Acute Lymphoblastic Leukemia
Karen R. Rabin
Maria M. Gramatges
Judith F. Margolin
David G. Poplack
INTRODUCTION
Advances in the study and treatment of pediatric acute lymphoblastic leukemia (ALL) are often cited as the paradigm of success in modern, clinical trial, and translational research-based medicine.1 Overall cure rates for pediatric ALL have improved from virtually zero in the 1950s, to current event-free survival (EFS) rates approaching 90% for this disease (Fig. 19.1). The improved overall cure rate is only part of the story, however, since the cure rates for specific ALL subgroups range from greater than 95% to as low as 20% to 30% (Figs. 19.7, 19.9, 19.10, 19.11, 19.13).2 Survival gains are attributable to the development of active chemotherapeutic agents, optimization of individual agent dosing and multiagent combination regimens, and advances in supportive care. Allocation of treatment based on complex integration of host, disease, and response-based prognostic factors has played a critical role as well, ensuring intensification of therapy for patients at high risk of relapse, while sparing toxicity for those with a high probability of cure. Future progress in this field will likely depend on advances in areas such as the recognition and therapeutic targeting of novel molecular abnormalities present in leukemia cells; innovations in development of signaling pathway inhibitors and immunotherapy; and tailoring of therapy on the basis of host pharmacogenetic factors.3,4,5 Although improved cure rates have been gratifying to patients, families, physicians, and researchers, this progress has sometimes come at a significant price in terms of both short- and long-term morbidity. It is hoped that future strategies to “personalize” ALL therapy by identifying individualized, targeted treatments will lead to further improvements in efficacy while minimizing toxicity.
 Figure 19.1 Improvement in survival of children with ALL. Overall survival probability by treatment era for 29,287 children with ALL who enrolled on trials from 1968 to 2005 conducted by the Children’s Oncology Group (COG), Pediatric Oncology Group (POG), and Children’s Cancer Group (CCG). The 2000 to 2005 curve includes CCG, POG, and COG. The 1995 to 1999 curve includes CCG and POG. All other curves are CCG. (Adapted from Hunger SP, Lu X, Devidas M, et al. Improved survival for children and adolescents with acute lymphoblastic leukemia between 1990 and 2005: a report from the children’s oncology group. J Clin Oncol 2012;30:1663-1669.) |
EPIDEMIOLOGY
ALL is the most common malignancy in children. It accounts for one-fourth of all childhood cancers and 72% of all cases of childhood leukemia.6 Approximately 4,900 children are diagnosed
with ALL each year in the United States, with an incidence of 2 to 5 cases per 100,000 U.S. children.7 The peak incidence of ALL occurs between 2 and 5 years of age. This is in marked contrast to acute myeloid leukemia (AML), which does not have a distinct peak age incidence in childhood (Fig. 19.2).
with ALL each year in the United States, with an incidence of 2 to 5 cases per 100,000 U.S. children.7 The peak incidence of ALL occurs between 2 and 5 years of age. This is in marked contrast to acute myeloid leukemia (AML), which does not have a distinct peak age incidence in childhood (Fig. 19.2).
 Figure 19.2 Age-specific incidence of ALL and AML in children and in adults (inset). (From Brown P, Hunger SP. Acute leukemia in children. In: Bope ET, Rakel RE, Kellerman RD, eds. Conn’s current therapy. Philadelphia, PA: Saunders, 2013:765-768, with permission.) |
The incidence of ALL is higher among boys than girls.8 In early studies, male sex was a distinctly poor prognostic factor. At least some of the inferior outcome in boys has been related to the higher incidence of adverse prognostic features, including a higher percentage of T-cell immunophenotype and fewer cases with a favorable DNA index. Although outcomes have improved with modern therapy, boys continue to have a slightly higher incidence and poorer prognosis than girls in most categories of disease.8 Available evidence suggests a link between higher risk for ALL and higher birth weight and maternal history of miscarriage,9,10,11 and lower risk for ALL in children who were breastfed.12
Childhood ALL is characterized by racial and ethnic disparities with regard to both cancer incidence and treatment outcomes (Figs. 19.3 and 19.4). Among children in the United States, ALL is more common in European Americans than in African Americans.13 This disparity appears to be due to the early surge in ALL incidence that occurs in European Americans, which may reflect differences in susceptibility, environmental exposures, or both.14 Among all ethnic subgroups, ALL is most common in those with Hispanic ethnicity.
 Figure 19.3 Kaplan-Meier estimates of EFS for a cohort of 8,447 children with ALL according to racial and ethnic distribution. (Adapted from Bhatia S, Sather HN, Heerema NA, et al. Racial and ethnic differences in survival of children with acute lymphoblastic leukemia. Blood 2002;100:1957-1964, with permission.) |
Genome-wide association studies have identified several germline variants in genes implicated in ALL pathogenesis,15,16,17,18,19 some of which are significantly more common in Hispanics or African Americans, which may explain the increased incidence of ALL observed in these groups. For Hispanics, this may be due, in part, to an effect from Native American ancestry, also associated with higher risk for ALL and for known ALL risk alleles (see Genetics section).20 Further research is needed to better understand racial and ethnic differences for genetic predisposition to ALL, as well as regarding genetic differences in drug metabolism and bioavailability. Racial and ethnic differences in ALL outcomes are detailed in the Prognosis Factors section.
 Figure 19.4 Genetic ancestry and risk of relapse in childhood ALL. Higher levels of Native American (NA) ancestry were linked to increased risk of relapse in patients with native American ancestry who did not receive delayed intensification (A) but not in those who did receive delayed intensification in the Children’s Oncology Group P9904/9905 trial (B). Red = NA ancestry <10%, blue = NA ancestry >10%. (Adapted from Yang JJ, Cheng C, Devidas M. Ancestry and pharmacogenomics of relapse in acute lymphoblastic leukemia. Nat Genet 2011;43:237-241, with permission.) |
GENETICS
ALL is proposed to arise from a combination of ALL susceptibility loci and subsequent somatic mutations in genes critical to lymphoid cell development.21 Evidence for this is based on several observations including (a) the demonstration of both random and nonrandom karyotypic abnormalities in the leukemic cells of the majority children with ALL (Table 19.1), (b) the association between various constitutional chromosomal abnormalities and childhood ALL, (c) the occurrence of familial leukemia, (d) the high incidence of leukemia in identical twins, and (e) the molecular epidemiologic evidence of the importance of various alleles in specific genes (see also Pharmacogenetics section).2 As discussed above, genome-wide association studies have recently identified several germline single-nucleotide polymorphisms (SNPs) associated with a significantly higher risk of developing ALL that also occur in genes involved in B-cell transcriptional regulation and differentiation.15,16,17 The affected genes include ARID5B and IKZF1. The ARID5B variant SNPs were specifically associated with development of the hyperdiploid subtype of ALL and also with increased accumulation of methotrexate polyglutamates, which together may contribute to improved response to therapy.15,16
Certain constitutional chromosomal abnormalities are associated with an increased incidence of childhood leukemia. Children with trisomy 21 (Down syndrome, DS) are 10 to 20 times more likely to develop leukemia than children without DS.22 Although both ALL and AML occur with increased frequency, ALL predominates in all but the neonatal age group.23 The common cytogenetic changes occurring in childhood ALL are less common in DS-ALL.24,25,26 Several novel somatic genetic alterations, however, have been recently identified to occur at higher frequency in DS-ALL: Janus kinase 2 (JAK2) activating mutations are found in 20% of DS-ALL, and translocations and interstitial deletions leading to high expression of cytokine receptor-like factor 2 (CRLF2) occur in 50% of DS-ALL.27,28,29 Clinical outcomes for DS patients have generally been poorer than for non-DS patients, primarily due to an increased risk of relapse, and also due to an increased risk of treatment-related mortality.24 Outcomes have improved with enhanced supportive care and a better understanding of the characteristic toxicities observed in DS patients, which include more frequent mucositis (partly attributable to increased sensitivity of DS patients to methotrexate), hyperglycemia, and infectious complications.25
Although other less common preexisting chromosomal abnormalities and specific inherited syndromes have been linked to leukemia, less than 5% of ALL cases (including those in patients with DS) can be linked to any specific genetic cause.2 Included among these are reports of ALL in children with germline BRCA2 mutations, neurofibromatosis, and those with Shwachman syndrome.30 An increased incidence of acute leukemia among those with Bloom syndrome, Fanconi anemia, Nijmegen breakage syndrome, and ataxia telangiectasia (AT) is well documented.30 These rare, autosomal recessive disorders are characterized by increased chromosomal fragility. T-ALL is more common in AT, such that a diagnosis of AT should be considered in children presenting with
this type of ALL and neurologic abnormalities.30 Although AML is more common in Bloom syndrome as well as in patients with Fanconi anemia, ALL may occur.
this type of ALL and neurologic abnormalities.30 Although AML is more common in Bloom syndrome as well as in patients with Fanconi anemia, ALL may occur.
TABLE 19.1 Selected Recurrent Genetic Alterations in Childhood Acute Lymphoblastic Leukemia (ALL) | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Multiple cases of leukemia within families have been reported, including aggregates among siblings and groups within the same generation or in several generations. Though ALL was not historically viewed as a cancer associated with Li-Fraumeni syndrome, recent genome-wide sequencing of hypodiploid ALL revealed germline TP53 mutations in numerous cases (see also Ploidy section), and an inherited TP53 mutation was also reported in a familial leukemia kindred.31,32 This finding suggests that ALL should be included as one of the cancers associated with Li-Fraumeni syndrome. In addition, mutations in the PAX5 gene, a key transcription factor required for normal hematopoiesis, have been associated with genetic predisposition to B-ALL in familial leukemia kindreds.33 With the increasing availability of whole-genome or whole-exome sequencing, mutations in other genes critical to hematopoiesis are likely to emerge as additional ALL predisposition factors.
The importance of in utero genetic events has been suspected for many years because of concordance studies on twins with leukemia.2,34 Leukemogenic translocations (e.g., t[4;11] and t[12;21]), hyperdiploidy, and other markers of clonality (e.g., evidence of the same specific immunoglobulin variable, D, and joining regions [VDJ] and T-cell receptor [TCR] recombination events) that are found in individual patients’ leukemic clones at diagnosis can be detected at birth as minor hematopoietic populations.2,35 Studies of stored cord blood and neonatal heel-stick (Guthrie) cards have shown that as many as 1 in 100 to 1 in 1,000 newborns have either preleukemic translocations and/or evidence of VDJ recombinant clones present at the level of 10-2 to 10-4 at the time of birth.2,35 Clearly, the majority of these children do not develop clinical leukemia, but this is evidence that important initiating events contributing to leukemogenesis may often occur in utero.
Current research in this area utilizes the newest tools of high-throughput sequencing, as well as mRNA and microRNA microarrays, DNA copy number alterations and loss of heterozygosity (LOH), and epigenetic changes (e.g., methylation) to uncover cooperating changes required for the preleukemic clones identified in the cord blood samples to become frankly leukemic.4 These studies show promise as tools not only for elucidating the biology of leukemogenesis and potentially identifying children at greatest risk for developing leukemia, but also for the development of new biomarkers and therapeutic targets for the next generation of leukemia therapies.4
The frequency of leukemia is higher than expected in families of leukemia patients.36 Siblings of children with leukemia, including ALL, have an approximately two- to fourfold greater risk of developing the disease than do unrelated children in the general population.36 Although the occurrence of leukemia in identical twins has been used to support the role of in utero and genetic factors in the disease, the extent to which this association implicates a genetic susceptibility is ambiguous. The concordance of acute leukemia in monozygotic twins is estimated to be as high as 25%, and although it is often the result of shared in utero circulation (since by molecular fingerprinting techniques the leukemic clone can be shown to be identical in the two twins), both patients share similar exposures to prenatal or postnatal leukemogenic factors.37 The risk for leukemia concordance among twins (both mono- and dizygotic) is highest in infancy, and this risk diminishes with age. After the age of 7 years, the risk to the unaffected twin is similar to that of persons within the general population.37
PATHOGENESIS
In addition to genetic influences, environmental factors, viral infection, and immunodeficiency may predispose children to leukemia.
Environmental Factors
ALL is relatively rare in North Africa and the Middle East where non-Hodgkin lymphoma (NHL) is the most common childhood malignancy. In India and China, ALL is somewhat more common, but its incidence is still considerably less than in the industrialized West. This geographic variation may reflect, in part, an underdiagnosis or underreporting of some or all forms of ALL in developing countries, where barriers in access to medical care may prohibit diagnosis before death from the disease.38 However, it is also possible that children in industrialized countries are uniquely exposed to environmental leukemogens that may influence the incidence of ALL.
The young age peak in ALL has appeared historically at different eras in different countries. It occurred initially in Great Britain in the 1920s, in the United States in the 1940s, and in Japan in the 1960s. The appearance of these peaks corresponds to major periods of industrialization in these countries, suggesting a reflection of periods of exposure to new environmental leukemogens.39 In support of these observations, the risk for childhood ALL has been associated with third trimester in utero exposure to polycyclic aromatic hydrocarbons, arsenic, benzene, and other compounds related to fuel combustion.40
Chronic chemical exposure (e.g., benzene) has been strongly associated with the development of AML in adults. In contrast, direct evidence linking specific chemical exposure to the development of childhood ALL has been difficult to establish and effect sizes have been notably smaller.41 Pre- and postnatal parental smoking is associated with an increased risk for ALL,42 a risk that may in part be related to DNA damage and carcinogens found in sperm. There also appears to be some association with postnatal exposure to household paint, polychlorinated biphenyls (PCBs), and professional pest control chemicals.43,44,45 Whether, and to what degree, any of these exposures contribute to the incidence of childhood ALL is controversial. The evidence for this is usually in the form of a mild increase in the odds ratio for developing the disease. To date, it appears unlikely that the bulk of ALL is due to any of these individual exposures.
Ionizing radiation represents one of the few, if not only, established causal exposures for childhood leukemia. The high incidence of leukemia in survivors of the atomic bomb explosions in Japan during World War II is well documented. The risk of leukemia was dose related and greatest for those closest to the explosion. For persons who received exposure doses greater than 100 cGy, the dose-response relationship to the incidence of leukemia was linear, with the type of leukemia observed related to the age at exposure, ALL in children, and AML in adults.
Among survivors of the atomic bomb, there was no increase in the incidence of leukemia in children exposed to radiation in utero. This experience contrasts with other reports of an increased risk of leukemia in children exposed to diagnostic irradiation in utero, particularly during the first trimester. However, changes in obstetrical practices (limiting x-ray exposure because of awareness of potential organ toxicity and carcinogenesis), coupled with improvements in x-ray methodologies and availability of alternate imaging technologies, have now significantly removed radiation or lowered radiation doses from routine prenatal procedures, such that these exposures are no longer considered to be significant contributors to leukemia risk.
Therapeutic irradiation for conditions such as neonatal thymic enlargement and tinea capitis (no longer recommended), and ankylosing spondylitis has been associated with a higher risk of acute leukemia. However, uncertainties surround the percentage of leukemia cases directly attributable to natural background radiation. The largest case-control study to date investigating ALL risk and background radiation exposure found that risk for ALL increased with increasing cumulative doses of gamma radiation,46 though this hypothesis remains somewhat controversial and requires confirmation by replication in other large populations. Concerns regarding the possibility that exposure to electromagnetic
fields (EMFs), routine emissions from nuclear power plants, or fallout from atmospheric nuclear testing may be causally related to the development of childhood ALL have not withstood close scientific scrutiny.
fields (EMFs), routine emissions from nuclear power plants, or fallout from atmospheric nuclear testing may be causally related to the development of childhood ALL have not withstood close scientific scrutiny.
Viral Infection
There has been intense interest in the possible role played by viral infection in the pathogenesis of human leukemia. This has been due, in large part, to the above-described peak age of disease onset between 2 and 5 years observed in industrialized countries, corresponding to a time when the immune system is developing and perhaps more vulnerable to the oncogenic effects of particular viruses. Two theories have been proposed to explain this observation: the delayed-infection hypothesis and the population-mixing hypothesis.47 Both hypotheses suggest that leukemia may occur as an aberrant immune response to common infections in genetically susceptible individuals. Greaves hypothesized that infants and children insulated from infectious exposures from an early age might be more “immunologically naïve” and thus predisposed to a pathological immune system response to infection, precipitating malignant transformation.47 In support of this hypothesis, there is now substantial evidence that early childhood daycare attendance is protective against development of childhood ALL48,49 and that the degree of protection may be proportional to the number and frequency of social contacts.48 Kinlen suggested that the migration of individuals between developing and industrialized societies (population mixing) may result in unprecedented infectious exposures in previously isolated populations, thus similarly resulting in an aberrant immune response predisposing to malignant transformation. Both theories conclude that the development of ALL may occur as an unusual response to any number of common infectious exposures occurring in childhood.
Specific viruses have been definitively linked to certain types of ALL. The Epstein-Barr virus (EBV) has been linked to cases of endemic Burkitt lymphoma, the L3 morphologic subtype of ALL, and to some cases of Hodgkin lymphoma (HL). The EBV association with ALL is discussed further in the subsequent section on molecular genetics, and the association of EBV with other cancers is discussed in Chapters 1, 22 to 24, and 37. The human lymphotropic viruses I and II (HTLV I and II) are retroviruses that are implicated in some cases of adult T-cell and hairy cell leukemia. Cases of childhood malignancies have been linked to human immunodeficiency virus (HIV) infection, but the spectrum of histologies is different from those seen in adult acquired immunodeficiency syndrome (AIDS) patients and does not usually include ALL.
Immunodeficiency
Children with various congenital immunodeficiency diseases, including Wiskott-Aldrich syndrome, congenital hypogammaglobulinemia, and AT (see the Epidemiology and Genetics sections), have an increased risk of developing lymphoid malignancies, as do patients receiving chronic treatment with immunosuppressive drugs. These are usually lymphomas with mature B-cell phenotypes. Although ALL may occur in these circumstances, it is uncommon.
Abnormalities of the immune system are frequently observed in newly diagnosed patients with ALL. For example, 30% of patients present with abnormally low serum immunoglobulin levels. Whether such abnormalities precede the development of leukemia or are a consequence of the disease is unclear. Similarly, abnormalities of the immune system may persist after therapy, although the effects of therapy versus effect of leukemia may be difficult to discern. In addition to infectious risks or complications, it is conceivable that aberrant immune status may not only be a factor in leukemogenesis, but also be an important influence on prognosis (see Response to Treatment section), susceptibility to relapse, and/or the development of second malignancies after completion of therapy.50
Clonal Pathogenesis
ALL, like other lymphoid malignancies, is believed to develop as a consequence of malignant transformation of a single abnormal progenitor cell that has the capability to expand (in a so-called clone of similar progeny cells) by indefinite self-renewal. It is not entirely clear where in the normal course of differentiation the leukemic “clonal event” occurs, and it may actually be highly variable. In pediatric ALL there is evidence that these events occur in committed lymphoid precursors, whereas in AML and Philadelphia chromosome-positive (Ph+) ALL (see section on recurrent translocations in B-ALL), it appears that they may occur in a less differentiated, more primitive hematopoietic cell, due to evidence for mutations in multiple cell lineages.51 The events that lead to malignant transformation are complex and multifactorial. It has been proposed that ALL results from spontaneous mutation(s), which may occur in lymphoid cells of B- or T-cell lineage or in their precursor stem cells. As noted previously, the causative mutations may occur in a small preleukemic clone years before the presentation of clinical leukemia. During normal lymphoid development, lymphocyte precursors may be at higher risk for spontaneous mutation. The risk of mutation is specifically increased during the process of immunoglobulin and TCR gene rearrangements (when DNA strands are being frequently broken and reannealed in new conFigurations). The risk is further compounded by the fact that there is often an increased rate of cellular proliferation/immune-based cell expansion in these very same cell populations. Many of the described molecular mutations bear evidence of IgG and TCR recombinase activity.
Other support for the clonal expansion theory comes from classic studies of glucose-6-phosphate dehydrogenase isotypes, cytogenetic profiles, and molecular characterizations.8 Rearrangement of immunoglobulin heavy chains (IgH) and TCR genes also has been studied as a marker of clonality in B-lineage ALL. In most cases, identical patterns of IgH and TCR gene rearrangement are observed in leukemic cells obtained at diagnosis and relapse and may reflect origins from the same ancestral clone.8 Deep-sequencing at the IgH locus has revealed an unexpectedly large number of low-frequency ALL subclones from diagnostic samples, in some cases numbering in the thousands, representing a marked degree of clonal heterogeneity.52 Such minor clones may reappear as the predominant clone at time of relapse,53 suggesting a previously unappreciated complexity to ALL as a heterogenous composition of numerous major and minor clones that may be differentially responsive to various chemotherapeutic agents (see also Molecular Alterations at Diagnosis and Relapse section).
The concept of cancer stem cells has been an increasing focus of research over the past decade. Cancer stem cells have been reported in a variety of solid tumors and in AML. The data for cancer stem cells in ALL is less well-developed but constitutes an active area of current research.51 Key features of leukemia stem cells, also termed leukemia initiating cells (LICs), have variously been defined as including the capacity for self-renewal, multipotential differentiation, relatively quiescent cell-cycle status, rarity within the total tumor cell population, a distinctive immunophenotype, and ability to initiate tumors in immunocompromised mice. Some data suggest that the immunophenotype and differentiation stage of LICs differ between different cytogenetically defined ALL subtypes.51
Conventional chemotherapy, which targets rapidly dividing cells, may not eradicate LICs, which are, by definition, capable of prolonged quiescence. Thus, leukemia stem cell research holds promise as a novel therapeutic approach; yet it also engenders significant challenges: LICs may have changeable and/or patient-specific immunophenotypic markers, and therapeutic responses may be difficult to assess as LICs do not make up the bulk of the leukemic cell burden. It has been suggested that the long-established maintenance backbone of ALL therapy consisting of methotrexate and 6-mercaptopurine (6-MP) may in fact constitute a form of stem cell therapy that already has been in use for decades.54
For many years, it has been assumed that cure of ALL necessitated the killing of all leukemic cells. This may not be true because multiple minimal residual disease (MRD) technologies occasionally show evidence of persistent viable leukemic cells late in therapy and even in off-therapy patients who do not subsequently clinically relapse. Whether these cells truly represent clonal cells identical to those at leukemic diagnosis, versus a vestige of a preleukemic clone, is controversial. Another hypothesis suggests these cells as being those of the original clone, and the reason that overt leukemia does not arise may be related to the fact that the patient’s own immune system (or an immune system that has been replaced by a bone marrow or stem cell transplant) has learned to control the proliferation of the remaining leukemic cells.55
CLASSIFICATION OF ALL
Morphologic, biochemical, and immunophenotypic characterizations of leukemic lymphoblasts have confirmed that ALL is a biologically heterogeneous disorder (see also Chapter 7). This heterogeneity reflects both the fact that different leukemias may arise from lymphoid progenitors at varied stages of differentiation and that leukemias often bear markers associated with more than one stage of normal development. The majority of pediatric ALL cases (80% to 85%) express markers that indicate origin from an early B-cell progenitor. These cases were formerly referred to as B-precursor ALL, and under the current WHO 2008 criteria they are termed B-ALL.56 T-lymphoblastic leukemia, originating from an early T-lineage progenitor, makes up approximately 15% of pediatric ALL, and mature B-cell or Burkitt leukemia, which is currently classified as a mature B-cell leukemia/lymphoma rather than a type of ALL, accounts for approximately 2% of cases.
Morphologic and Biochemical Classification
Historically, classification of ALL relied on morphologic criteria such as cell size, nuclear-to-cytoplasmic ratio, nuclear shape, number and prominence of nucleoli, nature and intensity of cytoplasmic staining, presence of cytoplasmic granules, prominence of cytoplasmic vacuoles, and the character of nuclear chromatin.57 Generally, these classification systems were unsuccessful because they were subjective, technically difficult to reproduce, and often lacked meaningful clinical correlations. Ultimately, morphologic classification has been replaced by immunophenotyping. However, prior to the advent of immunophenotyping, a system proposed by the French-American-British (FAB) Cooperative Working Group was accepted as the standard morphologic classification.
The FAB system defines three categories of lymphoblasts. L1 lymphoblasts are usually smaller, with scant cytoplasm and inconspicuous nucleoli. Cells of the L2 variety are larger and demonstrate considerable heterogeneity in size, prominent nucleoli, and more abundant cytoplasm. Lymphoblasts of the L3 type, notable for their deep cytoplasmic basophilia, are large, frequently display prominent cytoplasmic vacuolation, and are morphologically identical to Burkitt lymphoma cells. The L3 type correlates with mature B-cell leukemia as defined by immunophenotype, but the FAB L1 and L2 morphologic types do not correlate with immunophenotypic, cytogenetic, or microarray-based classification information (see later), and this level of classification is largely of historical interest.
The most important morphologic distinction is between ALL and AML, which may be useful in guiding initial urgent clinical management while awaiting a definitive diagnosis by immunophenotyping. Routine Wright-Giemsa staining is usually adequate for this task. Auer rods, if present, provide definitive evidence of myeloid lineage. However, no other features unequivocally distinguish between ALL and AML, and there are no morphologic features that reliably distinguish between B- and T-lineage ALL.
Cytochemical stains played an important role in distinguishing leukemia subtypes prior to immunophenotyping, but are no longer in widespread use. They remain useful in occasional challenging cases where the lineage is unclear or highly undifferentiated. Myeloperoxidase is particularly useful for identification of myeloid lineage. Terminal deoxynucleotidyl transferase (TdT) is a DNA-polymerizing enzyme that is useful for identification of mature B-cell leukemia versus a B- or T-lymphoblastic leukemia.
Immunophenotypic Classification
Immunophenotyping is now the standard accepted method for diagnosis of ALL (see also Chapter 7), due to advances in development of monoclonal antibodies, improvements in the enzymatic and fluorescent tagging of these antibodies, and the development of multiparameter FACS instrument technology.58 Leukemic transformation and clonal expansion can occur at different stages of lymphoid differentiation, and multicolor flow cytometry can be used to co-stain a sample with panels of antibodies to define clonal populations on the basis of marker co-expression. Hundreds of monoclonal antibodies are commercially available with specificity for antigens associated with the different hematopoietic lineages. Since the antigens used for routine clinical immunophenotyping are lineage-associated rather than absolutely lineage specific, a panel of multiple antibodies should be used. Those most helpful in the immunologic classification of ALL are shown in Table 19.2. The choice of the diagnostic panel may vary among institutions and laboratories, but for the characterization of lymphoid leukemias, generally includes antibodies for several T-lineage (e.g., CD2, 3, 4, 5, 8, and/or 7), B-lineage (e.g., CD10, 19, 20, 22, kappa, and lambda), and myeloid (e.g., CD13, 33, 64, and 117) antigens (also see Chapter 7 for more detailed discussion).
Using a panel of monoclonal antibodies associated with various stages of B-cell differentiation, along with information on the presence or absence of cytoplasmic and surface immunoglobulin, investigators have classified B-lineage ALL into discrete stages according to the degree of differentiation or maturation. However, leukemias often express antigens that are not synchronous with any of the normal stages of differentiation, and/or antigens associated with more than one lineage, so such assignment to an immunologic stage is necessarily approximate and not clinically useful. Although once thought to be an unfavorable prognostic sign, the presence of some myeloid antigen marker positivity in cells that predominantly mark as lymphoid has no prognostic or therapeutic implications.8 Aberrant expression of the myeloid antigens CD13 and CD33, for example, is common in B-ALL. Myeloperoxidase expression, in contrast, is not seen in B-ALL and is indicative of AML or mixed phenotype acute leukemia (see below). Several recurrent genetic abnormalities are associated with specific immunophenotypic patterns. MLL-rearranged ALL is often CD10–/CD15+; Ph+ ALL is often CD25+ and often expresses myeloid markers CD13 and/or CD33; ETV6-RUNX1+ ALL is often CD19+/CD10+/CD34+/CD9– with expression of myeloid marker CD13; and TCF3-PBX1+ ALL is often CD34–/CD9+/cytoplasmic mu chain. Infant ALL appears to arise from a very early multipotent hematopoietic precursor, lacking expression of CALLA (CD10) and mature B-cell antigens, and often expressing HLA-DR, and myeloid markers such as CD15. Early T-cell precursor ALL is a recently identified subgroup of T-ALL that is defined by its immunophenotypic features (see section on Molecular Alterations in T-ALL).59
Mixed phenotype acute leukemia (MPAL) is a diagnosis established by the WHO 2008 classification, replacing the former diagnosis of bilineal or biphenotypic acute leukemia by the European Group for the Immunological Classification of Leukemias (EGIL) and the WHO 2001 classification.56,60 MPAL comprises approximately 0.5% to 1% of acute leukemias, with combined lineage differentiation that is most often B- and myeloid, followed by T- and myeloid, and rarely B- and T-cell, trilineage, or natural killer cell
combinations. The WHO classification recognizes two specific genetic subgroups, characterized by BCR-ABL or MLL rearrangement, and a third subgroup for cases with any other chromosomal abnormalities, subclassified by the lineages present. There is little systematic data about this rare form of leukemia, but outcomes are generally poor (see also Treatment section).61
combinations. The WHO classification recognizes two specific genetic subgroups, characterized by BCR-ABL or MLL rearrangement, and a third subgroup for cases with any other chromosomal abnormalities, subclassified by the lineages present. There is little systematic data about this rare form of leukemia, but outcomes are generally poor (see also Treatment section).61
TABLE 19.2 Monoclonal Antibodies Commonly Used to Immunophenotype Leukemia | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
MOLECULAR GENETICS OF ALL
Genetic abnormalities are central to understanding of the biology and treatment of ALL.19,62 The molecular underpinnings of leukemogenesis are related to the alteration of key regulatory processes that control hematopoietic proliferation (self-renewal), differentiation, and apoptosis (the molecular signals that lead a cell to death). This can occur through a variety of mechanisms including alterations in cell signaling pathways, with mutations that affect the activity or expression of specific kinases and other proteins, aberrant expression of proto-oncogenes or silencing of suppressor genes, and the expression of chimeric transcription factors encoded by chromosomal translocations.8 Multiple recurrent chromosomal alterations linked to altered regulation and function of cellular oncogenes occur in pediatric ALL and are thought to play a pivotal role in leukemogenesis. The most significant in terms of frequency and/or prognostic significance are listed in Table 19.1 and Figure 19.5.
Cytogenetic analysis by karyotype and fluorescent in-situ hybridization (FISH) is an essential element in the modern diagnostic evaluation of ALL. Other newer technologies include array-based analysis of global gene expression, DNA copy number, SNPs, methylation, and genome-wide sequencing (Fig. 19.6). DNA copy number arrays have rapidly emerged as a powerful tool for genome-wide investigation of areas of gain and loss of genetic material.63 The benefits of these arrays include significantly higher resolution than routine karyotype; genome-wide scope in contrast to the targeting of select lesions by FISH; and lack of requirement for in vitro culture prior to analysis. Array CGH detects only DNA copy number alterations, whereas SNP arrays also detect copyneutral LOH, also known as uniparental disomy. Genome-wide sequencing has advanced rapidly and includes whole-genome sequencing, transcriptome sequencing (RNA-seq), and whole-exome sequencing.64 These genomic techniques were originally utilized as research discovery tools, but increasingly they are becoming established for clinical applications as well. These new technologies may yield powerful insights into disease biology, although significant challenges remain regarding test characteristics, validation, and interpretation.65
The genetic alterations reported in ALL involve both chromosomal number (ploidy) and structural rearrangements. The following sections discuss the molecular and clinical features and the prognostic impact of specific alterations. Therapeutic implications are discussed subsequently in the Treatment section.
Ploidy
Ploidy can be determined directly by the classic method of counting the modal number of chromosomes in a metaphase karyotype preparation or by an alternative indirect method of measuring DNA content by flow cytometry. The DNA content by flow cytometry is reported as the DNA index (DI), which is a ratio between the amount of fluorescence seen in a normal diploid cell and the fluorescent content of the bone marrow blasts (in G0/G1) at diagnosis. Normal diploid or pseudodiploid cells (cytogenetically abnormal but having a normal DNA content) have a DI of 1.0. Hyperdiploidy represents a chromosome number greater than 46 and DI greater than 1.0, and hypodiploidy represents a chromosome number less than 46 with a DI less than 1.0. Hyperdiploidy is further classified as “low” (47 to 50) and “high” (>50). Cases with more than 67 chromosomes are generally considered to be
near-tetraploid, a distinct category in terms of chromosomal doubling pattern.66 Near-tetraploidy is often associated with T-ALL or ETV6-RUNX2+ B-ALL. Early studies suggested a poor outcome, but more recent data suggest a favorable prognosis for near-tetraploid ALL.67 Cases in the pseudodiploid category (those with a DI = 1.0 or normal chromosome number but other abnormalities) have a relatively poorer prognosis. Those with diploidy and low hyperdiploidy (48 to 50 chromosomes) have a slightly worse prognosis than high hyperdiploid cases. In childhood ALL, approximately 20% of cases are high hyperdiploid and 1% are hypodiploid (Fig. 19.5). In T-ALL, the majority of cases are pseudodiploid or diploid.
near-tetraploid, a distinct category in terms of chromosomal doubling pattern.66 Near-tetraploidy is often associated with T-ALL or ETV6-RUNX2+ B-ALL. Early studies suggested a poor outcome, but more recent data suggest a favorable prognosis for near-tetraploid ALL.67 Cases in the pseudodiploid category (those with a DI = 1.0 or normal chromosome number but other abnormalities) have a relatively poorer prognosis. Those with diploidy and low hyperdiploidy (48 to 50 chromosomes) have a slightly worse prognosis than high hyperdiploid cases. In childhood ALL, approximately 20% of cases are high hyperdiploid and 1% are hypodiploid (Fig. 19.5). In T-ALL, the majority of cases are pseudodiploid or diploid.
 Figure 19.5 Cytogenetic and molecular genetic abnormalities in childhood ALL. Acute lymphoblastic leukemia with rearrangement of CRLF2 but without the BCR-ABL1-like transcriptional profile rarely presents with other classifying karyotypic alterations, but can be noted with high hyperdiploidy. Dicentric cases might have a range of translocations, including classifying translocations (e.g., ETV6-RUNX1). iAMP21, intrachromosomal amplification of chromosome 21; ETP, early T-cell precursor. (From Inaba H, Greaves M, Mullighan CG. Acute lymphoblastic leukaemia. Lancet 2013;381:1943-1955, with permission.) |
High hyperdiploidy is associated with a favorable prognosis. The chromosomes gained are nonrandom, with a distinct pattern of chromosomes gained at each modal number of chromosomes.66 Chromosome 21 is the most commonly gained, followed by X, 6, 14, 4, 18, 17, and 10.68 The molecular basis for hyperdiploid ALL remains unclear, but evidence suggests that the most common mechanism is via simultaneous gain of all extra chromosomes in a single abnormal cell division.69 Hyperdiploidy has long been recognized as a favorable prognostic finding in B-ALL (Fig. 19.7). Definitions of hyperdiploidy vary between studies, with some based on modal number of chromosomes determined by karyotype, some based on DNA index, and others based on presence of specific trisomies (e.g., trisomy 18 for the Medical Research Council in the United Kingdom; trisomies 10 and 17 for the Children’s Cancer Group; “double trisomy” of chromosomes 4 and 10 for the Pediatric Oncology Group and current COG trials; and “triple trisomies” of chromosomes 4, 10, and 17 in earlier COG trials). However it is defined, hyperdiploidy has been found to be an independent favorable prognostic factor on numerous cooperative group trials, with independent favorable prognostic significance preserved from the 1990s to present regimens.69 Cases with higher modal chromosome number (over 57 to 58) have consistently shown the most favorable outcomes.70 Patients with hyperdiploid ALL usually also share a number of other important favorable prognostic features (see Prognostic Factors section), including a favorable age, low initial leukocyte count, and a B-ALL phenotype often displaying CD10, or common acute lymphoblastic leukemia antigen (CALLA) positivity.
Trisomy 8 is prognostically neutral but warrants some specific comments. Trisomy 8 is the most common chromosomal numerical abnormality seen in AML, but occurs rarely in ALL and, when present, is associated with T-cell immunophenotype. When chromosome 8 of leukemic lymphoblasts (i.e., those with trisomies and those without) is examined by FISH, t(8:14)(q24:q32) translocations or duplications of the same 8q24 band may be identified. Chromosome 8q24 is the location of the c-myc gene, which is important for cell growth and the site of many leukemia-related translocations (see Table 19.1). The finding of trisomy 8 in leukemic lymphoblasts should prompt a reexamination of the available data to be sure that by morphology, immunophenotype, and karyotype, the patient has B-ALL and not mature B-cell ALL, T-ALL, or even AML.
The worst prognosis (by ploidy) occurs with hypodiploidy (chromosome number <45 and DI <1.0). Cases are generally
stratified by modal chromosome number into near haploid (24 to 31), low hypodiploid (32 to 39), and high hypodiploid (40 to 43) (Fig. 19.8).31 Cases with any chromosome number <44 have a dismal EFS of approximately 30%, while cases with 44 chromosomes have a somewhat better EFS of approximately 50% (Fig. 19.9).71 Hypodiploidy with 45 chromosomes (frequently associated with dicentric chromosomes 9p and/or 12p) has a relatively good prognosis.72 TP53 mutations have been identified in 91% of low-hypodiploid ALL cases, with nearly half of these mutations occurring in the germline as well (Fig. 19.8C, D).31 Thus, identification of low hypodiploid ALL should prompt investigation of the patient’s tumor and germline TP53 status and family history for cancers characteristic of Li-Fraumeni syndrome. Special mention should be made of cases that undergo doubling of the hypodiploid clone. In some cases, both clones are present, and in others, referred to as “masked hypodiploidy,” only the doubled clone is retained. Since masked hypodiploidy does not appear to differ from nonmasked hypodiploidy in clinical or prognostic
features, it is crucial to distinguish it from hyperdiploidy, which in contrast bears a favorable prognosis and might prompt de-escalation of therapy.31,73 Suspicion may be raised by a predominance of tetrasomies affecting characteristic chromosomes (most often 8, 10, 14, 18, and 21). The diagnosis may be made by microsatellite panel or SNP array, which demonstrate LOH for the disomic chromosomes of the doubled clone.
stratified by modal chromosome number into near haploid (24 to 31), low hypodiploid (32 to 39), and high hypodiploid (40 to 43) (Fig. 19.8).31 Cases with any chromosome number <44 have a dismal EFS of approximately 30%, while cases with 44 chromosomes have a somewhat better EFS of approximately 50% (Fig. 19.9).71 Hypodiploidy with 45 chromosomes (frequently associated with dicentric chromosomes 9p and/or 12p) has a relatively good prognosis.72 TP53 mutations have been identified in 91% of low-hypodiploid ALL cases, with nearly half of these mutations occurring in the germline as well (Fig. 19.8C, D).31 Thus, identification of low hypodiploid ALL should prompt investigation of the patient’s tumor and germline TP53 status and family history for cancers characteristic of Li-Fraumeni syndrome. Special mention should be made of cases that undergo doubling of the hypodiploid clone. In some cases, both clones are present, and in others, referred to as “masked hypodiploidy,” only the doubled clone is retained. Since masked hypodiploidy does not appear to differ from nonmasked hypodiploidy in clinical or prognostic
features, it is crucial to distinguish it from hyperdiploidy, which in contrast bears a favorable prognosis and might prompt de-escalation of therapy.31,73 Suspicion may be raised by a predominance of tetrasomies affecting characteristic chromosomes (most often 8, 10, 14, 18, and 21). The diagnosis may be made by microsatellite panel or SNP array, which demonstrate LOH for the disomic chromosomes of the doubled clone.
 Figure 19.6 Application of new techniques for characterizing ALL. A: DNA copy number profiling in 242 pediatric ALL cases. Each case is represented by a column. Pink represents diploid copy number, white deletion, and red amplification. HD > 50, hyperdiploidy with greater than 50 chromosomes; Ph, BCR-ABL1 positive ALL; Hypo, B-precursor ALL with hypodiploidy. (From Mullighan CG, Goorha S, Radtke I, et al. Genome-wide analysis of genetic alterations in acute lymphoblastic leukemia. Nature 2007;446:758-764, with permission.) B: Whole-genome sequencing (WGS) identifies recurring mutations in T-ALL. Data are shown for 106 T-ALL cases, including 12 cases subjected to WGS (arrowed), and 94 recurrence cases. Cases are grouped by early T-precursor status. Genes identified as novel targets of mutation in T-ALL are labeled in green. (From Zhang J, Ding L, Holmfeldt L, et al. The genetic basis of early T-cell precursor acute lymphoblastic leukemia. Nature 2012;481:157-163, with permission.) C: Gene-expression profiling identifies a novel BCR-ABL1-like subtype of ALL (left column of gray bar indicates BCR-ABL1+ cases, and right column of gray bar indicates BCR-ABL1-like cases). (From Den Boer ML, van Slegtenhorst M, De Menezes RX, et al. A subtype of childhood acute lymphoblastic leukemia with poor treatment outcome: a genome-wide classification study. Lancet Oncol 2009;10:125-134, with permission.) D: Genome-wide association study identifies germline single-nucleotide polymorphisms (SNPs) associated with susceptibility to childhood ALL. The x-axis shows the P values (-log10) for 307,944 germline SNPs. The allele frequency of the SNPs was compared between children with ALL in the discovery cohort (n = 317) and the combined non-ALL control cohort (n = 17,958). The SNP whose allele frequency differed most significantly between the two cohorts (p = 1.4 × 10-15) was localized to chromosome 10. The dashed line shows the threshold P value indicating genome-wide significance. (From Trevino LR, Yang W, French D, et al. Germline genomic variants associated with childhood acute lymphoblastic leukemia. Nat Genet 2009;41:1001-1005, with permission.) E: Mouse xenografts established by tail vein injection of luciferase-labeled leukemic blasts can be used to test in vivo efficacy of novel therapies. (Courtesy of Alexandra Stevens and Michele Redell, Baylor College of Medicine.) F: Phospho-flow cytometry can be used to demonstrate increased activity of specific signaling pathways in ALL blasts and to identify therapeutically useful agents that inhibit these pathways. |
 Figure 19.7 Kaplan-Meier analysis of event-free survival according to biological subtype of leukemia. (From Pui CH, Robison LL, Look AT. Acute lymphoblastic leukemia. Lancet 2008;371:1030-1043, with permission.) |
Structural Chromosomal Abnormalities
Structural chromosomal abnormalities also occur frequently in ALL. As with numeric abnormalities, they are limited to the leukemic cells, a finding consistent with the clonal nature of the disease. Among the structural abnormalities encountered, translocations are the most common. Generally, chromosomal translocations either juxtapose a proto-oncogene with a TCR or immunoglobulin locus causing its overexpression, or fuse the genes at the translocation breakpoints to form a novel chimeric protein with altered, oncogenic effects. Translocations that are detectable by standard Giemsa banding techniques occur in about 40% of cases, but the number rises to more than 75% with the use of more sensitive testing methods such as FISH, PCR, and array and next-generation sequencing platforms.64
Recurrent Translocations in B-ALL
The most common translocation in childhood ALL, t(12;21)(p13;q22), is associated with a favorable prognosis (Table 19.1, Figs. 19.5 and 19.7).74,75 Approximately 25% of B-ALL patients in the United States have this abnormality, although the prevalence varies widely among ethnic groups.76 In most cases, the translocation is cryptic, detectable by FISH but not by karyotype. It results in fusion of the coding regions of two transcription factors (ETV6 on chromosome 12p13, formerly known as TEL, and RUNX1 on chromosome 21q22, formerly known as AML1), resulting in an oncogenic fusion protein that promotes B-progenitor differentiation and self-renewal. This t(12;21) fusion appears to arise in utero and is thought to be necessary but not sufficient to cause the leukemia.77 The subsequent events that cause progression to true leukemia are not certain. Resistance of t(12;21)+ cells to TGF-β-mediated inhibition of proliferation is one possible mechanism, which may afford a selective proliferative advantage to a t(12;21)+ preleukemic clone in the setting of a dysregulated immune response to infection.78
The t(12;21) translocation occurs more often in children 1 to 10 years of age and in CD10 (CALLA)-positive cases. It is generally regarded as a favorable prognostic factor, although some studies have suggested an increased risk of late relapse.79 The presence of a t(12;21) translocation is not a guarantee of successful treatment, however, as 10% of relapsed cases continue to come from this group. Opinion is divided as to whether ETV6-RUNX1 retains independent prognostic significance or simply tends to occur in cases with other favorable features (e.g., age and response to therapy).80,81 Relapse in these cases tends to occur late, with excellent chemosensitivity and salvage rates.82 Relapse may in fact represent an independent “second hit” in the original preleukemic t(12;21)+ clone.83
 Figure 19.8 Molecular genetics of hypodiploid ALL. A: Genome-wide DNA copy number heatmap showing SNP 6.0 microarray data for pediatric hypodiploid ALL cases. Chromosomes 1-22, X and Y are depicted from top to bottom, and individual samples are shown from left to right. LH, low hypodiploid. Blue indicates deletions and red indicates amplifications. B: Unsupervised principal components analysis (PCA) of gene-expression data from all hypodiploid ALL cases with available high-quality RNA (n = 94) distinguishes near-haploid, low-hypodiploid, and near-diploid ALL subgroups. Right, hierarchical clustering analysis of gene-expression array data. C: Frequent mutations in TP53 in pediatric hypodiploid ALL are indicated on the p53 protein domain plot. Known Li-Fraumeni syndrome (LFS) alterations are indicated in red. Alterations present in nontumor cells are indicated by blue lines. D: Pedigree of a family with an inherited TP53 mutation. N, number of siblings. (From Holmfeldt L, Wei L, Diaz-Flores E, et al. The genomic landscape of hypodiploid acute lymphoblastic leukemia. Nat Genet 2013;43:242-252, with permission.) |
 Figure 19.9 Overall survival in hypodiploid ALL. Overall survival for 130 evaluable, non-Ph+ patients by modal chromosome number: 44 chromosomes, 40-43 chromosomes, 30-39 chromosomes, and 24-29 chromosomes. (Adapted from Nachman JB, Heerema NA, Sather H, et al. Outcome of treatment in children with hypodiploid acute lymphoblastic leukemia. Blood 2007;110:1112-1115, with permission.) |
The t(1;19)(q23;p13) is the second most common chromosomal translocation in childhood ALL, occurring in 6% of cases (Table 19.1, Figs. 19.5 and 19.7).8 This translocation results in the fusion of the transcriptional activation domain of the helix-loop-helix transcription factor TCF3 (formerly known as E2A) on chromosome 19p with the DNA-binding homeodomain of PBX1 located on chromosome 1, band q23. The t(1;19) translocation can occur either as a balanced or unbalanced translocation and does not appear to arise in utero. It was associated with a poor prognosis in the 1990s but no longer has prognostic significance on most current protocols,84,85,86 except for a reportedly increased risk of CNS relapse87 and is therefore no longer used for risk stratification by most cooperative groups.2
Another fusion partner of the TCF3 gene on chromosome 19p13 was originally described as a variant of the t(1;19)(q23;p13) (Table 19.1). The TCF3 fusion partner in this translocation is the hepatic leukemia factor gene (HLF), a transcription factor not usually expressed on hematopoietic cells. The t(17;19)(q22;p13) occurs in 1% of childhood ALL and appears to define a poor-prognostic group of adolescent patients. These patients have an unusual clinical presentation characterized by hypercalcemia, an increased risk of disseminated intravascular coagulation, and a low CD10 positivity pro-B immunophenotype.
The t(8;14)(q24;q32) can be identified in virtually every case of mature B-cell or Burkitt ALL (FAB L3) (Table 19.1). In this translocation, the MYC proto-oncogene, normally located on chromosome 8, is translocated near a transcriptional enhancer of the immunoglobulin heavy chain gene on chromosome 14. The resulting dysregulation of MYC expression is believed to be responsible for the uncontrolled proliferation of B cells characteristic of this disorder. In addition to the translocation of MYC coding sequences, mutations sometimes occur in the translocated sequences. Two variant translocations, the t(2;8)(p11-p12;q24) and t(8;22) (q24;11) involving the κ- and λ-light chains, respectively, are observed less commonly. The similarity in the molecular mechanisms associated with these translocations in mature B-cell ALL and those that occur in Burkitt lymphoma support the presumption that mature B-cell ALL represents a disseminated form of Burkitt lymphoma (also see discussion in Chapter 23).
Alterations of the MLL gene located at chromosome band 11q23 occur in 5% of pediatric ALL, in 70% of infant leukemia (both ALL and AML) in patients less than 1 year of age (also see Chapter 20), as well as in most therapy-related leukemias in patients who received topoisomerase II inhibitors (Table 19.1, Fig. 19.7).88,89 MLL abnormalities are nearly all in-frame fusions of the N terminus of MLL to a fusion partner to create a novel oncogene, although amplification and partial tandem duplication also occur occasionally, primarily in AML. More than 70 partner genes have been identified (see Chapter 3 and Table 19.1), but nine partners account for almost 90% of rearrangements.88,90 The most common translocation in ALL is t(4;11)(q21;q23), fusing MLL with AFF1 (formerly known as AF4), followed by fusion with MLLT1 (formerly known as ENL) and with MLLT3 (formerly known as AF9). The t(4;11) generally occurs in B-ALL and is associated with a characteristic immunophenotype (CD10–/CD15+/CD19+) and frequent co-expression of myeloid markers. MLL-rearranged leukemias show evidence of in utero origin and manifest fewer cooperating mutations than other B-ALL subtypes, suggesting that MLL rearrangement alone is a potent oncogene. These leukemias also exhibit a characteristic stem cell-like gene-expression profile with upregulation of HOXA genes and MEIS1.91 B-ALL patients with MLL rearrangements have significantly poorer treatment outcomes than those of clinically similar patients who do not demonstrate this cytogenetic abnormality, particularly in infant ALL (Fig. 19.10), but in T-ALL there does not appear to be a similar adverse prognostic impact.92
 Figure 19.10 Outcomes in infant ALL. Event-free survival for infants treated on the Interfant-99 trial, by MLL status, age, and initial white blood count. Group 1: patients with germline MLL gene. Group 2: patients with MLL gene rearrangement, age <6 months at diagnosis and initial white blood count >300 109 cells/L. Group 3: all other patients. (From Pieters R, Schrappe M, De Lorenzo P, et al. A treatment protocol for infants younger than 1 year with acute lymphoblastic leukaemia (Interfant-99): an observational study and a multicentre randomised trial. Lancet 2007;370:240-250, with permission.) |
The t(9;22)(q34;q11), one of the first leukemic translocations described, results in the formation of a small marker chromosome known as the Philadelphia chromosome (Ph) and gives rise to the BCR-ABL1 fusion oncogene (Table 19.1, Fig. 19.7). It is the most common translocation in adult ALL, occurring in 25% of cases, but substantially less frequent in pediatric ALL, where it occurs in only approximately 3% of cases. Until recently, it was associated with a dismal prognosis.93 Even with intensive chemotherapy, patients with Ph+ ALL demonstrated an increased risk of induction failure, CNS leukemia, and early relapse. Treatment of Ph+ ALL and CML were revolutionized by development of imatinib mesylate, a selective tyrosine kinase inhibitor active against the BCR-ABL1 fusion, which has improved survival from approximately 30%-40% to 80% (Fig. 19.11).94,95,96,97 Due to these survival gains, hematopoietic stem cell transplant (HSCT) in first remission is no longer considered standard of care for Ph+ ALL.
Most cases of Ph+ ALL are CD10+ B-ALL, although rare cases of T-ALL also occur. Co-expression of myeloid markers is frequent. Patients with Ph+ ALL tend to be older, present with higher leukocyte and peripheral blast counts, have CNS involvement, and historically have exhibited lower induction remission rates, shorter remission durations, and very poor survival. Presence of secondary aberrations in addition to the Ph chromosome may be associated with worse outcomes.96,98
The Philadelphia chromosome is formed by in-frame fusion of the 5′ portion of BCR (for breakpoint cluster region) on chromosome 22 to the 3′ portion of the tyrosine kinase C-ABL1 on chromosome 9, a proto-oncogene that is part of the RAS signaling pathway. The resulting fusion protein upregulates ABL1 tyrosine kinase activity. Two main fusion proteins occur, which differ in BCR breakpoint. Breaks within the 5.8-kb major breakpoint cluster region (M-BCR), occurring in CML and 25% of adult Ph+ ALL, form a 210-kD protein known as p210. In the remainder of adult ALL and the majority of pediatric ALL, the breakpoint occurs further upstream, in the minor breakpoint cluster region (m-BCR),
forming a 185-to-190-kD protein usually known as p190. An additional breakpoint generates a 230-kD protein associated with a rare CML variant with neutrophilia and occasionally with classic CML. All three transcripts can be detected at very low levels using sensitive PCR techniques. It has been suggested that the p190 protein arises de novo, whereas the p210 protein may represent the blast crisis of a previously unrecognized CML. Other features that may distinguish cases that originated as CML include basophilia, marked splenomegaly, and persistence of the BCR-ABL1 fusion protein in hematopoietic precursor cells of all lineages following remission.
forming a 185-to-190-kD protein usually known as p190. An additional breakpoint generates a 230-kD protein associated with a rare CML variant with neutrophilia and occasionally with classic CML. All three transcripts can be detected at very low levels using sensitive PCR techniques. It has been suggested that the p190 protein arises de novo, whereas the p210 protein may represent the blast crisis of a previously unrecognized CML. Other features that may distinguish cases that originated as CML include basophilia, marked splenomegaly, and persistence of the BCR-ABL1 fusion protein in hematopoietic precursor cells of all lineages following remission.
 Figure 19.11 Outcomes in Philadelphia chromosome acute lymphoblastic leukemia (Ph+ ALL). Comparison of disease-free survival in Ph+ ALL patients treated on Children’s Oncology Group study AALL0031 Cohort 5. Cohort 5 patients received continuous dosing of imatinib during intensive chemotherapy and intermittent dosing during maintenance therapy. The curves compare Cohort 5 patients who received chemotherapy only versus related and unrelated bone marrow transplant (BMT). (From Schultz KR, Carroll A, Heerema NA, et al. Long-term follow-up of imatinib in pediatric Philadelphia chromosome-positive acute lymphoblastic leukemia: children’s Oncology Group study AALL0031. Leukemia 2014;28(7):1467-1471, with permission.) |
Several recurrent translocations in ALL involve the immunoglobulin heavy chain locus (also known as IGH@), juxtaposing its enhancer region to transcription factor or cytokine receptor genes, resulting in dysregulated target gene expression (Table 19.1).99,100,101 These comprise approximately 4% of childhood ALL and involve partner genes that include BCL2, the CEBP family, EPOR, ID4, and CRLF2. Overall, the group of patients with IGH@ translocations tend to be older and may have a poorer prognosis, although this finding may be limited to adults.102,103
The most common IGH@ translocation involves CRLF2 and is associated with a poor prognosis. This and other alterations involving CRLF2 are discussed in further detail below. Another, rare IGH@ translocation which is noteworthy for its distinctive clinical presentation is the t(5;14)(q31;q32), which is associated with hypereosinophilia. In this translocation, the immunoglobulin heavy chain enhancer is juxtaposed to IL-3 (or less often IL-5 or granulocyte-stimulating factor) on chromosome 5.104 The eosinophils are reactive, resulting from elevated cytokine production, not malignant. Cases often present with prominent hypereosinophilia but a relatively low percentage of leukemic blasts in the bone marrow. Hence, according to the WHO 2008 classification of ALL, identification of this translocation is sufficient to confirm the diagnosis of ALL even if the bone marrow blast percentage is low.56
Other Recurrent Structural Molecular Alterations in ALL
Deletions of chromosomal band 9p21-22 are common in pediatric ALL, occurring in approximately 20% of B-lineage and 50% of T-ALL (Table 19.1).105 These deletions vary in size but the critical region involves two tumor suppressor genes, the cyclin D kinase inhibitors p16 and p15. Although sometimes reported as an adverse prognostic feature, deletions at this locus do not appear to have independent prognostic significance on current regimens.2,105
Intrachromosomal amplification of chromosome 21 (iAMP21) has been identified in approximately 2% of childhood ALL and is associated with a poor prognosis (Table 19.1, Figs. 19.5 and 19.12A).106 Patients tend to be older, with lower initial white blood cell and platelet counts. Clinical outcomes are poor, with relapse rates ranging from 38% to 61% depending on the treatment regimen.107,108,109 Both the size of the region that is amplified and the number of copies are variable, though by definition there must be at least three extra copies of the RUNX1 locus (Fig. 19.12A). iAMP21 has been incorporated on several current ALL treatment protocols as a factor, which stratifies patients to receive intensified chemotherapy.108,109
Approximately 40% of pediatric B-ALL cases harbor aberrations in genes regulating B lymphocyte development and differentiation (Table 19.1, Fig. 19.6A).63 This was identified in a landmark study by Mullighan et al.,63 in which 242 pediatric ALL samples were subjected to genome-wide analysis using high-resolution SNP arrays and genomic DNA sequencing. Notably, PAX5 alterations were most frequent, affecting 31.7% of cases. Deletions were also detected in other mediators of B-cell pathways including TCF3, EBF1, LEF1, IKZF1, and IKZF3. Most have not been reported to be prognostically significant, except for IKZF1, which encodes the hematopoietic transcription factor IKAROS. Recurrent deletions and inactivating mutations in IKZF1 occur in 15% of ALL cases overall and are associated with a markedly inferior prognosis (Fig. 19.12B).110,111 IKZF1 alterations occur more frequently in high risk and Ph+ ALL, as well as in Ph-like ALL (described below).
Gene-expression profiling has identified a subgroup of ALL, comprising approximately 9% of childhood ALL, with a gene-expression signature resembling that of Ph+ ALL (Figs. 19.5 and 19.6C).110,112 Approximately 50% of Ph-like cases harbor CRLF2 rearrangements (Fig. 19.12D), but the remainder appear to have a variety of driver lesions. Whole-genome and mRNA sequencing of a cohort of 15 cases identified alterations in those cases lacking CRLF2 rearrangements that included rearrangements of PDG-FRB, ABL1, JAK2, and EPOR, and deletion or mutation of SH2B3 and IL7R.113 These cases share in common the activation of downstream kinase signaling pathways such as JAK/STAT, mTOR, PI3K, MEK, and ERK and may be responsive to JAK and ABL1/PDGFRB inhibitors.114,115
CRLF2 and IL7RA form a heterodimeric cytokine receptor that mediates B-cell precursor proliferation and survival through activation of downstream JAK/STAT and PI3K/mTOR pathways.116 Several alterations of CRLF2 have been noted in ALL, all of which lead to CRLF2 overexpression. These include a focal interstitial deletion of the pseudoautosomal region of the sex chromosomes, which creates the P2RY8-CRLF2 fusion, an IGH@ translocation (IGH@-CRLF2) and the CRLF2 F232C point mutation (Table 19.1, Fig. 19.12D).28,29,117 CRLF2 alterations have been noted in 5% to 8% of pediatric114 and 10% to 15% of adult118 ALL and have a higher incidence in high-risk ALL, in patients of Hispanic/Latino ethnicity, and in patients with Down syndrome where the incidence is approximately 50%.28,29 IL7R gain of function mutations are seen in 6% to 10% of ALL and are noted in both B and T phenotypes.119,120
Related posts:
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
