Acute Lymphoblastic Leukemia in Adults

Epidemiology

It is estimated that in the year 2019, approximately 5930 new cases of ALL were diagnosed and 1500 deaths occurred due to the disease in the United States. ALL is primarily a cancer of childhood; the peak incidence (7.7 in 100,000) occurs between the ages of 1 and 4 years, and approximately 60% of the patients are diagnosed before the age of 20 years. The incidence of ALL begins to decline with increasing age after the first decade of life. A second upward trend starts to emerge in the sixth decade of life, and a much smaller peak is seen in patients older than 85 years of age (1.8 in 100,000). Men have a slightly higher incidence of ALL than do women (male-to-female ratio, 1.4:1). The overall age-adjusted rate of ALL has increased from 0.93 in 100,000 in the year 1975 to 1.38 in 100,000 in the year 2015. Similar increases in the incidence of ALL have also been reported in Scandinavia, the United Kingdom, and Italy. Although most investigators agree that this increase in the incidence is beyond what might be expected from better reporting, the actual cause(s) of this increased incidence remains largely speculative.

Etiology

A small minority of ALL cases (<5%) are associated with predisposing inherited syndromes such as Down syndrome, Bloom syndrome, ataxia telangiectasia, and Nijmegen breakage syndrome. However, the underlying etiology is not known in most cases. Although alcohol use, exposure to pesticides or solvents, and cigarette smoking have all been implicated, only ionizing radiation has been significantly linked to an increased risk of developing ALL. The vast majority of attempts to identify causative agents for ALL have been at best correlative; most of it borders on pure speculation and conjecture. Only recently has it been appreciated that the interaction between genetic predisposition and environmental factors involved in leukemogenesis is complex and may vary for different subtypes of ALL. Epidemiologic studies designed to identify environmental agents involved in leukemogenesis will therefore have to take into account the biologic subsets defined by morphology, immunophenotype, karyotype, and the molecular abnormalities and consider the possibility that for each of these subtypes the causative agent may well be different. To adequately investigate these issues, future studies will require collaborative efforts studying large patient populations.

Despite the limitations described, it is relevant to review some of the recent advances in our understanding of the etiology of ALL. Greaves and colleagues have used DNA obtained from neonatal blood spots to demonstrate that ALL-associated chromosomal translocations were acquired in utero (and hence a preleukemic clone) because they could be demonstrated in neonatal blood spots. Furthermore, these ALL-associated chromosomal translo cations were present at a 100-fold higher rate than the incidence of leukemia. These investigators hypothesized that overt leukemia evolves as a consequence of an abnormal lymphoid proliferation that occurs in response to exposure to an as yet unidentified infectious agent(s). Supporting this hypothesis are data that suggest that day care attendance associated with early exposure to common infectious agents is associated with a lower incidence of ALL. Similarly, industrialization associated with improved socioeconomic status and exposure to common infectious agents later in life have been postulated to result in abnormal and excessive lymphoid proliferation and leukemic transformation. Finally, an infectious etiology can also be evoked to explain leukemic clusters that occur when a previously unexposed community is exposed to infectious agents brought into the community by a large influx of residents, as happens during urbanization of rural communities. However, it should be reiterated that all of these studies are at best correlative and not supported by direct experimental evidence. Furthermore, the infectious or environmental agent(s) responsible for this abnormal lymphoid proliferation remains elusive.

More recently genome-wide association studies (GWASs) demonstrated that susceptibility to ALL is polygenic. Several GWASs identified single nucleotide polymorphisms (SNPs) in eight loci influencing ALL risk at 7p12.2 (IKZF1) , 9p21.3 (CDKN2A) , 10p12.2 (PIP4K2A) , 10q26.13 (LHPP) , 12q23.1 (ELK3) , 10p14 (GATA3) , 10q21.2 (ARID5B) , and 14q11.2 (CEBPE) . ALL is biologically heterogeneous, and subtype associations have been identified for 10q21.2 (ARID5B) associated with high-hyperdiploid B-acute lymphoblastic leukemia (B-ALL) (i.e., >50 chromosomes) and 10p14 (GATA3) associated with B-cell receptor (BCR) ABL-like B-ALL. Vijayakrishnan and colleagues also applied GWASs to identify risk loci for B-ALL at 8q24.21 and for ETV6-RUNX1 fusion-positive B-ALL at 2q22. Although these SNPs have been identified as risk factors for B-ALL, they do not appear to correlate with survival outcomes based on these early data.

Retrospective analyses have also reported cases of therapy-related ALL (t-ALL) in patients who received chemotherapy and/or radiation for a prior malignancy. Several institutions in the United States jointly reported a higher frequency of poor-risk cytogenetic features, including KMT2A rearrangements and myelodysplastic syndrome–like abnormalities (e.g., monosomal karyotype) in t-ALL. In addition, they observed a variety of mutations among t-ALL patients, with the majority of patients exhibiting mutations that were more common in myeloid malignancies (e.g., DNMT3A, RUNX1, ASXL1), whereas others had ALL-type mutations (e.g., CDKN2A, IKZF1). Median overall survival (OS) was significantly shorter in the t-ALL cohort compared with patients with de novo ALL. Patients who were eligible for allogeneic stem cell transplant had improved long-term survival. Washington University reported similar findings in a retrospective analysis at their institution, and a meta-analysis showed t-ALL had a higher risk of relapse and death than de novo ALL but outcomes were better in those who underwent transplant.

Thus predisposing genetic alterations and prior chemotherapy and/or radiation exposure are emerging as some of the few known causes of ALL.

Clinical Manifestations

The clinical presentation of ALL encompasses a wide spectrum of symptoms that correlate with the degree of bone marrow (BM) involvement and the resultant cytopenias, as well as the leukemic cell burden. Typical symptoms include fatigue, anorexia, night sweats, pallor, shortness of breath, bone pain, fever, and bleeding diathesis. Involvement of extramedullary sites may present with lymphadenopathy, hepatomegaly, or splenomegaly. Less commonly, ALL can involve the central nervous system (CNS), leading to headache, vomiting, lethargy, and cranial nerve palsies. Other extramedullary sites of involvement include the testis, tonsils, adenoids, breast, and gastrointestinal (GI) tract. Precursor T-cell ALL often presents with a large mediastinal mass with associated respiratory distress or possible signs of superior vena cava syndrome. Burkitt leukemia/lymphoma is frequently associated with CNS involvement and bulky adenopathy.

Clinical and Laboratory Evaluation

The initial work-up for patients with suspected ALL is detailed in Table 68.1 . All patients should undergo a detailed history and physical examination. Family history should be ascertained. Laboratory evaluation should include a complete blood count (CBC) with differential; and comprehensive metabolic panel, including liver function tests, lactate dehydrogenase (LDH), and uric acid. A coagulation profile should also be obtained, although coagulation parameters are frequently normal at diagnosis. Human leukocyte antigen (HLA) typing should be performed for patients who are potential candidates for autologous stem cell transplantation (ASCT).

All patients should undergo BM aspiration and biopsy for confirmation of diagnosis and for cytogenetic and molecular genetic evaluation. It is particularly important to send an aspirate (or peripheral blood if lymphoblasts are present) to the molecular oncology diagnostic laboratory to evaluate for the presence of the BCR-ABL fusion transcript using reverse-transcriptase polymerase chain reaction (PCR) because these patients will receive frontline therapy that includes a tyrosine kinase inhibitor (TKI). Skin biopsy for fibroblast culture for germline testing can also be obtained at the time of diagnostic BM biopsy if there is concern for a familial leukemia predisposition syndrome. Increasingly, chromosomal microarrays, gene expression profiling, RNA sequencing (RNAseq), and whole-genome sequencing are being incorporated into the work-up of ALL patients.

A lumbar puncture should be performed at diagnosis to determine CNS involvement. In the event of increased risk of bleeding caused by severe thrombocytopenia or risk of cerebrospinal fluid contamination caused by high peripheral blood blasts, the lumbar puncture should be performed by an experienced operator. It is prudent to administer intrathecal chemotherapy at the time of the diagnostic lumbar puncture after obtaining the necessary samples. A white blood cell (WBC) count of greater than 5/μL with morphologically identifiable blasts is considered as evidence of cerebrospinal fluid involvement by leukemia ¹ .

Approach to Diagnosis

Although evaluation of morphology and immunophenotype are sufficient for making the diagnosis, current risk stratification relies on additional cytogenetic and molecular genetic information. Data from these tests are therefore an important adjunct to the initial diagnostic work-up.

Initial laboratory evaluation starts with a CBC and morphologic evaluation of a Giemsa-stained peripheral blood smear. An abnormality of at least one of the CBC parameters is detected in more than 90% of ALL patients at the time of diagnosis. Anemia and thrombocytopenia are common. The anemia is usually a normochromic, normocytic anemia accompanied by reticulocytopenia. The hemoglobin levels range from 30 to 174 g/L, and almost 50% of the patients have hemoglobin levels below 100 g/L. The median platelet count at presentation is approximately 55 to 60 × 10 ⁹ /L, and almost 60% to 70% of patients have platelet counts below 100 × 10 ⁹ /L. Although the total WBC count may be low, normal, or elevated, neutropenia is commonly present. In a Cancer and Leukemia Group B (CALGB) study, the median WBC count at presentation was 19.3 × 10 ⁹ /L. Almost one-third of the patients are likely to present with a WBC count greater than 30 × 10 ⁹ /L. Blasts account for a variable proportion of the circulating WBCs, and the percent blast population can range from 0% to 100%. A leukoerythroblastic picture can sometimes be seen. In an extreme form, immature myeloid precursors and myeloblasts constitute the vast majority of cells in the peripheral blood. This should be kept in mind when attempting to make a diagnosis exclusively from peripheral blood. Eosinophilia as a presentation of ALL is extremely uncommon and is seen in association with specific chromosomal abnormalities, including t(5;14)(q31;q32) or, even less frequently, with 8p11-associated ALL (see Chapter 57 ). Eosinophilia associated with t(5;14) is reactive and due to overexpression of interleukin-3 on chromosome 5 driven by the immunoglobulin H (IgH) promoter on chromosome 14. The eosinophilia can be extremely pronounced and mask the blast population in this subset of patients.

Several metabolic abnormalities are present at the time of diagnosis and frequently reflect tumor burden. For example, LDH levels are frequently elevated, and almost 50% of the patients have levels between 300 and 1000 U/L. Elevated serum levels of calcium, potassium, and phosphorous have been noted. More importantly, elevated serum uric acid levels are frequently present and reflect tumor burden. Hyperuricemia needs to be carefully monitored and aggressively corrected to avoid renal failure, especially at the time of starting induction therapy.

Morphology

Romanowsky-based stains such as Wright-Giemsa and Giemsa provide the greatest cytoplasmic detail for evaluation of cytomorphology of the cells in the peripheral blood smear, BM aspirate smear, and touch imprints ( Fig. 68.1 ). Most frequently, lymphoblasts are small to intermediate in size and have scant, agranular cytoplasm. B lymphoblasts are morphologically indistinguishable from T lymphoblasts, and this distinction relies on immunophenotyping. The nuclei are usually round, with uniformly dispersed “smudgy” chromatin and inconspicuous nucleoli (see Fig. 68.1C ). However, variations in morphology are common, and larger cells with abundant bluish gray cytoplasm, larger, somewhat irregular nuclei, and variably prominent nucleoli can be frequently seen (see Fig. 68.1D ). Even though the nuclear chromatin of these cells can be fine, it is never as finely dispersed as in a myeloblast. At the other end of the morphologic spectrum are smaller cells with uniformly condensed, mature lymphocyte-like chromatin (see Fig. 68.1E ), and distinction from mature B-cell malignancies relies on immunophenotyping. Coarse azurophilic granules (see Fig. 68.1F ) can be seen in a subset of blasts in 5% to 8% of childhood ALLs and even more frequently in adult ALL patients. These have been reported in association with Philadelphia chromosome-positive (Ph ⁺ ) ALL and in ALL in Down syndrome patients. The granules are coarser than the granules seen in myeloblasts and are invariably myeloperoxidase negative (see later discussion of cytochemistry). Morphologic distinction between the L1 and L2 category recommended in the French-American-British (FAB) classification has proven to be poorly reproducible and of little prognostic value. It has been abandoned in the current World Health Organization (WHO) classification.

Cytoplasmic vacuolation (see Fig. 68.1H ) can be seen in as many as 28% of childhood ALL patients. These lymphoblasts can be distinguished from leukemic presentation of Burkitt lymphoma (BL) (see later discussion) based on other morphologic features such as a smaller cell size, lack of deep blue cytoplasm, and less coarse chromatin. However, when morphology is confounding, the distinction relies on immunophenotyping of the malignant cells. In contrast to BL cells, ALL blasts are precursor B cells that express terminal deoxynucleotidyl transferase (TdT) and lack surface immunoglobulin (Ig) expression (see later discussion).

The trephine biopsy sections show hypercellular BM ( Fig. 68.2 ). The sections are evaluated after staining with hematoxylin and eosin. The BM is usually packed with a relatively uniform population of small round blasts with round to oval nuclei. Less frequently, the blasts can be more pleomorphic, with indented, convoluted, and variably sized nuclei. The chromatin is described as being finely dispersed or stippled, and the nucleoli are usually not conspicuous. Brisk mitotic activity is almost always present. The effacement of the BM space is almost complete and uniform at the time of initial presentation. Minimal residual hematopoiesis is present; in most instances this is represented by a few megakaryocytes and some erythropoiesis. Normocellular or even hypocellular BMs at presentation have been described but are uncommon. Rarely, the initial presentation of ALL can be with an aplastic or markedly hypocellular BM. Making the diagnosis in this hypocellular context can be particularly challenging because of a paucity of material available for supporting studies such as cytogenetics and immunophenotyping.

The BM biopsy can reveal partial or complete necrosis (see Fig. 68.2C ). When extensive necrosis is present, making a diagnosis can be challenging or almost impossible. A repeat BM biopsy should be attempted and will usually provide diagnostic material. Some increase in reticulin fibrosis is present in 60% to 70% of ALL patients. When fibrosis is extensive, an aspirate cannot be obtained, limiting material available for ancillary studies such as flow cytometry and cytogenetics. For these patients, the immunophenotyping can be performed by immunoperoxidase immunohistochemistry on the bone core biopsy. A second core can be obtained and submitted without fixation for cytogenetic analyses.

Organs other than the BM can be frequently involved. Extramedullary or lymphomatous presentation is more common with T-acute lymphoblastic leukemia (T-ALL) than B-ALL. The cytomorphology of the malignant cells in extramedullary disease is similar to that described in the bone core biopsy. Lymph node involvement is usually diffuse but can be partial with sparing of the follicles.

Cytochemistry

The use of cytochemistry to assign lineage has been largely replaced by flow cytometry evaluation of the leukemic blast immunophenotype. However, when available, the myeloperoxidase reaction ( Fig. 68.3A and B ) permits a rapid distinction from acute myeloid leukemia (AML). The reaction detects the myeloperoxidase enzyme in the primary granules of myeloblasts and is specific for the myeloid lineage. An acute leukemia in which 3% or more of the cells are myeloperoxidase positive is considered myeloid. There is excellent concordance between the myeloperoxidase reaction detected by cytochemistry and the myeloperoxidase molecule detected by flow cytometry.

Immunophenotype

Based on large cooperative group studies in Europe and the United States, B-ALL comprises 75% to 80% and T-ALL from 15% to 25% of cases of ALL. B lymphoblasts cannot be distinguished from T lymphoblasts by morphology. Extensive immunophenotypic characterization is therefore required for the appropriate classification of ALL and indeed distinction from certain subtypes of AML. When adequate material is available, immunophenotyping should be performed using multicolor flow cytometry so that multiple antigens can be detected simultaneously on the lymphoblasts ( Fig. 68.4 ). When interpreting the immunophenotypic data, it is important to remember that no single antigen is specific for any given lineage and multiple antigens need to be evaluated to establish the correct diagnosis. The panel of antibodies used for flow cytometry of a new leukemia and the pattern of expression seen in B-ALL and T-ALL are shown in Table 68.2 . In addition, the combination of markers expressed on the B or T lymphoblasts can be reflective of the stage of development at which the transformation happened ( Table 68.3 ). Of note, expression of myeloid antigens is seen frequently in B-ALL and T-ALL, as is the expression of T-cell antigens in B-ALL and B-cell antigens in T-ALL. Expression of individual myeloid antigens should not be a deterrent to making the diagnosis of ALL. Of particular significance is the subgroup of T-ALL that lack CD1a, CD8, and CD5 but show expression of one or more myeloid or stem cell markers and have been recently designated as early T-cell precursor ALL . Although these leukemias can have a gene expression profile and spectrum of genetic mutations that overlap significantly with biphenotypic leukemia, most studies suggest improved outcomes using T-ALL therapy. It is therefore best to recognize this type of ALL as a type of T-ALL rather than mixed lineage leukemia. The criteria for diagnosis of acute leukemia of ambiguous lineage, which would include mixed phenotype acute leukemia, have been extensively revised in the current WHO classification.

Cytogenetics and Molecular Genetics

Chromosomal abnormalities can be detected in almost 80% of B-ALLs and 70% of T-ALLs (see Chapter 57 ). Cytogenetic classification remains the single most important prognostic factor in both pediatric and adult ALL. Numerical abnormalities, as well as structural abnormalities that disrupt the function of transcription factors involved in hematopoietic development and differentiation, are common. These genetic abnormalities define the biology of the disease and have an impact on treatment outcome. In addition, specific cytogenetic or molecular abnormalities are associated with unique phenotypic characteristics and are amenable to targeted therapy. BCR-ABL1 ⁺ (referred to as Ph ⁺ ) ALL was the genetically defined subtype of B-ALL incorporating molecularly targeted TKI therapy in the frontline treatment. Studies over the past decade have identified the existence of a subset of B-ALL with a gene expression profile similar to that of BCR-ABL1 ⁺ patients but without the BCR-ABL1 fusion. This group of patients is referred to as BCR-ABL1 –like (or Ph-like) and is also likely to benefit from targeted TKI therapy (see discussion on Philadelphia chromosome–like ALL [Ph-like ALL] later). The 2016 revision of WHO classification added the category of B-lymphoblastic leukemia/lymphoma BCR-ABL1–like to the previously recognized genetically defined entities ( Table 68.4 ).

Hyperdiploidy, defined by the presence of more than 50 chromosomes, is seen in almost 25% of pediatric patients and 4% to 5% of adult ALL patients. In addition to routine karyotyping or fluorescence in situ hybridization (FISH) analysis, hyperdiploid DNA content can be determined by flow cytometry using DNA-binding fluorescent dyes and corresponds to a DNA content between 1.16 and 1.6. However, some studies have demonstrated that hyperdiploidy resulting from duplication of specific chromosomes (4, 10, and 17) is a better indicator of a favorable prognosis than the actual ploidy or the DNA content. Careful analysis of the specific pattern of chromosomal gains and losses or flow cytometry peaks for DNA content is required to distinguish true hyperdiploid cases from near-haploid cases that have undergone endoreplication (vide infra). In contrast to hyperdiploidy, a hypodiploid karyotype is associated with an adverse prognosis. Hypodiploid ALL with near-diploid chromosome numbers (44 to 45) is associated with somewhat distinct biology (loss of sex chromosome, dicentric chromosomes, presence of recurrent cytogenetic abnormalities such as ETV6-RUNX1) and is associated with a better outcome. In contrast, hypodiploidy with chromosome numbers less than 44 is invariably associated with poor outcome. The most common chromosome complements within the hypodiploid group are near haploid (24 to 31 chromosomes) and low hypodiploid (32 to 39 chromosomes) and are both associated with a poor outcome. Of note, low-hypodiploid ALL is frequently associated with germline TP53 mutations and at least in a subset of cases appears to be a manifestation of Li-Fraumeni syndrome (see Chapter 66 ). Both near-haploid and low-hypodiploid ALL have activation of phosphatidylinositol 3-kinase (PI3K)/mammalian target of rapamycin (mTOR) and MEK-ERK signaling that can potentially be targeted by PI3K inhibitors.

A structural abnormality seen commonly in the pediatric age group but extremely rarely in the adult age group is a cryptic translocation, the t(12;21)(p13;q22) ( ETV6-RUNX1 ). When present, it is associated with a favorable prognosis. Leukemias that harbor rearrangements of the mixed-lineage leukemia ( MLL ; now referred to as KMT2A ) gene at chromosome 11q23, most notably t(4;11)(q21;q23), present with high WBC counts and frequent CNS involvement and are associated with poor clinical outcomes. Ph ⁺ ALL associated with t(9;22)(q34;q11.2) is seen in 25% of adults diagnosed with ALL. The translocation results in BCR-ABL1 fusion, and the resulting leukemia has a poor outcome without the inclusion of TKIs into frontline treatment. Intrachromosomal amplification of chromosome 21 (iAMP21) is defined as a gain of at least three copies of regions of chromosome 21 that include RUNX1 . The abnormality seen in 2% of B-ALL patients is associated with an older age at diagnosis and poor outcome with standard-risk therapy. iAMP21 is usually the sole cytogenetic abnormality but can be seen in patients with Ph-like ALL (see later). In addition, use of technology to assess copy number variation has identified submicroscopic genomic alterations (see Chapter 57 ) that have significant impact on the biology of B-ALL. These involve transcription factors such as IKZF1 (Ikaros), EBF , and PAX5 . The role of these transcription factors in B-cell development and the biology of B-ALL are discussed in greater detail in Chapter 66 .

One of the most critical and clinically relevant discoveries with respect to ALL has been the description of high-risk ALL with a gene expression profile similar to that of Ph ⁺ ALL but without the BCR-ABL1 fusion gene. This subtype of leukemia is referred to as BCR-ABL1 –like or Ph-like ALL , and the kinase-activating alterations are a result of a diverse group of genetic lesions that result in dysregulated cytokine receptor and tyrosine kinase signaling. The lesions can be grouped into a surprisingly small number of broad categories. The most frequent category involves CRLF2 rearrangements that account for almost 50% of Ph-like ALLs. The rearrangements involve either translocation of IGH2 at locus on chromosome 14q32 to CRLF2 on Xp22.3/Yp11.3 (pseudoautosomal region 1) or a 320-kb interstitial deletion centromeric of CRLF2 resulting in P2RY8-CRLF2 fusion. CRLF2 overexpression is frequently associated with activating mutations of JAK1/JAK2 , IL7R , or deletion of the JAK-STAT negative regulator SH2B3 , resulting in activation of JAK-STAT signaling. JAK-STAT signaling activation also occurs in a second category of Ph-like ALL with JAK2/EPOR rearrangements. The third broad category of Ph-like ALL involves ABL1 , ABL2 , CSF1R , and PDGRB , described together as ABL1 class rearrangements. In addition, other kinases such as NTRK3 , PTK2B , and TYK2 have also been shown to be involved in rearrangements in sporadic cases. As preclinical studies and case reports continue to document dramatic responses to targeted therapy for this group of patients, it is imperative that these lesions be identified prospectively. Because a wide variety of lesions result in a Ph-like phenotype, the laboratory diagnosis of Ph-like ALL remains a challenge. One approach to work-up of Ph-like ALL is highlighted in Fig. 68.5 .

In addition to identification of Ph-like B-ALL, recent studies that use multiple techniques including chromosomal microarrays, gene expression profiling, RNAseq, and whole-genome sequencing have identified novel genetic subtypes that had were not detectable by karyotype and FISH analyses. These genetic subtypes show alteration in PAX5, MEF2D, ZNF384, and DUX4 genes (reviewed in references and ). The incidence and impact on outcome for these abnormalities remain an area of active investigation.

Similar to B-ALL, translocations involving transcription factors are common in T-ALL. The most commonly involved genes include HOX11 and HOX11L2 . Other genes that are involved in T-ALL are MYC , TAL1 , LMO2 , and LYL1 . Similar to t(12;21), translocations involving the TAL1 gene are cryptic and require detection by molecular techniques (see Chapter 57 ). Activating mutations of the NOTCH1 gene are detected in almost 50% of ALL patients (see Chapter 66 ). Although activating mutations of NOTCH1 appear to be associated with disease pathogenesis, they do not appear to be associated with an adverse prognosis; in fact, some pediatric studies suggest that NOTCH1-mutated T-ALL may have a relatively favorable response to current treatment regimens. Deletions of the CDKN2A gene on chromosome 9p are also particularly frequent as are mutations in the FBXW7 gene, but neither has clear prognostic significance as single abnormalities. The significance of these mutations remains an area of active research.

Differential Diagnosis

ALL blasts can be easily distinguished from the reactive lymphocytes which accompany viral infections, because of the precursor phenotype of these cells. Low to weak expression of CD45 and expression of one or more precursor antigens such as TdT or CD34 is useful. In addition, precursor T-ALL cells express only cytoplasmic CD3 and no surface CD3, and B-ALLs frequently lack expression of CD20 while being CD19 positive. As mentioned previously, ALL cells can have some cytoplasmic vacuoles and need to be distinguished from leukemic presentation of BL. Unlike ALL, BL cells are mature B cells with bright surface Ig expression, with very strong CD20 expression, and that lack expression of CD34. Diagnosis of BL is confirmed by the presence of an MYC translocation using FISH or karyotypic analyses. Other entities that require distinction from ALL depend on the age of presentation. In the pediatric age group, ALL blasts need to be distinguished from hematogones. Hematogones are normal B-cell precursors present within the BM. These are more abundant in childhood and decrease with increasing age. Hematogones may also be increased during hematopoietic regeneration, particularly after chemotherapy or BM engraftment after stem cell transplant. Hematogones possess a distinct pattern of antigen expression that recapitulates progressive B-cell maturation. This is reflected in progressive loss of antigens such as CD34, TdT, and CD10 and acquisition of CD20 and surface Ig expression ( Fig. 68.6 ). Other diseases that need to be morphologically distinguished from ALL include small blue cell tumors, including Ewing sarcoma, neuroblastoma, and medulloblastoma. Ancillary studies, including immunophenotyping and cytogenetics, are helpful in making the distinction. In older adults, entities that can morphologically mimic ALL include blastoid mantle cell lymphoma, chronic lymphocytic leukemia, and prolymphocytic leukemia (see Chapter 76, Chapter 86 ). These latter are all mature B-cell malignancies that can be distinguished from ALL based on the mature B-cell phenotype, including consistent expression of CD20 and surface Igs.

Prognosis

Prognostication based on clinical and biologic risk factors has been useful in making informed decisions about postremission treatment options. Established risk factors for a poor prognosis with current chemotherapeutic approaches include age older than 60 years, elevated WBC count at diagnosis (>30,000/μL for B-cell ALL; >100,000/μL for T-cell ALL), pro–B-cell or early T-cell immunophenotype, and cytogenetics (t[4;11][q21;q23] and other MLL rearrangements, hypodiploidy, or a complex karyotype). The presence of the Philadelphia chromosome, t(9;22)(q34;q11.2) resulting in the BCR-ABL fusion gene, was previously associated with very poor treatment outcomes. Recent addition of ABL kinase inhibitors (molecularly targeted therapy), discussed in detail later, has improved the prognosis for these patients. Time to achievement of complete remission (CR) longer than 4 weeks has also been associated with a poor clinical outcome ( Table 68.5 ).

Although different study groups have used slightly different variations of these risk factors, the presence of any one of the following is generally accepted as high risk (HR): high WBC count at diagnosis (>30,000/μL in B-cell ALL or >100,000/μL in T-cell ALL); cytogenetic abnormalities [hypodiploidy, t(4;11), t(9;22)]; age older than 60 years; pro–B-cell phenotype; and time to remission longer than 4 weeks. All other patients are considered standard risk (SR).

Nevertheless, despite being labeled as SR, up to 40% to 50% of these adults eventually relapse. Thus there is a need for refinement of prognostic markers for ALL patients. Recently, an early T-cell phenotype, the expression of CD20 ⁺ , and several recently described genetic mutations ( IKAROS , CRLF2 , and JAK2 ; reviewed earlier in this chapter) have been identified as being associated with adverse outcomes from retrospective analyses.

Minimal Residual Disease

The identification and quantification of measurable or MRD (below the level of morphologic disease detection) are now considered the key prognostic marker of disease-free survival (DFS). A variety of methods are used to detect MRD, including multiparameter flow cytometry (MFC), quantitative PCR, and next-generation sequencing (NGS); highest sensitivity is yielded from BM samples as compared with peripheral blood. While flow cytometry–based methods have issues with lack of standardization, they can detect aberrant patterns of cell surface antigen expression unique to the leukemic clone. PCR methods detect particular fusion transcripts such as BCR-ABL translocation, where its use has been most widespread and very well standardized as an outstanding marker for MRD detection, or clone-specific rearrangements as quantitative PCR (clone-specific IgH, or T-cell receptor [TCR] rearrangements) which are well standardized in European centers. NGS methods identify similar clone-specific gene rearrangements with a higher threshold of sensitivity at 10 ⁻⁶ . Regardless of method, many prospective studies have been performed and demonstrate the utility of MRD monitoring for prognosis and treatment stratification. The NGS methodology perhaps holds the most promise for future MRD testing given its ability to standardize methods, its higher limit of sensitivity, specificity for residual disease, and ability to track minor subclones of evolving disease. Consensus opinion has recommended MRD testing at the end of induction, in early consolidation, and approximately for every 3 months thereafter for at least 3 years, as well as immediately prior to stem cell transplantation (SCT) and serially thereafter.

A recent meta-analysis of MRD in pediatric and adult ALL demonstrated that MRD negativity is associated with improved event-free survival (EFS) and OS at 10 years, when analyzing studies from 2000 to 2014. Among studies of pediatric ALL, MRD negativity was associated with an EFS of 77% at 10 years and those with detectable MRD had an EFS of 32%. Among studies of ALL in adults, MRD negativity was associated with an EFS of 64% at 10 years, compared with 21% among those with detectable MRD. These findings, importantly, held true across modalities of MRD detection method, timing of MRD testing, karyotype of underlying disease, and phenotype of ALL—with none of these factors holding significance on the effect of MRD on hazard ratios of EFS or OS.

Individual studies of MRD among specific phenotypes of ALL have revealed new insights on prognosis and potential future study and/or treatment. Adult patients with Ph-like ALL, MLL gene rearrangements, and early T-cell precursor ALL have had relatively poor outcomes with standard aggressive chemotherapy regimens and are more likely to have persistent MRD positivity after intensive therapy, guiding many physicians to intervene with allogeneic stem cell transplant during first CR for these subsets. By contrast, with the addition of novel TKIs and targeted immune therapies, the absence of MRD during intensive frontline treatment of Ph ⁺ ALL may be changing the general recommendation away from allogeneic stem cell transplant in first remission for all eligible patients. Among Ph ⁺ ALL patients receiving a BCR-ABL1 TKI who achieved a complete molecular remission (CMR), defined as absence of detectable MRD, after intensive chemotherapy and a TKI, who additionally did not undergo a stem cell transplant, a 4-year OS rate of 66% was achieved. Of note, the median OS for those achieving a CMR at 3 months was a statistically significant 127 months versus 38 months for those who did not achieve CMR by real-time-PCR detection of BCR-ABL1 transcripts. Similarly, Ph ⁺ ALL patients who received lower-intensity regimens after achieving CMR have been shown in a GIMEMA (Gruppo Italiano Malattie Ematologiche) study to have superior survival outcomes on TKI maintenance therapy post induction, despite not continuing on to higher-intensity chemotherapy and/or allogeneic stem cell transplant.

In addition to its prognostic relevance, MRD is becoming more widely accepted as an acceptable clinical end point in clinical trials and drug approval. Blinatumomab, the bispecific monoclonal T-cell engager antibody binding CD3 and CD19, has been approved for treatment for MRD-positive disease. The use of MRD as an end point or prognostic marker may facilitate more rapid evaluation of new, emerging therapies, which might allow one to assign highest priority for early-phase trials which demonstrate low rates of MRD for continued development and approval in phase III studies to evaluate survival.

Treatment of Acute Lymphoblastic Leukemia