relapses (i.e., >6 months after cessation of therapy) but in only approximately 10 percent of those with early relapse. The persistence of minimal residual disease after reinduction treatment also portends a very poor prognosis.²³⁵ In patients who develop hematologic relapse while on therapy or shortly thereafter, and in those with residual disease after remission induction for relapse, allogeneic hematopoietic stem cell transplantation is the treatment of choice.²³⁶ Autologous transplantation as postinduction treatment offers no substantial advantage over chemotherapy.^190,192,237 For patients without histocompatible related donors, transplantation of stem cells from cord blood or marrow from matched unrelated donors is recommended.^{201,238,239,240,241} For patients with ALL that has relapsed after allogeneic transplantation, a second transplant or donor T-lymphocyte infusion occasionally results in sustained remission.²⁴²

Central Nervous System Relapse

Although extramedullary relapse is frequently an isolated clinical finding, many occurrences are associated with recurrent disease detectable in the marrow. CNS relapses are associated with higher level of minimal residual disease in the marrow than testicular relapses.²⁴³ Submicroscopic marrow involvement at a level of 10⁻⁴ or higher at the time of overt extramedullary relapse confers a very poor outcome.²⁴³ Hence, patients with extramedullary relapse and detectable disease in marrow require intensive systemic treatment to prevent subsequent hematologic relapse. The efficacy of salvage therapy in children with an isolated CNS relapse depends partly on duration of first complete remission and partly on whether CNS irradiation was previously performed. The strategy of delaying cranial or craniospinal irradiation for 6 to 12 months to allow initial intensification with systemic chemotherapy has yielded long-term second EFS of 70 to 80 percent in children with isolated CNS relapse.^244,245 In one study, 12 months of intensive systemic chemotherapy and reduced-dose cranial irradiation (18 Gy) resulted in an excellent 4-year EFS rate among children with precursor B-cell ALL who had not received cranial irradiation during initial treatment and who had an initial remission duration of 18 months or more.²⁴⁶ Notably, in this study, a favorable age group of 1 to 9.9 years plus a low presenting leukocyte count (<50 × 10⁹/L) at diagnosis of ALL was an independent favorable prognostic factor. For patients, especially those with T-cell ALL, in whom relapse develops during therapy and who had previously undergone cranial irradiation, the remission rate generally does not exceed 30 percent.^244,245 Adults with isolated CNS relapse fare much more poorly than children. However, the recommended treatment strategy remains the same–combining CNS-directed treatment with additional systemic chemotherapy.

Testicular Relapse

One-third of patients with early testicular relapse and two-thirds of patients with late testicular recurrence became long-term survivors after salvage chemotherapy and testicular irradiation.^232,247,248 In one study, some patients with late isolated testicular relapses were successfully treated with chemotherapy that included very high-dose methotrexate, without the addition of radiation therapy.²⁴⁹ The optimal treatment and prognosis for patients with relapse at unusual extramedullary sites are unclear. However, the same principles that apply to the clinical management of CNS or testicular relapse probably apply to this subgroup.

TREATMENT SEQUELAE

Despite the increasing intensity of curative treatment for childhood ALL, judicious use of supportive care reduced the rate of early death from 8 percent in the early 1970s to less than 2 percent in the 1990s.^16,17 Currently, the induction mortality ranges between 2 percent and 11 percent in adult ALL, with increasing age associated with higher death rate.^{66,67,150,151,152,153} Most deaths are caused by bacterial or fungal infections. The death rate among older patients receiving remission induction therapy can be as high as 30 percent because of increased hematologic and nonhematologic toxicities (e.g., hepatotoxicity and cardiotoxicity).^111,112 This poor tolerance for chemotherapy and consequent reduction of dose intensity largely account for the poor clinical outcome in older patients.

Patients previously treated for ALL are best monitored in dedicated cancer survivorship clinics where both early and late complications can be identified and managed. Table 91–6 summarizes common side effects associated with antileukemia therapy. Hyperglycemia develops in 10 to 20 percent of children during induction therapy with prednisone, vincristine, and l-asparaginase but has no long-term consequence nor prognostic implication; in some cases, short-term insulin treatment is required.²⁵⁰ Adolescent and older adult age, obesity, a family history of diabetes mellitus, and Down syndrome are associated with increased susceptibility to hyperglycemia.^250,251 This induction regimen causes a hypercoagulable state leading to cerebral thrombosis, peripheral vein thrombosis, or both, in as many as 5 percent of patients. Cerebral thrombosis should be distinguished from transient ischemic lesions (posterior reversible encephalopathy syndrome) which are associated with acute hypertension and severe constipation.²⁵² These lesions are located at the watershed areas between the major cerebral arteries and generally are reversible. Cerebral thrombosis can be readily distinguished from transient ischemic lesions by magnetic resonance imaging or computed tomography (Fig. 91–5). Occasionally, cerebral thrombosis may not be apparent by diagnostic imaging until a few days after the onset of symptoms and signs. Thrombotic complications (especially in leg veins or inferior vena cava) are also common in adults receiving asparaginase.^{136,137,138,139,140}

Emphasis on the intensive use of vincristine, methotrexate and glucocorticoids has led to an increased frequency of neurotoxicity,^252,253 and of osteonecrosis.²⁵⁴ Many long-term survivors of childhood ALL, especially those who received high cumulative doses of glucocorticoid, methotrexate, or cranial irradiation, have developed severe osteoporosis.^{254,255,256,257} Early identification of bone lesions and the introduction of therapy to prevent fractures is recommended. Treatment with anthracyclines can produce severe cardiomyopathy, especially when anthracyclines are given in high cumulative and peak doses to children, and particularly young girls.^258,259,260 Prolonged infusion did not appear to reduce late cardiotoxicity compared to bolus administration.²⁶¹ The existence of a safe cumulative dose of anthracycline is controversial. Cardiac abnormalities are persistent and progressive years after anthracycline therapy.²⁶² In one study, dexrazoxane prevented or reduced anthracycline-induced cardiotoxicity without interfering with antileukemic activity.^263,264 In current pediatric trials, limited doses of anthracyclines are used, even for high-risk cases, to decrease the risk of subsequent cardiomyopathy. Most regimens for adult ALL restrict cumulative anthracycline dosage to levels associated with less than 5 percent risk of congestive heart failure.

Cranial irradiation has been implicated as the cause of numerous late sequelae in children, including second cancer, neurocognitive deficits, and endocrine abnormalities that can lead to obesity, short stature, precocious puberty, and osteoporosis.^{265,266,267,268,269} In general, these complications are seen in girls more often than in boys and in young children more often than in older children or adults. Long-term followup studies of survivors of childhood ALL reveal a greater than 10 percent cumulative risk of second neoplasms after 30 years of observation, and a higher-than-average mortality rate among patients who had received cranial irradiation.^266,267 Patients who had been irradiated also had a high unemployment rate and, among women, a low marital rate. Many children with profound deficiencies of growth hormone receive hormone replacement therapy, which permits attainment of acceptable final heights without an increased chance of relapse.²⁷⁰

The most devastating complication is the development of brain tumors and therapy-related myeloid leukemia.^271,272 Children who undergo cranial irradiation at age 6 years or younger are most susceptible to development of brain tumors.²⁷³ Intensive use of antimetabolites before and during cranial irradiation also increases the risk of brain tumor.¹⁷⁷ The median latency period for high-grade brain tumors is 9 years; it is 20 years for low-grade tumors (e.g., meningioma).^266,273

Therapy-related myeloid leukemia has been linked to intensive treatment with epipodophyllotoxins (teniposide and etoposide). The risk of disease depends on treatment schedule, concomitant use of other agents (e.g., l-asparaginase, alkylating agents, perhaps antimetabolites), and host pharmacogenetics.^176,274 The long-term survival rate for patients with this complication is very low, even when the patients undergo allogeneic stem cell transplantation. No evidence indicates an increased incidence of cancer or birth defects among the offspring of adult survivors of childhood ALL.^275,276

PROGNOSTIC FACTORS

The cornerstone of the modern therapeutic approach to childhood ALL has been careful assessment of the risk of relapse so that only high-risk or very-high-risk patients are treated with intensive therapy.^277,278 Less-toxic treatments (usually antimetabolites) are reserved for low-risk or standard-risk patients. By contrast, almost all adult patients are candidates for intensive therapy. Of the many variables that influence prognosis, treatment is the most important. Some of the factors that emerged as useful prognostic indicators in the past have disappeared as treatment has improved; others have shown predictive strength in one or several trials, but not in others. For example, T-cell ALL, once associated with a poor prognosis, now has long-term response rates of 70 to 85 percent in children^2,17,190 and 50 to 60 percent in adults^66,67,279 as a result of effective intensive chemotherapy. Outcomes for mature B-cell ALL, also a poor prognostic subset in the past, have improved in both children and adults and 80 percent or more are cured with short but intensive chemotherapy treatments.²⁸⁰

Age and leukocyte count continue to be used for risk classification in almost every pediatric clinical trial involving precursor B-cell ALL. In a workshop sponsored by the National Cancer Institute, participants agreed on a presenting age of between 1 and 9 years and a leukocyte count of less than 50 × 10⁹/L as the minimum criteria for low-risk ALL. These criteria apply only to precursor B-cell ALL and not to T-cell ALL. Among adults, the outcome of therapy worsens with increasing age and leukocyte count. Age younger than 35 years and leukocyte count less than 30 × 10⁹/L are considered favorable prognostic indicators (Table 91–7).^67,150,152 In general, age younger than 60 years is considered a practical guide for selecting candidates who might benefit from intensive therapy, including allogeneic transplantation. Any decision to begin aggressive treatment in patients older than age 60 years must be weighed against the risk of increased morbidity and mortality.^111,112

Male sex has long been recognized as an adverse prognostic factor in childhood ALL but has less influence in adult ALL. Its prognostic significance was abolished in a number of childhood studies in which overall outcome was improved. Black race conferred a poor outcome in the national clinical trials,^281,282 but in a single-institution study with equal access to effective treatment regimens, race had no prognostic significance.²⁸³

Primary genetic abnormalities have important prognostic significance. Hyperdiploidy (>50 chromosomes) and ETV6-RUNX1 fusion—seen primarily in children ages 1 to 9 years but very rarely in adults—are associated with a favorable prognosis.^2,17 MLL rearrangements, which occur in 70 to 80 percent of infants younger than age 1 year and in 10 percent of adults, and Philadelphia chromosome with BCR-ABL1 fusion, which is found in 3 percent of children but in 25 to 30 percent of adult patients, historically confer a poor outcome.^1,2,17,199 Interestingly, there is a marked influence of age on the prognosis of genetic subtypes of ALL. For example, Philadelphia chromosome–positive ALL is associated with a poor outcome in adolescents but a relatively favorable outcome in children ages 1 to 9 years old who have a low leukocyte count at presentation.¹⁹⁸ Once associated with a dismal outcome, the early treatment outcome of Philadelphia chromosome–positive ALL in both children and adults has improved substantially with the advent of BCR/ABL1 tyrosine kinase inhibitors.^{113,114,200,201,210,211,212,213,214,215,216} Among patients with MLL-rearranged ALL, infants younger than age 1 year fare considerably worse than older children.²⁰² The basis of these differences may be related to some combination of secondary genetic events, the developmental stage of the target cell undergoing malignant transformation, and the pharmacogenetics or pharmacokinetic features of the patient. In T-cell ALL, NOTCH1 or FBXW7 mutation identifies a subgroup of childhood or adult cases with a favorable outcome.^284,285,286

A useful adjunct in risk assessment is the response to early treatment, as measured by the rate of clearance of leukemia cells from the blood or marrow with the use of flow cytometric detection of aberrant immunophenotype or analysis by PCR of clonal antigen receptor gene rearrangements.^{83,123,124,125,126,127,128,129,190,287,288} This measure reflects the drug sensitivity or resistance of leukemia cells and the pharmacodynamics of the drugs, which is affected by the pharmacogenetics of the host. Current techniques allow measurement of minimal residual disease in all patients, and it has become the most important prognostic factor (Table 91–8). The expectation is that alteration of treatment intensity according to the level of minimal residual disease at early time points will improve the long-term outcome of patients with ALL. The level of measureable residual leukemia is also a strong predictor of treatment outcome before allogeneic stem cell transplantation for relapsed leukemia.^235,289 Pediatric trials have shown better outcomes when treatment decisions have been based in part on measurable levels of residual disease. However, similar benefit remains to be shown convincingly in adults.