Project no ENCCA. EUROPEAN NETWORK for CANCER research in CHILDREN and ADOLESCENTS. Network of Excellence. Deliverable Number: 9.

Transcription

1 Project no ENCCA EUROPEAN NETWORK for CANCER research in CHILDREN and ADOLESCENTS Network of Excellence Deliverable Number: 9.1 Title: Common guidelines for diagnostic approaches to leukemias Due Delivery Date: M24 Actual Delivery Date: M24 Start date of project: 1 January 2011 Duration: 48 months Project co-funded by the European Commission within the Seventh Framework Programme ( ) Dissemination Level PU Public 1

2 Summary In this deliverable we describe consensus recommendations on the diagnosis of childhood acute leukemia from a European viewpoint based on surveys of the groups participating to the AIEOP-BFM ALL 2009 study, a 7-country-based international frontline treatment trial on acute lymphoblastic leukemia (ALL) in children and adolescents, a survey of the work group DCOG Molecular Research of the Dutch Childhood Oncology Group, and the Biology & Diagnosis Committee of the International BFM Study Group, the largest initiative on childhood leukemias worldwide which unifies cooperative national study groups on childhood leukemia from more than 30 different, mostly European countries. Topics covered relate to cytomorphology, immunophenotyping, cyto- and molecular genetics, and the evaluation of measures of treatment response. The recommendations provide a common ground for systematic implementation of molecularly-based diagnostic strategies in acute leukemias to be developed within ENCCA and, therefore, can be seen as a prerequisite for the implementation of rational molecular information-based treatment approaches in pediatric hematology and oncology. The diagnostic guidelines will be published on the website of the international BFM study group ( 2

3 Description Common guidelines for diagnostic approaches to leukemias Correspondence to: Susanne Kilian, PhD, MLaw University Medical Center Schleswig-Holstein Department of General Pediatrics Campus Kiel Schwanenweg Kiel, Germany Tel.: +49 (0) Fax: Disclaimer: While the advice and information in these guidelines is believed to be true and accurate, neither the authors, the International BFM Study Group nor the European Network for Cancer Research in Children and Adolescents accept any legal responsibility for the content of this document. Date of recommendation review: January 2013 Writing group: ENCCA WP9 and I-BFM Biology & Diagnosis Committee members Declarations of Interest: None of the authors are considered to have a conflict of interest with regards to the recommendations made in this document. 3

4 Contents Deliverable in context of WP9 5 Recommendations for the diagnosis of childhood acute lymphoblastic leukemia 6 General information Clinical presentation and diagnosis Prognostic factors and risk-adapted treatment Diagnosis of childhood acute lymphoblastic leukemia 12 Diagnostic methods Cytomorphology Immunophenotyping Cytogenetics and FISH Molecular genetics Treatment response evaluation Biobanking Recommendations for the diagnosis of childhood acute myeloid leukemia 19 General information Clinical presentation and diagnosis Prognostic factors and risk-adapted treatment Diagnosis of childhood acute myeloid leukemia 22 Diagnostic methods Cytomorphology Immunophenotyping Cytogenetics and FISH Molecular genetics Treatment response evaluation Biobanking References 24 4

5 Deliverable in context of WP9 WP9 aims to establish a harmonized and integrated approach to the rational introduction of molecularly targeted treatment in clinical trials on leukemias. In order to achieve this ultimate goal, the development of standardized comprehensive diagnostic approaches as well as biobanking, and the establishment of a common pipeline for molecular diagnostics in a European virtual laboratory on leukemias are necessary. These recommendations are neither obligatory nor is it necessary to follow all of them. 5

6 Recommendations for the diagnosis of childhood acute lymphoblastic leukemia (ALL) General information ALL represents the malignant proliferation of lymphoid cells blocked at an early stage of differentiation and is the most common malignancy in children. It accounts for approximately 25% of all childhood cancers and about 80% of childhood leukemias. The annual incidence rate of childhood ALL varies world-wide between approximately one and four new cases per 100,000 children younger than 15 years, with a peak incidence at approximately two to five years of age. More affluent countries tend to have higher incidence rates. However, incidence rates for childhood ALL do not only vary between countries, but also by ethnicity within countries: in the USA rates are highest in Hispanic children and higher in white compared to black children. More than 60% of patients diagnosed with ALL are children. Treatment results in childhood ALL are one of the true success stories of clinical oncology with current overall cure rates of approximately 80% in developed countries. These results are reached by application of intensive multiagent chemotherapeutic regimens and in specific patient subgroups additional radiotherapy and/or hematopoietic stem cell transplantation (HSCT). Modern treatment regimens consist of at least four phases: (1) an induction period aiming at an initial remission induction within approximately 4 to 6 weeks through the use of multiple cancer chemotherapeutic drugs; (2) consolidation/intensification and reinduction segments to eradicate residual leukemic blasts in patients who are in remission by morphologic criteria; (3) extracompartment therapy such as central nervous system (CNS) preventive therapy, and (4) a maintenance period to further stabilize remission by suppressing re-emergence of a drug-resistant clone through continuing reduction of residual leukemic cells. As certain clinically and biologically distinct patient subgroups with ALL have a particular poor outcome on standard ALL treatment, clinical protocols specifically addressing the potential therapeutic needs of these subgroups have been initiated in the recent past (e.g., hybrid protocols for infants, and imatinibincluding regimens for BCR/ABL1-positive ALL). Clinical presentation and diagnosis The initial clinical presentation of a child with ALL largely depends on the extent of the leukemic infiltration of the bone marrow and extramedullary sites. Typical clinical signs are fever, pallor, fatigue, bruises, enlargement of liver, spleen and lymph nodes, and pain (e.g., bone pain). In most patients, white blood cell counts show anemia, thrombocytopenia and granulocytopenia with or without concommitant leukocytosis. The diagnosis of ALL is usually made by cytomorphological and cytochemical examination of a bone marrow aspirate and in difficult cases by trephine biopsy and is established when at least 25% lymphoblasts are present in the marrow. CNS involvement (CNS3 status) is diagnosed by the presence of blasts in the cerebrospinal fluid (CSF; for definition see Table 1) or if intracerebral infiltrates are detected by cross-sectional radiological imaging. Initial diagnostics are complemented by flowcytometry-based immunophenotyping to gain information on the blasts expression of lymphoid differentiation-associated antigens as measure by the reactivity to specific monoclonal antibodies and to determine the cellular 6

7 DNA content of leukemic cells. In addition, a combined approach using cytogenetic and molecular genetic techniques is used for the detection of genetic aberrations, such as nonrandom recurrent chromosomal translocations or their molecular equivalents (e.g., the t(9;22) or the BCR/ABL1 fusion transcript). Molecular-genetic techniques and/or flow cytometry are also used to monitor disease burden during therapy by measuring minimal residual disease (MRD). A last important issue addresses the definition of what is called complete remission and relapse: complete remission is defined as the absence of leukemic blasts in blood and CSF, fewer than 5% lymphoblasts in bone marrow aspiration smears, and no evidence of localized disease. Relapse is defined as the recurrence of lymphoblasts or localized leukemic infiltrates at any site. Prognostic factors and risk-adapted treatment Continuing research on the clinical and biological aspects of ALL has identified numerous features with prognostic potential some of which are displayed in Table 1. On modern protocols, risk-adapted therapy reflecting the probability of treatment failure has become a common feature in the clinical management of childhood ALL. For this purpose, the initially assessed prognostic factors are used to estimate an individual patient s risk of relapse and to adjust the required treatment intensity by therapy stratification into different risk groups (e.g., standard/low, intermediate, high). The prognostic significance of an inadequate early reduction of leukemic blasts in the peripheral blood was first described by the BFM study group and confirmed by several other study groups. Of importance, the specificity of response evaluation might vary with the composition of the induction regimen and the time point of response evaluation. However, although a poor early response to induction therapy as described above is highly predictive of treatment failure, the majority of recurrences occur in the large group of patients with an adequate morphological response to treatment. Of advantage in this context, the sub-microscopic assessment of MRD is approximately 1000 to fold more sensitive compared to methods based on morphological detection and provides excellent prognostic information. Although most of the experience on MRD in clinical settings was gained through DNA-PCR-based detection of leukemic clone-specific immunoglobulin and/or T-cell receptor gene rearrangements, it was shown that flowcytometry-based analyses by detection of specific antigen patterns of the leukemic clone yield sensitive and reliable results comparable to PCR-based methods. Remission induction treatment Contemporary treatment approaches for childhood ALL aim at an initial remission induction and restoration of normal hematopoiesis within approximately 4 to 6 weeks. In most study groups this goal is achieved in approximately 98% of patients through the systemic application of three drugs (glucocorticoid, vincristine, L-asparaginase) to which an anthracycline may be added as a fourth drug. On ALL-BFM protocols, remission induction is initiated by a 7-day monotherapy with orally administered prednisone (and one intrathecal dose of intrathecal methotrexate on day 1), which isparticularly useful in avoiding complications related to extensive tumor cell lysis. Undoubtedly, the dose intensity of the induction phase can have a major impact on the overall treatment outcome. Nevertheless, in specific subgroups of childhood ALL, the necessity of a four-drug induction regimen is 7

8 subject to debate and it is, for example, unclear if addition of an anthracycline to a threedrug induction regimen is of real benefit to certain low- or intermediate-risk patients. Another frequently discussed issue addresses the choice of the glucocorticoid for optimal induction. Despite some debate on a truly equivalent dose, compared to prednisolone, dexamethasone appears to have a stronger antileukemic effect in vitro and has been shown to provide better leukemic CNS control and lower relapse rates. However, dexamethasone was also associated with increased side effects including severe infectious complications. The 2% of patients not in remission after induction therapy will either have died of treatment- or disease-related complications or display nonresponsive disease. The latter group includes patients that will achieve only delayed remission or show resistant disease. Because of the poor prognosis of this minor non-responsive patient population, alternative therapeutic approaches should be considered early during the disease process. Consolidation/intensification and reinduction treatment Eradication of residual leukemic blasts in patients who are in remission by morphologic criteria is the primary aim of consolidation/intensification treatment. Consolidation/intensification treatment is necessary as patients successfully induced into remission, but not given additional treatment, usually relapse within months. A so-called reinduction or delayed intensification treatment can further enhance the effect of previous consolidation/intensification therapy. As consolidation/intensification phases administered in protocols of the large study groups on treatment of childhood ALL are variable and may differ, for example, with regard to amounts, timing, and number of drug doses, drug composition and overall treatment context, the direct contributions of most of these consolidation/intensification strategies and/or their individual components are difficult to assess and associated with limited generalizability. Today, most protocols use high-dose methotrexate (combined with folinic acid rescue) together with 6-mercaptopurine (6-MP) and/or prolonged administrations of asparaginase in consolidation/intensification. Reinduction treatment mainly consists of a late repetition of the initial remission induction and early intensification phases. A randomized trial by the Children s Cancer Group applying an augmented BFM protocol showed that intensified consolidation and doubledelayed intensification can further improve the outcome of high-risk patients with a slow initial treatment response. Of interest in this context, in a recent subsequent trial on higherrisk patients with a rapid marrow response to induction therapy by the same group, the investigators demonstrated an improved event-free survival for more intensive but not for longer postinduction intensification treatment. Unfortunately, further intensification of treatment and the related higher doses of glucocorticoids have been associated with a high incidence of osteonecrosis, especially in older children. Consequently, some investigators suggest glucocorticoid administration in intensification/consolidation on alternate weeks for children older than 10 years to improve complication rates. Central nervous system-directed therapy CNS-directed therapy has become a prerequisite for successful treatment of childhood ALL. Before its introduction in the 1960s, more than 50% of children with ALL suffered from disease recurrence originating from the CNS. This high rate could be reduced to less than 8

9 5% through the introduction of cranial irradiation, intrathecal chemotherapy with methotrexate alone or in combination with other drugs (cytarabine, hydrocortisone), and systemic application of chemotherapeutics with adequate penetration into the CNS (highdose methotrexate, dexamethasone, high-dose cytarabine). The intensity of CNS-directed treatment is adjusted according to the risk of ALL relapse originating from the CNS, the most important risk factor being overt CNS involvement at diagnosis (CNS3). Additional risk factors include a high initial white blood cell count, pro-b or precursor T-cell immunophenotype, t(9;22) or t(4;11), and a traumatic lumbar puncture with identifiable blast cells present at diagnosis. CNS-directed therapy may differ in the number of intrathecal injections and/or intrathecally applied drugs, as well as in the inclusion of cranial irradiation at different doses. Excluding infants, most clinical protocols administering intensive systemic therapy still recommend preventive cranial irradiation (12 or 18 Gy) for high-risk patients and/or those with a precursor T-cell immunophenotype - at least for those with white blood cell counts of /µl or more at diagnosis. Patients with CNS2 status or a traumatic lumbar puncture are recommended to receive additional therapeutic doses of intrathecal chemotherapy. Also CNS3 patients receive more intense intrathecal chemotherapy and, in addition, are subject to therapeutic cranial irradiation (18 or 24 Gy when two years of age; younger children should receive reduced doses). All other patients (precursor B-cell ALL, CNS1, non HR) should receive preventive intrathecal chemotherapy. Allogeneic hematopoietic stem cell transplantation Results of frontline and relapse protocols have improved over time. At the same time, the experience gained also led to advancements in HSCT procedures. The continuous parallel developments in both fields complicate the description of the exact role of HSCT in childhood ALL and elucidate the strong need for prospective clinical trials. Therefore, in 2003, the ALL-BFM and the ALL-REZ BFM study groups initiated a prospective, international, multicenter trial (ALL-SCT-BFM 2003), which will now be extended to a larger international consortium. In this trial exactly defined procedures on HLA-typing, donor selection, conditioning regimen, graft versus host disease prophylaxis and therapy as well as standards of supportive care ensure a high degree of standardization with regard to all relevant components potentially associated with the heterogeneity in outcome observed in the context of HSCT. It is expected that the results of such prospective trials will more precisely determine the potential of the different HSCT procedures in high-risk or relapsed childhood ALL. Meanwhile, HSCT in children with ALL in first remission should be confined to patients whose disease is associated with poor prognostic features such as the t(9;22) or a poor response to remission induction therapy. Maintenance therapy Hypothetically, maintenance treatment aims at a further stabilization of remission by suppressing the re-emergence of a drug-resistant clone through consistently reducing the pool of residual leukemic cells. The current standard of maintenance therapy consists of up to two or three years of treatment (from initial time of diagnosis) with daily oral 6-MP and weekly oral methotrexate. The combination of 6-MP with methotrexate acts synergistically as methotrexate inhibits purine de novo synthesis, leading to a higher intracellular Availabi- 9

10 lity and increased incorporation of phosphorylated thiopurines in DNA and RNA. During maintenance treatment, 6-MP and methotrexate doses are adjusted according to absolute leukocyte or neutrophil and platelet counts. Important to note and a potential source of heterogeneity with regard to outcome analyses, the starting dose as well as dose adjustment guidelines while on therapy may differ between the different study groups. As several reports suggested an improved outcome with bedtime administration, 6-MP is commonly administered in the evening hours. Also, 6-MP should not be given in combination with milk since the xanthine oxidase activity contained in milk decreases the bioavailability of 6-MP. Of utmost clinical importance, at St. Jude Children s Research Hospital researchers have demonstrated that maintaining the highest tolerable dose of daily 6-MP in maintenance therapy is an important prognostic factor in childhood ALL. Intensification of maintenance treatment by the administration of vincristine/dexamethasone pulses was recently shown to provide no extra benefit. The reduction of maintenance below two years (from the time point of initial diagnosis) was associated with an increased frequency of leukemic relapses. Although it was proven disadvantageous to shorten maintenance treatment, whether or not extended maintenance of up to three years is offering any beneficial effect for particular subgroups in the context of different treatment strategies is not completely evaluated. With regard to the debate on the better thiopurine, three randomized studies compared the toxicity and efficacy of 6-thioguanine (6-TG) with 6- MP in interim maintenance and maintenance therapy of childhood ALL. However, due to the observation of dose-dependent high rates of severe hepatotoxic side effects associated with the application of 6-thioguanine, the current thiopurine drug of choice for maintenance treatment remains 6-MP. 10

11 Table 1. Important prognostic factors a and their approximate incidences in childhood ALL Factor Favorable prognostic factors and their approximate incidence (%) Unfavorable or less favorable prognostic factors and their approximate incidence (%) Age at diagnosis 1 and < 10 years (77%) < 1 year (3%) or 10 years (20%) Gender female (45%) male (55%) White blood cell count at diagnosis < /µl (80%) /µl (20%) Immunophenotype CD10-positive precursor B-cell ALL (83%) CD10-negative precursor B-cell ALL (4%), T-ALL (13%) CNS disease b CNS 1 (80%) CNS 3 (3%), TLP+ (7%) Genetic features c hyperdiploidy (20%), TEL/AML1 positivity (20%) hypodiploidy (1%), t(9;22) or BCR/ABL1 positivity (2%), t(4;11) or MLL/AF4 positivity (2%) Prednisone response d < 1000/µl blood blasts (90%) 1000/µl blood blasts (10%) Early bone marrow response Remission status after induction therapy in the bone marrow (morphologically assessed) Minimal residual disease e in the bone marrow (molecularly assessed ) < 5% blasts (M1) on day 15 of induction treatment (60%) < 5% blasts (M1) after 4 to 5 weeks of induction treatment (98%) < 10-4 blasts after 5 weeks of induction treatment (40%) 25% blasts (M3) on day 15 of induction treatment (15%) 5% blasts (M2 or M3) after 4 to 5 weeks of induction therapy (2%) 10-3 blasts after 12 weeks of treatment (induction and consolidation) (10%) a prognostic factors are treatment dependent and, therefore, the selection presented in the table above cannot be entirely comprehensive; it reflects the current recommendations of AIEOP-BFM ALL study group and the Biology & Diagnosis Committee of the I-BFM Study Group. b CNS1 (puncture nontraumatic, no leukemic blasts in the cerebrospinal fluid (CSF) after cytocentrifugation); CNS3 (puncture nontraumatic, >5 leukocytes/µl CSF with identifiable blasts); TLP+ (traumatic lumbar puncture with identifiable leukemic blasts); a TLP with no identifiable blasts is not an adverse factor; the prognostic impact of CNS2 status (puncture nontraumatic, 5 leukocytes/µl CSF with identifiable blasts) is debated. For cytomorphological examination, CSF samples should be analyzed after cytospin preparation, a method through which cellular components within the CSF are concentrated by centrifugation. c hyperdiploidy defined as the presence of more than 50 chromosomes or a DNA index (the ratio of DNA content in leukemic G0/G1 cells to that of normal diploid lymphocytes) 1.16; hypodiploidy defined by <45 chromosomes; the prognostic value of MLL gene rearrangements other than MLL/AF4 and presence of the E2A/PBX1 fusion transcript are debated. d after 7 days induction with daily prednisone and a single intrathecal dose of methotrexate on treatment day 1. e assessed by molecular genetic techniques or flow cytometry; markers required to have a sensitivity of at least

12 Diagnosis of childhood acute lymphoblastic leukemia The diagnosis of ALL should be made, or reviewed, in a laboratory providing the necessary expertise and adequate technical equipment. A diagnostic setting in a clinical trial environment on acute leukemias requires access to different resources which can be centralized in a single place or be present in a decentralized structure in form of a laboratory network structure. Diagnostic methods The diagnostic work-up of childhood ALL requires cytomorphology with cytochemistry, immunophenotyping, karyotyping, FISH, and molecular genetic analyses of bone marrow and/or peripheral blood leukemic cells. To tailor the intensity of CNS-directed treatment, appropriate assessment of CNS involvement needs to be performed. The definition of CNS disease is explained in Table 1. Cytomorphology The morphologic classification of ALL should be based on the French-American-British (FAB) classification. Cytochemistry may help to assess lineage affiliation (e.g., myeloperoxidase [MPO]-positivity). In the presence of ambiguous morphology and cytochemistry, immunophenotyping should be employed to define lineage. Reference cytomorphology in the context of clinical trials may increase the quality of cytomorphological assessments. Immunophenotyping In the context of ALL, international efforts within the I-BFM Study Group have led to a consensus panel to be assessed for all ALL cases which should include icd3, icd22, icd79a, iigm, ilysozyme, impo as intracellular markers (prefix i stands for intracellular staining; in combination with CD45) as well as CD2 (PE recommended), CD3, CD5, CD7; CD10, CD19, CD20; CD11b, CD11c, CD13, CD14, CD33, CD64, CD65 (available only in FITC), CD117; CD34, CD56, and HLA-DR as surface markers (in combination with CD45). In an integrated approach, the respective lineage-associated MRD-tubes should then be added according to the primary lineage and, in addition, markers for lineage subclassification be added as appropriate (e.g., if T-ALL: CD1a, TCRab, TCRgd, CD4, CD8; if B-ALL: CD15; if B-IV suspected: Kappa and Lambda (on surface after pre-washing or intracellular); if biphenotypic acute leukemia (BAL) according to general panel: CD24, itdt). In case of double-platform approaches, screening tubes must include in all cases the following markers: CD19, CD10, icd79a, icd22, icd3, CD7, impo, and CD45. Permeabilization reagents for intracellular stainings should be applied in a standardized fashion in clinical trial groups. Bone marrow is preferred over peripheral blood for analyses. The latter may be used for assessing the immunophenotype if bone marrow is not available or of poor quantity/quality. There is consensus that lysis preparation is preferred. Additional usage of SYTO-16 or -41 is recommended to distinguish non-nucleated events (erythrocytes) from blasts which is particularly relevant in cases of very low CD45 expression of blasts. Immunological blast gating should be performed based on a 12

13 CD45/SSC backbone strategy and lineage-defined gating wherever appropriate. The percentage of blasts among total nucleated cells of the sample must be recorded. Notably, formal diagnosis of leukemia and treatment indication usually depends on morphologic blast enumeration respecting conventional thresholds (e.g. 25% as per AIEOP-BFM ALL 2009 protocol, or as per current WHO definition which respects a usual threshold of 20%). This threshold requirement is, however, independent of the process of immunophenotyping of the leukemia. Hence, there is no formal cut-off of blast percentage below which the immunophenotype may not be determined the leukemic immunophenotype can be determined as long as the blast population can be clearly distinguished from normal background cells. In such case with a malignant blast percentage <25% by flow cytometry, however, the immunophenotypic report should not anticipate formal diagnosis of ALL but state e.g. malignant lymphoblasts suggestive of ALL with low percentage of blasts. In leukemia with maturing cells in addition to immature blasts (usually not an issue in ALL not relating to B-I to B-III transition; but common in AML with maturation, e.g., FAB M2 or M4), detailed immunophenotype reporting relates only to the population of immature blasts (most frequently defined by CD34 and/or CD117 expression), but the phenotypic peculiarities of the maturing leukemia cells (as far as unambiguously belonging to the leukemia) should be reported in the descriptive summary (including % among all nucleated cells). In case of truly bilineal leukemia, the detailed phenotype should be reported for the larger clone, and the details of the second blast population should be summarized in the descriptive summary (including % among all nucleated cells, lineage and subtype assignment, antigen expression pattern as different from the larger clone). A minimum of nucleated events should be acquired per tube. The definition of antigen distribution and expression levels should employ ratings which use assessment of antigen expression on blasts relative to an appropriate negative population. By determining the degree of overlap (prevalence) of the blast population with the negative control subset, expression is to be rated as either negative (overlap in 90%, i.e. non-overlap in <10% of blasts), weak positive (non-overlap=positivity in 10% to <50% of blasts, i.e. low expression with a major overlap with negative cells or positive in only a minor separate subset of blasts), or strong positive (non-overlap=positivity in 50% of blasts, i.e. all kinds of expression with non-overlap of the majority of blasts). In addition to that a more detailed further description of antigen distribution in combination with expression intensity relative to the appropriate control population is a recommended optional output. This fine-rating should be employed according to adapted Bethesda criteria which make use of a descriptive and semiquantitative scale rather than an exact enumeration procedure. Hence, the focus in interpretation of phenotyping data on the descriptive scale lies more on robustness in daily routine and less on biologic accuracy. Among weak positive distinguish dim positive from partially positive in <50% of blasts (= partially positive 1 ). Among strong positive distinguish medium, bright, heterogeneous or partially positive in 50% of blasts (= partially positive 2 ). 13

14 Flow cytometry for determination of DNA index Determination of DNA index by flow cytometry is frequently employed to indicate the presence of hidden high hyperdiploid clones (not detected by cytogenetics). These poor risk abnormalities are often accompanied by a population comprising a doubling of the hypodiploid clone, which may represent the predominant population. In near-haploid cases the doubled population may be misclassified as classical high hyperdiploidy. As these two abnormalities represent poor and good risk groups, respectively, their accurate distinction is vital and the technique recommended to be employed in the diagnostic wotk-up of clinical trials on treatment of childhood ALL. Cytogenetics and FISH Cytogenetic analysis of bone marrow blasts allows for the identification of significant numerical and structural chromosomal abnormalities in acute leukemias. In childhood ALL, the poor risk group currently comprises patients with t(9;22)(q32;q11), translocations involving the MLL gene, primarily t(4;11)(q21;q23), near-haploidy and iamp21. FISH is one of the most successful techniques to be integrated into routine clinical practice recently. It offers a valuable complementary approach to cytogenetics and one main advantage is that it provides information in addition to the result of the specific test for which it is applied. For diagnostic purposes, FISH should always be carried out in accredited laboratories with a high throughput of samples. Preparation and analysis should always be performed in accordance with national quality assessment schemes. In diagnostic assessment procedures, highest priority should be given to the high-risk abnormalities, in particular: BCR-ABL1; MLL rearrangements, particularly t(4;11)(q21;q23) giving rise to the MLL-AFF1 fusion; iamp21; and near-haploidy. The good risk abnormalities, ETV6-RUNX1 fusion and hidden high hyperdiploidy, are also recommended to be assessed with high priority for testing as they are cryptic at the cytogenetic level. In addition, their presence does not only impact on patient management but the same probes are used for the detection of the poor risk anomalies: iamp21 and near-haploidy. Molecular genetics Molecular genetic techniques have revolutionized leukemia diagnostics and have the advantage of making use of DNA or RNA extracted directly from bone marrow without the need for cell culture to produce metaphases or laborious chromosomal analysis. With increasing numbers of significant chromosomal translocations and their partner genes identified, techniques of reverse transcriptase PCR (RT-PCR) were developed for their detection. The advantage of this approach is that it requires only small amounts of RNA to demonstrate the expression of the fusion transcript arising from the significant translocations. RT-PCR is used as a complementary or alternative technique to cytogenetic analysis in a number of protocols for the rapid detection of specific translocations. It provides a highly sensitive and accurate method, even in a multiplex reaction, for the identification of specified abnormalities among patients with a failed cytogenetic result or variant and complex chromosomal rearrangements. One major disadvantage is that cooperating abnormalities cannot be simultaneously identified. Recently, Multiplex Ligation- 14

15 dependent Probe Amplification (MLPA) became an important technique in the molecular work-up of specific losses and gains of genetic material in acute leukemias. Due to the integrated use of cytogenetic and molecular genetic techniques in a complementary way to assess the most relevant recurrent genetic aberrations, both technical approaches are described below in context of those recurrent aberrations recommended to be assessed. 15

16 BCR-ABL1 fusion Routine cytogenetic analysis and RT-PCR readily detects the t(9;22)(q34;q11)/bcr-abl1 fusion in the majority of patients. Those with a normal or ill-defined karyotype, a failed cytogenetic result or a discrepancy between the cytogenetic and molecular result should be tested by FISH. The use of a dual colour, dual fusion probe is recommended. The detection of two fusion signals indicates a positive result with both fusion products present, for which the false positive rate is extremely low. At the same time these probes will show the presence of a second copy of the Philadelphia (Ph) chromosome, which is a characteristic feature of Ph-positive ALL and indicate deletions from the derivative chromosome 9 involving the reciprocal ABL1-BCR fusion. These probes will also detect the NUP214-ABL1 fusion, found as a secondary change in T-ALL and other rarely occurring ABL1 translocations, which may be responsive to imatinib treatment. It could be argued that the presence of the BCR-ABL1 fusion is the significant event; therefore FISH or RT-PCR could replace cytogenetic analysis for the detection of this abnormality. However, for the purpose of monitoring patients following treatment with imatinib, it is important to be aware of secondary or chromosomal changes acquired during therapy. MLL fusion For risk stratification in current childhood ALL treatment trials, the translocation t(4;11)(q21;q23)/mll-aff1 fusion is classified as high-risk, although the prognosis of the other MLL partners may become significant in the future. As for BCR-ABL1, the t(4;11)(q21;q23)/mll-aff1 fusion is readily detectable by cytogenetics and RT-PCR. Multiplex molecular approaches may be applied for the detection of this fusion in conjunction with other common MLL translocations. A dual colour breakapart probe provides a highly effective approach for the detection of all chromosomal rearrangements involving MLL by FISH, including variant t(4;11) translocations and cryptic insertions. This FISH probe also detects 3 deletions associated with MLL translocations. This confirms, that as for BCR-ABL1, those samples with a normal or ill-defined karyotype, a failed cytogenetic result or a discrepancy between the cytogenetic and molecular result should be tested by FISH. ETV6-RUNX1 fusion This fusion arises from the translocation t(12;21)(p13;q22). Although the fusion transcript can be detected by RT-PCR, the translocation is kryptic at the cytogenetic level. An extra signal or dual colour fusion probe is usually used for detection by FISH. The advantage of the FISH approach is that fusions arising from rare variant breakpoints are also identified, which may not be detected by RT-PCR. In ETV6-RUNX1 positive cases, FISH provides information on the status of the second ETV6 homologue: whether it is retained or deleted, as well as indicating the number of RUNX1 and fusion signals. Deletion of the second ETV6 allele and additional fusion signals are important secondary events in this ALL subtype, which may be linkable to outcome in future stuies and, therefore, are recommended to be assessed. 16

17 Hidden high hyperdiploidy In ETV6-RUNX1-negative cases, additional RUNX1 signals are frequently observed in interphase. In high hyperdiploid cases, these usually correspond to the additional copies of chromosome 21 seen by cytogenetic analysis, which are a characteristic finding in this cytogenetic subgroup. Thus it can be inferred that additional RUNX1 signals seen in interphase cells of patients with normal of failed cytogenetic results, when negative for the three abnormalities: ETV6-RUNX1, BCR-ABL1 fusions and MLL rearrangements, represent additional copies of chromosomes 21, likely within a hidden high hyperdiploid clone. When using the ETV6-RUNX1 probe, the ETV6 signals will indicate the number of copies of chromosome 12, acting as an internal control for RUNX1. Chromosomes 9, 11 and 22 are less frequently gained than chromosome 21 within high hyperdiploid karyotypes, therefore the relative numbers of signals seen with the BCR-ABL1 and MLL probes will rule out the presence of triploid or tetraploid populations. Subsequent FISH hybridisation with centromeric probes specific for the simultaneous trisomies of chromosomes 4, 10, and 17 is recommended to identify the good risk group of high hyperdiploid patients as defined by the Children s Oncology Group. iamp21 iamp21-positive samples are negative for the fusion, while in addition to the two normal copies of the ETV6 signal, show multiple RUNX1 signals (3 or more additional signals) with the ETV6-RUNX1 probe. In metaphase, one signal is located to the normal chromosome 21, while the others are seen in tandem duplication along an abnormal chromosome 21. In interphase, the signals are clustered together, except for one signal representing the normal chromosome 21 which is usually located apart. The observation that the commonly amplified region always includes the RUNX1 gene, confirms that probes directed to RUNX1 provide a reliable detection method for this abnormality. It was shown by FISH, using a probe specific to the 21q subtelomeric region, and acgh that the abnormal chromosome 21 frequently shows a deletion of the subtelomeric region. Alternatively a normal pattern is seen. In rare cases where a 21q subtelomere is gained, the level of RUNX1 amplification is always greater. This means that accurate distinction of iamp21 from the gains of whole chromosomes 21 or isochromosomes 21 in interphase can be made by calculating the ratio of RUNX1 to subtelomeric signals. When whole chromosomes 21 are gained or an isochromosome 21 is present, the number of RUNX1 and 21q subtelomeric signals will be equal in number, with a ratio of 1. In iamp21 the number of RUNX1 signals will always exceed the number of signals from 21q, with a ratio greater than 1. Thus application of these two probes directed to chromosome 21 provides an accurate interphase FISH diagnostic test for iamp21. Hidden near haploidy Although near-haploidy is classified as a high risk abnormality in childhood ALL, it is rare. Its characteristic features are the gain of specific chromosomes onto the haploid chromosome set and, in the majority of patients, the presence of a population of cells with an exact doubling of this chromosome number. This doubled population is often the predominant one. The majority of near-haploid cases are identified from cytogenetic analy- 17

18 sis or their presence is inferred by the finding of the doubled population. This mixture of near-haploid and doubled populations makes FISH detection of hidden populations difficult. Probes specific for chromosomes which would be expected to be present as a single copy in the near-haploid clone need to be applied in parallel with probes specific for chromosomes which would be expected to have four copies within the doubled population. As chromosomes 3, 7, 9, 11 and 22 are rarely gained in near-haploidy, centromeric probes to chromosomes 3 and 7, in addition to the BCR-ABL1 and MLL probes should identify the near-haploid clone. In these cases flow cytometry to measure the DNA index of the two populations may provide a more appropriate and accurate detection method. IGH@ translocations Translocations involving IGH@ at 14q32 are emerging as a significant subgroup in childhood ALL. Dual colour breakapart probes can identify these translocations and in metaphase will identify their partners allowing studies to inform of their biological and prognostic significance. This probe will also indicate gains of chromosome 14 in hidden high hyperdiploid cases. t(17;19)(q22;p13) The translocation, t(1;19)(q23;p13)/tcf3-pbx1, fusion is now regarded as a standard risk abnormality in childhood BCP-ALL. It has been shown that not all translocations involving t(1;19)(q23;p13) at the cytogenetic level involve this fusion. This has been demonstrated by FISH using a breakapart probe to TCF3. In addition this probe is able to detect the presence of variant translocations involving TCF3. Of note is the t(17;19)(q22;p13) with the TCF3-HLF fusion, which even on current therapies has a very dismal outcome. Although this translocation is rare, it is usually visible by cytogenetic analysis and primers exist for its detection by RT-PCR. However, in cases with failed and normal cytogenetic results FISH with the TCF3 probe would be valuable to ensure that all cases are identified in view of the high risk associated with this abnormality. CDKN2A deletions FISH with a commercially available probe to CDKN2A provides a reliable method for the detection of deletions around this gene. Although this is known to be an important secondary abnormality in both BCP- and T-ALL, its relative incidence among the different cytogenetic subgroups and its relative prognostic significance remain unclear. Routine FISH screening within different treatment protocols will be of value to inform the design of future treatment protocols. IKZF1 and CRLF2 As further evidence is accumulating regarding their potential inclusion in future riskstratification algorithms, deletions of IKZF1 and the PAR region on chromosome X including CRLF2 is recommended to be assessed by MLPA. In addition, other techniques for the assessment of aberrant IKZF1 and/or CRLF2 (e.g., real-time quantitative PCR, fusion gene expression bei RT-PCR, and cell surface expression by flow cytometry are strongly encou- 18

19 raged to be evaluated in clinical trials on ALL as their exact prognostic value is yet to be defined. TPMT To prevent prolonged hematopoietic toxicity for TPMT deficient individuals upon exposure towards thiopurine drugs (6-MP, 6-TG), all patients with ALL should be genotyped for the most common variant TPMT alleles before introduction of thiopurine medications (TPMT*2, and *3A, *3C). Treatment response evaluation Response to treatment is the most important indicator of outcome in childhood ALL. Most study groups evaluate treatment response by cytomorphology in the bone marrow after treatment days 8, 14, or 15, and after the end of induction therapy (e.g., days 28 or 33). Measures of cytomorphological treatment response are relevant for stratification on many protocols. The difficulties of assessing hypoplastic bone marrow are well acknowledged here. Besides cytomorphological response assessment, MRD analyses by mainly two technical approaches, DNA-based PCR of clone-specific Ig or TCR gene rearrangements or assessment of an aberrant immunophenotype by flow cytometry (see also section on immunophenotyping), have become the main stratification elements in childhood ALL. It has to be stressed that MRD analyses should be incorporated in the context of clinical trials and, if appropriate, for treatment stratification. However, the latter can only be recommended in association with the existing international study groups on MRD to assure comparable and quality-controlled procedures. Biobanking Biobanking of spare diagnostic material in the context of clinical trials according to standardized operating procedures is recommended. Consent issues should be addressed already in the trial protocols. 19

20 Recommendations for the diagnosis of childhood acute myeloid leukemia (AML) General information Childhood AML is rarer than ALL and accounts for up to 20% of childhood leukemias. The annual incidence rate of this disease varies around a half new case per 100,000 children younger than 15 years. Mainly through joint efforts of national cooperative trial groups, survival rates during the last decades increased to 70%. These results are reached by application of intensive multiagent chemotherapeutic regimens and in specific patient subgroups additional radiotherapy and/or hematopoietic stem cell transplantation (HSCT). Treatment regimens basically consist of three main components: (1) an induction period; (2) consolidation/intensification elements; and (3) central nervous system (CNS) preventive therapy. Treatment duration is approximately 18 months. Certain clinically and biologically distinct patient subgroups with AML receive treatment specifically adapted to their needs (e.g., ATRA-based maintenance treatment in promeyelocytic leukemia, adapted chemotherapeutic protocols for patients with Down syndrom). Clinical presentation and diagnosis Similar to ALL, the initial clinical presentation of a child with AML largely depends on the extent of the leukemic infiltration of the bone marrow and extramedullary sites. The diagnosis of ALL is usually made by cytomorphological and cytochemical examination of a bone marrow aspirate. CNS involvement (CNS3 status) is diagnosed by the presence of blasts in the cerebrospinal fluid (CSF) or if intracerebral infiltrates are detected by crosssectional radiological imaging. Initial diagnostics are complemented by flowcytometry-based immunophenotyping to gain information on the blasts expression of myeloid differentiationassociated antigens as measured by the reactivity to specific monoclonal antibodies. In addition, a combined approach using cytogenetic and molecular genetic techniques is used for the detection of genetic aberrations, such as non-random recurrent structural and, less frequently, numerical chromosomal aberrations (e.g., different MLL rearrangements, t(8;21), inv(16), t(15;17), t(7;12), t(6;9), monosomy 7). Molecular-genetic techniques are employed to detect additional somatic aberrations (e.g.; FLT3 internal tandem duplications (ITD), mutations of WT1, N-RAS, K-RAS, PTPN11, and c-kit). Prognostic factors and risk-adapted treatment Large research efforts on the clinical and biological aspects of AML have led to the identification of several prognostic factors mostly genetic aberrations which are used for risk-adapted therapeutic treatment stratification in the clinical management of childhood AML. Response to the first course of treatment and genetics are the most important prognostic factors and are instrumental for most of today s risk group stratification strategies. Most study groups evaluate treatment response morphologically in the bone marrow after the first and second induction course (e.g.; treatment days 15 and 28). MRD in AML can be analyzed by morphology, immunophenotyping, and quantification of molecular aberrations as well as specific gene expression. The definite clinical implications of the different MRD measurements are still under debate. 20

21 Remission induction treatment Contemporary treatment approaches for childhood AML mostly apply two courses of induction therapy which routinely consist of three drugs an anthracycline (eg.; daunorubicin or idarubicin), cytarabine, and etoposide or 6-TG and lead to achievement of complete remission in more than 85% of patients. Consolidation/intensification treatment Eradication of residual leukemic blasts in patients who are in remission by morphologic criteria is the primary aim of consolidation/intensification treatment. In childhood AML this is achieved by application of chemotherapy courses consisting of drug combinations similar to those given during induction. The value of applying more than three consolidation courses is unclear. Central nervous system-directed therapy CNS-directed therapy has become a prerequisite for successful treatment of childhood AML. Similar to ALL, preventive CNS treatment is given to all patients and traditionally consisted of intrathecal chemotherapy (cytarabine or methotrexate, or triple drug combinations of cytarabine, methotrexate, and hydrocortisone). Depending on the protocols, intrathecal chemotherapy was given alone or combined with cranial irradiation. During the last decade, however, systemic chemotherapy regimens with adequate CNS intensity together with the lack of evidence of superiority for cranial irradiation led to abandonment of irradiation by all major study groups. A CNS3 status in AML is not a crucial factor within the AML risk group stratification because it does not affect overall survival. However, those with CNS involvement relapse more frequently in the CNS and, therefore, require intensified intrathecal therapy. Although most study groups have added CNS irradiation to the regimen of these patients, recent observations suggest that frequent intrathecal chemotherapy combined with intensive systemic chemotherapy may yield similar results. Allogeneic hematopoietic stem cell transplantation The role of allogeneic HSCT in first complete remission is not clearly defined, but HSCT in second CR is generally considered. 21

22 Diagnosis of childhood acute myeloid leukemia The diagnosis of AML should be made, or reviewed, in a laboratory providing the necessary expertise and adequate technical equipment. A diagnostic setting in a clinical trial environment on acute leukemias requires access to different resources which can be centralized in a single place or be present in a decentralized structure in form of a laboratory network structure. Diagnostic methods The diagnostic work-up of childhood AML requires cytomorphology with cytochemistry, immunophenotyping, karyotyping, FISH, and molecular genetic analyses of bone marrow and/or peripheral blood leukemic cells. To tailor the intensity of CNS-directed treatment, appropriate assessment of CNS involvement needs to be performed. The definition of CNS disease is explained in Table 1. Cytomorphology AML classification is based on lineage-associated phenotype (undifferentiated, myeloid, monoblastic, erythroblastic, or megakaryoblastic) and defined according to FAB. Cytochemistry confirms lineage affiliation and classifies myeloid (MPO-positivity) and monoblastic differentiation (nonspecific esterase-positivity). In the presence of ambiguous morphology and cytochemistry, immunophenotyping may further support the lineage definition. Acute megakaryoblastic leukemia (AMKL, FAB M7) and minimally differentiated AML (FAB M0) have to be confirmed by immunophenotyping, although the former may show typical morphologic features. The presence of myelofibrosis frequently associated with acute megakaryoblastic leukemia, and consequent sampling problems may lead to an underestimation of blasts by both morphology and immunophenotyping. In case of a low blast count (< 20%), repeated bone marrow sampling, including biopsy, needs to be performed. Immunophenotyping At present, there is no standardization of antibody panels used for immunophenotyping among the large trial groups. However, upcoming standards suggest the use of multicolor monoclonal antibody combinations that include CD45 to enable optimal gating and analysis of the blast population within the complex context of residual hematopoiesis. In addition, these procedures might improve flow cytometry techniques for assessing MRD in childhood AML. Another methodologic transition concerns the interpretation of immunophenotypic expression data. Rigid cut-off points for determining marker positivity (e.g., 10% or 20%) are being replaced by biologically more meaningful semiquantitative estimates of blast population appearances compared with those of normal populations. The mandatory minimal panel required to fulfill WHO and EGIL criteria for AML includes CD34, CD117, CD11b, CD11c, CD13, CD14, CD15, CD33, CD64, CD65, impo, i-lysozyme, CD41, and CD61; and for MPAL: CD19, icd79a, icd22, CD10, and icd3. 22