SEARCH

SEARCH BY CITATION

Keywords:

  • acute lymphoblastic leukaemia;
  • childhood;
  • prognostic factors;
  • chemotherapy;
  • minimal residual disease

Abstract

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Conclusion
  7. Acknowledgments
  8. References

Published studies of the prognostic value of the early response to induction treatment in childhood acute lymphoblastic leukaemia (ALL) were analysed. Three criteria were used to judge the early treatment response: persistence of peripheral blasts (PPB) or of bone marrow blasts (PBMB) during induction therapy and minimal residual disease (MRD) after completion of induction therapy. Studies with more than 50 patients, published between 1980 and 2000, were reviewed. Among 13 659 distinct articles published on ALL, we identified only 43 applicable studies. Within- and between-laboratory variations were evaluated in only one study. Treatment modalities differed among, and sometimes within, studies. The cut-off points used in the statistical analyses were never discussed, and in many studies appeared to be selected after multiple tests. The proportion of missing data was > 30% in almost all studies of MRD, as a result of technical difficulties and not missing samples. PPB and PBMB were associated with shorter survival in, respectively, 13 out of 14 and 15 out of 16 studies. Detection of MRD was associated with poor outcome in 12 of the 13 studies. Because none of the parameters used to measure the early response to induction therapy for childhood ALL have been properly assessed as prognostic factors, we conclude that they should be considered only as candidate prognostic indicators pending more thorough studies.

In the absence of treatment, acute lymphoblastic leukaemia (ALL) is invariably fatal. The prognosis depends on the response to treatment. This response depends on the characteristics of the disease, its sensitivity to the drugs used, the patient's general condition (nutrition, infections, etc.), individual pharmacological characteristics and compliance with treatment. Many prognostic factors have been identified in childhood ALL, but most of the prognostic information is contained in four characteristics, namely age, sex, leucocytosis and cytogenetics (Donadieu et al, 1998). The prognostic value of these factors has been demonstrated in numerous studies and they serve as the basis for the most widely agreed patient classification (Smith et al, 1996). However, they have two important weaknesses. First, they have low explanatory power, and many relapses occur in patients who do not have unfavourable characteristics. Second, they have low heuristic or cognitive power. For example, there is a clear statistical link between hyperleucocytosis and poor prognosis (this has been known since the 1950s when treatments were clearly less effective) (Tivey, 1952), but there is still no biological explanation for this link.

New prognostic factors have thus been sought for childhood ALL. Most relevant studies have aimed at establishing a link between the early response to treatment and the prognosis. The persistence of circulating blasts after 7–10 d of therapy is commonly referred to as steroid resistance, even though steroid-based regimens may contain other agents. The persistence of bone marrow blasts after 7, 14 or 21 d after a combination of several treatments is commonly termed chemoresistance. Studies of residual disease after the completion of induction treatment are based on a variety of techniques (Campana & Pui, 1995), all of which are more sensitive than cytological examination; detection limits of one cell in 102−106 have been claimed. Two types of technique are available currently. Detection with immunological markers identifies cells with the same immunophenotypic characteristics as the initial bone marrow disease. Detection by molecular biology is based on two types of molecular probe. The first type of molecular probe is based on chimaeric genes which reflect molecular events in the early stages of leukaemogenesis. The TEL-AML 1, E2A-PBX, MLL AF4, BCR-ABL, P 16 MTS1 and SIL TAL 1 genes are found at diagnosis in nearly 35% of childhood leukaemias. The second type of molecular probe is based on specific rearrangements of either the immunoglobulin genes or the genes coding for the different T cell receptor (TCR) chains (α, β, γ and δ).

The evaluation of new prognostic factors must satisfy basic quality criteria, which were described by Simon & Altman (1994). During an extensive review of the literature, we examined previous studies of prognostic factors in childhood ALL from a methodological standpoint.

Materials and methods

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Conclusion
  7. Acknowledgments
  8. References

All articles on childhood acute lymphoblastic leukaemia (ALL) published between 1980 and the end of 2000 that considered the persistence of circulating blasts, the persistence of bone marrow blasts or minimal residual disease (MRD) were taken into account. The Medline data base was scanned with the key words ‘Acute lymphoblastic leukaemia’ and ‘Minimal residual disease’ and ‘Prognostic factors’. The bibliography of each paper was used to identify additional papers.

The articles were classified into three categories according to the factor studied, i.e. persistence of circulating blasts, persistence of bone marrow blasts and MRD, evaluated at the end of induction therapy by means of either immunophenotyping or molecular biology.

Only articles including > 50 patients in the survival analysis were taken into account. Where one team had published several updates of the same study, we used the most recent paper. Among 13 659 published articles about ALL and 413 focusing on MRD, 43 studies fulfilled the criteria for this study.

Each paper was examined using the criteria proposed by Simon and Altman (1994) for the assessment of new prognostic factors (Table I). Points 1 and 2 deal with the reproducibility of the assays. Points 3 and 4 deal with the population studied. Points 5–11 deal with the statistical techniques. To determine if a multivariate prognostic analysis had adequate power, we computed the ratio of the number of events to the number of potential prognostic factors. According to Harrell et al (1985), this ratio should be > 10.

Table I.   Guidelines for phase III prognostic studies (Simon & Altman, 1994).
1Within- and between-laboratory reproducibility of assays should be documented
2Laboratory assays should be performed blinded to clinical data and outcome
3A clear inception cohort of patients should be assembled with few (e.g. < 15%) patients unassessable because of missing material or data. The referral pattern and eligibility criteria should be described so that the generalizability of the results can be evaluated
4Treatment (or absence of treatment) should be standardized or randomized and accounted for in the analysis and interpretation
5Hypotheses to be tested should be stated in advance. The hypotheses should include specification of end points, cut-off values of prognostic variables, subsets of patients, treatment, standard prognostic factors or classifications relevant to the hypotheses
6The number of patients and number of ‘events “ should be sufficiently large so that statistically reliable results can be obtained. Statistical power calculations that incorporate the number of hypotheses to be tested and the appropriate subset of patients (e.g. node-negative) for each hypothesis should be described
7Analyses should be tested whether new assays add predictiveness once outcome is adjusted for the effect of standard prognostic factors
8The analyses should be adjusted for the number of hypotheses to be tested
9Analyses should be based on pre specified cut-off values for prognostic variables or cut-offs should avoided
10Confidence intervals should be provided to indicate the uncertainty in estimates
11Claims of subset-specific treatment effects should be documented by a test of the single global null hypothesis that there is no treatment specificity involving any subsetting variables

Results

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Conclusion
  7. Acknowledgments
  8. References

Persistence of peripheral blasts

Peripheral blasts are counted after a short initial treatment course composed of steroids and various other drugs. A total of 14 studies (Riehm et al, 1986; Rautonen et al, 1988; Reiter et al, 1994; Arico et al, 1995; Gajjar et al, 1995; Thyss et al, 1997; Schrappe et al, 1998, 2000; Conter et al, 2000; Harms & Janka-Schaub, 2000; Toyoda et al, 2000; Tsuchida et al, 2000; Vilmer et al, 2000) reported in 13 papers were identified (Table II).

Table II.   Persistence of peripheral blasts (PPB) main results.
   Treatment before assessment of peripheral blasts†
Study Inclusion criteria* nS (mg/m2/d)VCR (nb of pulses)DNR (mg/m2)Asp (1000/m2)day of IT MTX
BFM 83 (Schrappe et al, 2000)65360*7000d1
Helsinki (Rautonen et al, 1988)16760*9100d1
BFM 86 (Schrappe et al, 2000)99860*7000d1
BFM 90 (Schrappe et al, 2000)217860*7000d1
AEIOP 8703 (Arico et al, 1995)T ALL5440*7000d1
AEIOP 8803 (Arico et al, 1995)T ALL4910–60*7000d1
AEIOP 88 (Conter et al, 2000) 43810–60*7000d1
AEIOP 91 (Conter et al, 2000) 126710–60*7000d1
St Jude XI (Gajjar et al, 1995) Peripheral blast at diag. > 029740*712530d1§
EORTC 58881 (Vilmer et al, 2000) 206560*7000d 1 −7
BFM–AEIOP (Schrappe et al, 1998)t(9; 22)6110–60*7000d 1
COALL 92(Harms & Janka-Schaub, 2000) 538003600
TCCSG L89–12 (Tsuchida et al, 2000) 41860*70000
TCCSG L92–12 (Toyoda et al, 2000)34760*70000 

Reproducibility of the assay No reproducibility studies of peripheral blast assays have been published. Centralized review of blood smears was mentioned in the German Berlin–Frankfurt–Münster (BFM) studies and the measure was described as reproducible (Schrappe et al, 1996). Initial treatment was rarely the same in the different studies. The German BFM group (Reiter et al, 1994; Schrappe et al, 2000) measured peripheral blasts after 7 d of prednisone treatment at a dose of 60 mg/m2/d, combined with an intrathecal methotrexate injection on the first day of steroid administration. In the AEIOP (Associazione Italiana Ematologia Oncologia Pediatrica) 8803 study (Arico et al, 1995), prednisone was started at 10 mg/m2, and increased to 60 mg/m2 as soon as possible before d 7, whereas in the AEIOP 8703 study (Arico et al, 1995) the steroid dose was 40 mg/m2. The EORTC (European Organization for Research and Treatment of Cancer) team (Thyss et al, 1997) showed the very clear influence of the day of intrathecal methotrexate injection on the proportion of resistant patients, which was 10% when intrathecal methotrexate was delivered on d 1, 23% when it was administered between d 2 and d 6 and 30% when it was administered after d 6. In the EORTC and BFM studies, the intrathecal injection contained only methotrexate, whereas a steroid and cytarabine were added in other co-operative studies. The day on which resistance was evaluated varied from d 7 to d 10.

Population All the studies were based on data collected during prospective therapeutic trials. Some included all patients with ALL. Two studies (Gajjar et al, 1995; Thyss et al, 1997) excluded patients with no peripheral blast cells at diagnosis. Two studies were restricted to patients with T-cell leukaemia (Arico et al, 1995), or with the t(9; 22) translocation (Schrappe et al, 1998). The percentage of missing resistance data ranged from 0 to 7%, except for one study where it reached 50% (Harms & Janka-Schaub, 2000).

Statistical techniques Two studies (Rautonen et al, 1988; Gajjar et al, 1995) defined resistance as > 0 blasts/µl, whereas all others used a cut-off point value of 1000/µl. The cut-off point of 1000 blasts/µl appears to result from an exploratory analysis of different cut-off points (Riehm et al, 1986). In some studies the definition of persistence of peripheral blasts was stated in advance (Reiter et al, 1994; Schrappe et al, 2000), whereas in the others (Riehm et al, 1986; Rautonen et al, 1988; Arico et al, 1995; Gajjar et al, 1995; Thyss et al, 1997; Schrappe et al, 1998) it was established retrospectively. Two of the four multivariate analyses (Reiter et al, 1994; Gajjar et al, 1995) had an adequate number of events and two (Rautonen et al, 1988; Schrappe et al, 1998) had too few events to study 19 and 5 covariates respectively.

Comments on the results All but one of the studies showed a shorter event-free survival (EFS) in patients with persistence of peripheral blasts (after various times and therapies) than in patients without persistence of peripheral blasts (Table II). The only negative study measured the persistence of blasts after a daunorubicine pulse instead of a steroid-based regimen (Harms & Janka-Schaub, 2000), The four multivariate analyses (<Rautonen et al, 1988; Reiter et al, 1994; Gajjar et al, 1995; Schrappe et al, 1998) demonstrated an independent value of this prognostic factor (relative risk in resistant patients versus others, respectively, 5·2, 1·9, 2·9 and 6·7).

Persistence of bone marrow blasts

Table III.   Persistence of bone marrow blasts (PBMB): main results. Multicentre studies except for St Jude XI.
   Treatment before assessment of PBMB‡
StudyInclusion criteria* nS (mg/m2/d)VCR (no. of pulses)DNR (mg/m2)C (mg/m2)Asp (1000/m2)Number of IT MTX
CCG 141 (Miller et al, 1983)+ FAB 38804013200 vs. 1300 for HR301
CCG 160 (Miller et al, 1989)29874013200361
CCG 193 P (Gaynon  et al, 1990, 1988)HR209607125018
CCG 105 (Tubergen et al, 1993)SR16064013 vs. 60*1320 vs. 500361
BFM 90 (Schrappe et al, 2000)217860*141300101
UKALL X (Eden et al, 2000)161240*132900361
CCG123 (Steinherz et al, 1996)HR67860*61120 vs. 251200 vs. 00
CCG123 (Steinherz et al, 1996)HR67860*132120 vs. 501200 vs. 00
UKALL XI (Lilleyman  et al, 1997)141940*710 vs. 900181
FRALLE 83 (Donadieu  et al, 1998)LR9640*132 VDS vs. 2 VCR000
FRALLE 83 (Donadieu  et al, 1998)SR and LR529100*51 VDS vs. 1 VCR24060010 (U/kg)
FRALLE 87/89 (Donadieu  et al, 1998)891100*51DNR 240 vs. RBZ 480600250
St Jude XI (Pui et al, 2000)35840*13250060
St Jude XII (Pui et al, 2000)18840*13250060
DCLSG ALL6 (Kamps  et al, 2000) 1906 mg/m2 DXM*112000
Vienna group  (Panzer-Grümayer et al, 2000)22360*141300101

The persistence of bone marrow blasts was measured by cytological bone marrow smear analysis at various times during induction therapy. The idea of assessing the response to initial treatment dates back to the first effective therapies for ALL (Jacquillat et al, 1973), but most data come from the Children's Cancer Group (CCG) trials (Miller et al, 1983, 1989; Gaynon et al, 1990; Tubergen et al, 1993; Steinherz et al, 1996). In addition to these studies, the St Jude (Rivera et al, 1991; Pui et al, 2000), United kingdom Acute Lymphoblastic Leukaemia (UKALL, Lilleyman et al, 1997; Eden et al, 2000), French Acute Lymphoblastic leukaemia (FRALLE, Donadieu et al, 1998), BFM (Schrappe et al, 2000), Dutch (Kamps et al, 2000) and Viennese (Panzer-Grümayer et al, 2000) groups examined bone marrow blast persistence in studies involving usually more than 200 patients.

Reproducibility of the assays Centralized bone marrow slide review was mentioned in three studies (Miller et al, 1989; Lilleyman et al, 1997; Schrappe et al, 2000) and the results were reported in one (Lilleyman et al, 1997). Inter-observer agreement was > 90% when there were < 5% or > 80% of blasts. Inter-observer and intra-observer variations were larger when there were between 5 and 80% blasts. The treatment received before the evaluation of bone marrow blast persistence was highly variable, with differences in the dose and type of steroids, the number of vincristine pulses, and combination with an anthracycline (mainly daunorubicin, from 25 to 240 mg/m2) and asparaginase (cumulative dose ranging from 18 000 U/m2 to 250 000 U/m2). The intrathecal injections consisted of methotrexate alone or in combination with steroids and cytarabine. The day of the assessment also varied from d 7 to d 20. The effect of the timing of the assessment (relative to the first day of treatment) was studied by the CCG (Steinherz et al, 1996). The percentage of resistant patients depended on the timing, and tended to be lower on d 20 than on d 7. These differences were such that there were practically no two studies with the same definition; as a result, the percentage of patients with persistence of bone marrow blasts ranged from 2% to 73% according to the study (Lilleyman et al, 1997; Donadieu et al, 1998).

Population All the relevant studies were based on data collected during prospective therapeutic trials. Some studies included all patients, whereas others were restricted to either high-risk (Gaynon et al, 1990; Steinherz et al, 1996) or standard-risk patients (Tubergen et al, 1993). Data were missing for 15–69% of patients in most studies, and for < 1% of patients in two studies. In some studies, patients with persistence of bone marrow blasts received different therapies from other patients.

Statistical techniques Persistence of bone marrow blasts was defined as > 5%, 20% or 25% of bone marrow blasts, depending on the study, and three studies divided the population into three groups using cut-off points of 5% and 25%. The cut-off points were clearly stated in advance in only a few studies (Gaynon et al, 1988; Miller et al, 1989; Tubergen et al, 1993; Steinherz et al, 1996).

Univariate analyses were reported in each study. Multivariate analyses were reported in six studies (Miller et al, 1983, 1989; Gaynon et al, 1990; Chessells et al, 1995; Steinherz et al, 1996; Donadieu et al, 1998) and the number of events appeared to be adequate in all but two studies (Gaynon et al, 1990; Steinherz et al, 1996) (Table III).

Comments on the results Despite the diverse definitions of persistence of bone marrow blasts, the results of all the univariate prognostic analyses agreed, except for one study (Kamps et al, 2000). The UKALL IX study (Lilleyman et al, 1997) showed the influence of the overall therapy: the proportion of patients with > 25% of blasts on d 7 was 73% without daunorubicin and 39% after daunorubicin (45 mg/m2 on d 2). There was no difference in long-term relapse-free survival between the two groups with or without daunorubicin but, in each group of patients, those with persistence of bone marrow blasts had a worse outcome than other patients. The multivariate analyses always showed an independent, unfavourable prognostic value of bone marrow blast persistence.

Residual disease

Three techniques were used to quantify residual disease: immunophenotyping and two molecular methods to amplify a nuclear transcript, which was either a fusion transcript (a molecular event corresponding to translocation between chromosomes) or a somatic rearrangement of the immunoglobulin genes (characteristic of the B-lymphocyte lineage) or the T-cell receptor genes (characteristic of the T-cell lineage).

Residual disease can only be measured in highly specialized laboratories. This explains why, in the two multicentre studies (Cave et al, 1998; Van Dongen et al, 1998), the tests were performed on only 246 of 900 patients and 240 of 625 patients respectively.

The definition of residual disease depends very closely on the probes chosen and the initial treatment received. No two studies used the same technique or the same treatment, i.e. the same definition. The reproducibility of each technique was never studied in > 10 patients (Pongers-Willemse et al, 1999). The information on MRD was missing for 10–77% of patients, either because there was no bone marrow available at the end of induction therapy or because no rearrangement had been identified.

Residual disease studied by immunophenotyping

The markers studied by means of immunophenotyping are membrane proteins potentially characteristic of the abnormal clone present at diagnosis. Most articles either solely described the technical approach (n = 31), or reported survival data on < 50 patients (n = 9), and were therefore excluded from this analysis. Two papers reported clinical data on > 50 patients (<Farahat et al, 1998; Coustan-Smith et al, 2000) but the data were too complex for tabulation.

Reproducibility of the assay

The reproducibility of the assay was never studied. Variations in the expression of surface proteins on the malignant clone, thought to be characteristic of the disease, is one of the technical difficulties encountered. In a study of 40 patients compared at diagnosis and relapse (Van Wering et al, 1995), antigen expression varied in nearly 75% of cases. In the two clinical papers considered here, the probes used to follow the patients were TdT, CD10, CD19 (Farahat et al, 1998) or the same probes plus CD19/7·1; CD19/p53; CD34, CD13, CD33, CD65, CD21, CD56, IgM (Coustan-Smith et al, 2000). The probes were tested singly or in combination. Lastly, residual disease was measured in terms of the number of molecules per cell in one study (Farahat et al, 1998), and the ratio of pathological cells among all cells examined in the other study (Coustan-Smith et al, 2000).

Population

One study (Coustan-Smith et al, 2000) was based on data collected during a single-centre prospective therapeutic trial, whereas the other (Farahat et al, 1998) was based on retrospective data. This second study included 15 adults among the 53 patients, and involved a variety of treatments.

In one study (Coustan-Smith et al, 2000), a leukaemia-associated immunophenotype was identified in 58% of patients in an unselected cohort. After induction, information on MRD was evaluated in, respectively, 48% (Farahat et al, 1998) and 42% (Coustan-Smith et al, 2000) of the patients.

Statistical techniques

The cut-off points used to define positivity were different for each marker and also differed between the studies: TdT < 105; CD10 > 5 × 104; CD19 > 1·1 × 104 (molecules/cell) in one (Farahat et al, 1998) and a detection limit of 1 cell in 10 000 in the other (Coustan-Smith et al, 2000). Furthermore, the cut-off points were not stated in advance.

Univariate analyses were performed in the two studies and multivariate analysis was performed in the second (Coustan-Smith et al, 2000).

Comments on the results

Residual disease at the end of induction therapy was associated with an unfavourable prognosis in both univariate analyses. In one study, the prognostic value of this variable persisted after adjustment for age, leucocytosis and cytogenetics (Coustan-Smith et al, 2000).

Residual disease detected with fusion transcript markers None of the fusion transcript studies involved more than 50 patients.

Residual disease detected with molecular rearrangement markers MRD is detected by looking for clonal rearrangement of genes characteristic of lymphoid cells. In physiological conditions, lymphocytes possess a very extensive repertoire of rearrangements between these different genes, whereas leukaemia is characterized by an abnormal clone that can be identified and quantified. If, in theory, all cases of lymphoid leukaemia are caused by a single lymphocyte clone, it should be possible to identify the disease with these markers in each patient. We studied 179 articles published since the early 1980s and many literature reviews (Campana & Pui, 1995). More than 70 publications gave clinical data, but only 11 involved more than 50 patients and were thus considered in the present analysis (Table IV).

Table IV.   Minimal residual disease (MRD) at the end of induction therapy, using rearrangement of immunoglobulin or T-receptor genes. Publications with > 50 patients, by order of publication. The treatments were too diverse to be reported. The two studies using immunophenotyping to assess MRD are described in the text.
Study Inclusion criteria n MRD cut- off points % MRD + patients (%) Missing data for MRDMedian follow-up (months) (Pt in CCR) EFS* in %: .(MRD + vs. MRD-) No. events/ patients Relative risk MRD+/MRD-
  • EORTC: European Organization for Research and Treatment of Cancer; BFM: Berlin Frankfurt Münster German group

  • WBC: white blood cell, NA: not available

  • *

    EFS: event-free Survival

  • †U Univariate analysis; My: Multivariate analysis taking into account y covariates.

  • This study assessed the interest of MRD on d 15, and the value of MRD at the end of induction, measured at two close time points.

Philadelphia, Pennsylvania (Wasserman et al, 1992)Pre B ALL516 10−33921645-y EFS: 32 vs. 6220/391·8 M5
Australian groupWBC < 100 ×100/l18110−64351≥ 48NA32/889 M2 or 4
(Brisco et al, 1994)no t(9–22)        
Amsterdam (Steenbergen et al, 1995)Pre B ALL5010−4393261NA16/3418·3 U
Paris Necker-TrousseauPre B ALL7610−2357717Events/Patients:2/17NA
(Landman-Parker et al, 1996)      2/5 vs. 0/12  
Brussells (Jacquy et al, 1997)Pre B ALL5110−32633203-y EFS: 27 vs. 1005/34NA
Houston, Texas (Roberts et al, 1997) Pre B ALL No t9–22, t4–11, t1–19745 10−6836845NA7/24MRD+ not predictive of relapse
Australian group (Brisco et al, 1997) Pre B ALL Rearrangement at diagnosis8110−357 NDNANANA
All cytogenetic abnormalities except t(9–22)        
Bristol/UK (Goulden et al, 1998)Rearrangement at diagnosis WBC < 50 ×100/l6610−4561069Relapse rate:33/59NA
Age 2–15 years No t(9,22), t(4,11)     69 vs. 19  
EORTC/Paris (Cave et al, 1998)2461·5 10−4413938 55 vs. 865 y EFS:32/1515·7 U
BFM/Rotterdam (Van Dongen et al, 1998)24010−4573048 61 vs. 973 y RFS:42/1691·9 M
Vienna group (Panzer-Grümayer et al, 2000)Rearrangement at diagnosis, and sample available at d 5 and at the end of induction6810−4547066 100 vs. 304 y RFS:16/68NA

Reproducibility of the assays

The method used to quantify MRD was different in the 11 studies, in terms of the type of sample (stored marrow smears, frozen or fresh samples), the choice of primers and probes, and the number of polymerase chain reaction (PCR) cycles, which varied between 25 (Cave et al, 1994) and 40 (Pongers-Willemse et al, 1999). Only one study (Van Dongen et al, 1998; Pongers-Willemse et al, 1999) mentions standardization among four laboratories in four countries with samples from six patients, but the difficulties encountered were not precisely stated, except for a ‘difference in sensitivity’ between the laboratories. All the other studies were carried out in a single laboratory, with no mention of controls in other laboratories or of reproducibility studies.

All the patients were never treated with the same protocol in a given study: instead, patients were classified in at least three prognostic groups, each receiving a different treatment. The differences in initial treatments among the studies concerned the dose of prednisone (40 or 60 mg/m2 for 21 or 28 d), the total dose of anthracycline (0–160 mg/m2 daunorubicin), the cumulative dose of asparaginase (60 000 U/m2 to 130 000 U/m2), the use of cyclophosphamide and cytarabine and the number and content of intrathecal injections. Owing to the wide diversity of the therapeutic schedules, this information is not reported in Table IV.

Population

Only two studies (Cave et al, 1998; Van Dongen et al, 1998) were based on data collected during prospective multicentre therapeutic trials, whereas the others were based on data collected in one or two institutions and included patients receiving various treatment modalities. Five studies involved patients with a T and pre-B immunophenotype (Brisco et al, 1994; Cave et al, 1998; Goulden et al, 1998; Van Dongen et al, 1998; Panzer-Grümayer et al, 2000), whereas the other six were restricted to patients with the pre-B immunophenotype (Wasserman et al, 1992; Steenbergen et al, 1995; Landman-Parker et al, 1996; Brisco et al, 1997; Jacquy et al, 1997; Roberts et al, 1997). With the exception of three studies (Cave et al, 1998; Van Dongen et al, 1998; Panzer-Grümayer et al, 2000), patients considered to be at a high risk (mainly on the basis of age or initial leucocytosis) were excluded (Table IV). All the studies except one (Landman-Parker et al, 1996) were restricted to patients in complete remission.

Statistical methods

The choice of cut-off point for residual disease varied among the studies from 10−2 to 10−6, and the percentage of positive patients ranged from 26% to 83% at the end of induction (Table IV). Residual disease was assessed between week 4 and 7. The choice of cut-off points resulted from exploratory analyses rather than pre-established working hypotheses.

Univariate prognostic analysis was always used. The number of events was never sufficient for reliable multivariate studies with four or more covariates.

Comments on the results

Eight studies showed an unfavourable influence of MRD in univariate analysis, whereas one study showed no effect (Roberts et al, 1997).

One study showed a link between the t(9;22) translocation and residual disease (Brisco et al, 1997). In the other studies, either cytogenetic features were not taken into account, or patients with these abnormalities were excluded from the analysis. One study considered simultaneously the persistence of peripheral blasts, the persistence of bone marrow blasts and residual disease by molecular rearrangement on d 15 of induction and at the end of induction, but did not compare the prognostic value of these different indicators (Panzer-Grümayer et al, 2000). The multivariate analyses in four studies showed an independent prognostic value of MRD at the end of induction therapy but, as mentioned above, the number of events was inadequate.

Discussion

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Conclusion
  7. Acknowledgments
  8. References

We analysed 43 studies examining the prognostic value of the early response to initial treatment for childhood ALL, assessed in terms of peripheral blast persistence, bone marrow blast persistence or MRD. None of these studies fulfilled the quality criteria stipulated by Simon and Altman (1994) for the acceptability of prognostic markers in clinical practice.

Reproducibility

The reproducibility of the different assays was not demonstrated in 42 of the 43 studies, and was only partially reported in one (Lilleyman et al, 1997).

Assessment of peripheral blast persistence depends on the reproducibility of the peripheral blast count. Total circulating white cell counts are now carried out automatically. Sample dilution, storage temperature and storage duration and the nature of the anticoagulant can lead to variations (Henderson & Wood, 1986; Nelson et al, 1989). Two methods can be used to evaluate the percentage of blasts among leucocytes: the visual method, which is subjective and has poor reproducibility (Rumbke, 1985), and automated methods, the results of which depend on identifying blasts cells among young forms (D'Onofrio et al, 1987; Kawarabayashi et al, 1987).

The evaluation of bone marrow blast persistence requires a bone marrow sample. Cytological diagnosis of leukaemia is perfectly standardized and qualitatively reproducible, but evaluation of the proportion of abnormal cells is less reliable. It depends on the amount of marrow sampled (Batinic et al, 1990), the quality of the smear, the heterogeneity of the disease from one bone marrow site to another (Bernard & Mathe, 1951; Hann et al, 1977; Jacobs, 1977) and the examiner. This is reminiscent of the practical problems encountered during marrow sampling for therapeutic purposes. The volume of marrow taken on each occasion and the puncture site influence the cellularity of the sample (Bacigalupo et al, 1992). Finally, it should be noted that centralized marrow smear review does not eliminate all sources of variability.

All studies of residual disease by means of immunophenotyping or molecular biology, with the exception of one (Brisco et al, 1997), were based on marrow samples. Only one study, based on an animal model (leukaemic rats), explored the reproducibility of a MRD assay based on molecular biology (Maartens et al, 1987), and showed major variability among and within animals. Technical considerations such as the type and number of probes and the number of PCR cycles may also contribute to the lack of reproducibility (Freeman et al, 1999).

The study of residual disease by means of the limiting-dilution method is associated with other difficulties, especially the estimation of titres and their variability (Loyer & Hamilton, 1984; Ouspenskaia et al, 1995).

Another limitation of MRD studies based on gene rearrangements is the lack of stability of the rearrangement itself, which may change between diagnosis and relapse in between 25% and 65% of cases (Beishuizen et al, 1994; Baruchel et al, 1995).

In most of the studies, organized during prospective therapeutic trials, the laboratory assays appear to have been organized blindly to the clinical data and outcome. In the retrospective studies, this information was lacking and could not be inferred from the papers (Brisco et al, 1994; Farahat et al, 1998; Goulden et al, 1998).

Study population

Many studies were restricted to subgroups of patients defined by the use of several markers (white blood cell count, cytogenetic abnormalities, immunophenotype, etc.), making it difficult to extrapolate their results. Most teams reported few missing data on peripheral blast persistence. As regards persistence of bone marrow blasts and the three techniques used to study residual disease, the percentage of missing data was almost always > 25% and sometimes > 50%, that is to say, far > 15% recommended by Simon & Altman (1994).

Choice of cut-off points and timing of the measurement

All the prognostic factors considered here were continuous. Multiple cut-off points were examined (Miller et al, 1983; Cave et al, 1998; Van Dongen et al, 1998), leading to the use of multiple statistical tests and resulting in over-optimistic P-values (Altman et al, 1994; Donadieu et al, 2000). No adjustments were made for the number of hypotheses tested.

The most frequent definition of persistence of peripheral blasts (i.e. > 1000 blasts on d 8 of treatment) was derived from a pilot study (Riehm et al, 1986). This choice was subsequently endorsed in other studies and can be considered as a standard. As regards persistence of bone marrow blasts, both the cut-off points and the assay dates varied. MRD assessments based on immunophenotyping and molecular methods used cut-off points ranging from 10−2 to 10−6 and assay dates ranging from 4 to 7 weeks after the start of treatment.

Number of patients/Number of events and statistical power

Using the Harrell rule (Harrell et al, 1985), two studies of persistence of peripheral blasts, five studies of bone marrow blasts and no studies of MRD had a sufficient number of patients for reliable multivariate analysis controlling for standard prognostic factors such as age, the white blood cell, cytogenetics and gender.

Definition of early response: taking treatment into account

The wide variety of initial treatments affects the very definition of persistence of peripheral blasts, persistence of bone marrow blasts and MRD. Few studies have examined this question. Only two studies carefully noted the effect of therapy on the definition of persistence of peripheral blasts (Thyss et al, 1997) and persistence of bone marrow blasts (Lilleyman et al, 1997). In both studies, apparently minimal differences in treatment markedly affected the results.

From prognostic factors to treatment selection

Invariably, the authors of prognostic studies recommend using their results as a basis for adapting subsequent therapy. But there is a difference between determining which factors affect outcome and which factors identify groups in which relative treatment effects are different. The determination of treatment effects in different groups is the next step after good prognostic studies. Jumping from the results of a prognostic study to the selection of therapy is a gamble. Only randomized trials including different risk groups can provide a reliable answer.

Conclusion

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Conclusion
  7. Acknowledgments
  8. References

Almost all studies have concluded that the early response to induction therapy for childhood ALL is strongly predictive of outcome. Unfortunately, given the variable definition of this early response, together with the lack of reproducibility studies and certain statistical weaknesses, the studies analysed here can only be considered as preliminary. Indeed, none met the criteria laid down by Simon and Altman (1994) and none of the new prognostic factors identified in these studies should therefore, in principle, be used in clinical practice. To improve future studies, it will be necessary (i) to standardize technical considerations through consensus conferences. This has been started by some teams (Pongers-Willemse et al, 1999; Van Dongen et al, 1999), and should be generalized. (ii) To study the reproducibility of measurements; and (iii) to standardize initial induction treatment. Such a proposal may appear unrealistic, but the use of a common treatment strategy would help to improve the reliability of studies on the early response to therapy. If stratification is envisaged for induction therapy, the National Cancer Institute (NCI) proposal to classify B-cell leukaemia, excluding forms with major cytogenetic abnormalities t(9;22) and t(4;11), on the basis of age and the white blood cell count appear to be appropriate. The three groups are defined by age below 1 year, age between 1 and 9 years and a white blood cell count < 50 × 109/l, and age > 9 or both age between 1 and 9 years and a white blood cell count > 50 × 109/l (Smith et al, 1996).

Unless these steps are taken, many studies run a risk of wasting resources and raising false hopes, without achieving the pragmatic aim of improving the management of each patient.

Acknowledgments

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Conclusion
  7. Acknowledgments
  8. References

We are indebted to Maurizio Arico, Judith Landman-Parker and Guy Leverger for their advice and criticism, and to David Young for editing the manuscript. His work was supported by a grant from the Association pour la Recherche sur les Maladies Hématologiques de l'enfant.

References

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Conclusion
  7. Acknowledgments
  8. References
  • Altman, D.G., Lausen, B., Sauerbrei, W., Schumacher, M. (1994) Danger of using ‘optimal’ cutpoints in the evaluation of prognostic factors. Journal of National Cancer Institute, 86, 829835.
  • Arico, M., Basso, G.F.M., Rizzari, C., Colella, R., Barisone, E., Zanesco, L., Rondelli, R., Pession, A., Masera, G. (1995) Good steroid response in vivo predicts a favorable outcome in children with T-cell acute lymphoblastic leukemia. Cancer, 75, 16841693.
  • Bacigalupo, A., Tong, J., Podesta, M., Piaggio, G., Figari, O., Colombo, P., Segno, G., Tedone, E., Moro, F., Van Lint, M.T. (1992) Bone marrow harvest for marrow transplantation: effect of multiple small (2 ml) or large (20 ml) aspirates. Bone Marrow Transplantation, 9, 467470.
  • Baruchel, A., Cayuela, J.M., MacIntyre, E., Berger, R., Sigaux, F. (1995) Assessment of clonal evolution at Ig/TCR loci in acute lymphoblastic leukaemia by single strand conformation polymorphism studies and highly resolutive PCR derived methods: implication for a general strategy of minimal residal disease detection. British Journal of Haematology, 90, 8593.
  • Batinic, D., Marusic, M., Pavletic, Z., Bodganic, V., Uzarevic, B., Nemet, D., Labar, B. (1990) Relationship between differing Volumes of bone marrow aspirates and their cellular composition. Bone Marrow Transplantation, 6, 103107.
  • Beishuizen, A., Verhoeven, M.A., Van Wering, E.R., Hahlen, K., Van Dongen, J.J. (1994) Analysis of Ig and T-cell receptor genes in 40 childhood acute lymphoblastic leukemias at diagnosis and subsequent relapse: implications for the detection of minimal residual disease by polymerase chain reaction analysis. Blood, 83, 1355.
  • Bernard, J. & Mathe, G. (1951) La discordance des moelles au cours des leucoses. Bulletin de la Societe Medicale Des Hopitaux, 67, 12851291.
  • Brisco, M.J., Condon, J., Hughes, E., Neoh, S.H., Sykes, P.J., Seshadri, R., Toogood, I., Waters, K., Tauro, G., Ekert, H., Morley, A.A. (1994) Outcome prediction in childhood acute lymphoblastic leukaemia by molecular quantification of residual disease at the end of induction. Lancet, 343, 196200.
  • Brisco, M.J., Sykes, P.J., Dolman, G., Neoh, S.H., Hughes, E., Peng, L.M., Tauro, G., Ekert, H., Toogood, I., Bradstock, K., Morley, A.A. (1997) Effect of the Philadelphia chromosome on minimal residual disease in acute lymphoblastic leukemia. Leukemia, 11, 14971500.
  • Campana, D. & Pui, C.H. (1995) Detection of minimal residual disease in acute lymphoblastic leukemia: methodologic advances and clinical significance. Blood, 85, 14161434.
  • Cave, H., Guidal, C., Rohrlich, P., Delfau, M.H., Broyart, A., Lescoeur, B., Rahimy, C., Fenneteau, O., Monplaisir, N., D'Auriol, L., Elion, J., Vilmer, E., Grandchamp, B. (1994) Prospective monitoring and quantitation of residual blasts in childhood acute lymphoblastic leukemia by polymerase chain reaction study of delta and gamma T-cell receptor genes. Blood, 83, 18921902.
  • Cave, H., Van Der Werff ten Bosch, J., Suciu, S., Guidal, C., Waterkeyn, C., Otten, J., Bakkus, M., Thielemans, K., Grandchamp, B., Vilmer, E. (1998) Clinical significance of minimal residual disease in childhood acute lymphoblastic leukemia. European Organization for Research and Treatment of Cancer – Childhood Leukemia Cooperative Group. New England Journal of Medicine, 339, 591598.
  • Chessells, J.M., Bailey, C., Richards, S.M. (1995) Intensification of treatment and survival in all children with lymphoblastic leukaemia: results of UK medical research council UKALL X. Lancet, 345, 143148.
  • Conter, V., Arico, M., Valsecchi, M.G., Basso, G., Biondi, A., Madon, E., Mandelli, F., Paolucci, G., Pession, A., Rizzari, C., Rondelli, R., Zanesco, L., Masera, G. (2000) Long-term results of the italian association of pediatric hematology and oncology (AEIOP) acute lymphoblastic leukemia studies. 1982–95. Leukemia, 14, 21962204.
  • Coustan-Smith, E., Sancho, J., Hancock, M.L., Boyett, J.M., Behm, F.G., Raimondi, S.C., Sandlund, J.T., Rivera, G.K., Rubnitz, J.E., Ribeiro, R.C., Pui, C.H., Campana, D. (2000) Clinical importance of minimal residual disease in childhood acute lymphoblastic leukemia. Blood, 96, 26912696.
  • Donadieu, J., Auclerc, M.F., Baruchel, A., Leblanc, T., Landman-Parker, J., Perel, Y., Michel, G., Cornu, G., Bordigoni, P., Sommelet, D., Leverger, G., Hill, C., Schaison, G. (1998) Critical study of prognostic factors in childhood acute lymphoblastic leukaemia: differences in outcome are poorly explaine by the most significant prognostic variables. British Journal of Haematology, 102, 729739.
  • Donadieu, J., Auclerc, M.F., Baruchel, A., Perel, Y., Bordigoni, P., Leblanc, T., Cornu, G., Sommelet, D., Leverger, G., Schaison, G., Hill, C. (2000) Prognostic study of continuous variables (white blood cell count, peripheral blast cell count, haemoglobin level, platelet count and age) in childhood acute lymphoblastic leukaemia. Analysis of a population of 1545 children treated by the French Acute Lymphoblastic Leukemia Group. British Journal of Cancer, 83, 16171622.
  • D'Onofrio, G., Mancini, S., Leone, G., Bizzi, B., Mango, G. (1987) Identification of blast cells in peripheral blood through automatic assessment of nuclear density: a new tool for monitoring patients with acute leukaemia. British Journal of Haematology, 66, 473.
  • Eden, O.B., Harrison, G., Richards, S., Lilleyman, J.S., Bailey, C.C., Chessells, J.M., Hann, I.M., Hill, F.G.H., Gibson, B.E.S. (2000) Long-term follow-up of the United Kingdom Medical Research Council protocols for childhood acute lymphoblastic leukemia. 1980–97. Leukemia, 14, 23072320.
  • Farahat, N., Morilla, A., Owusu-Ankomah, K., Morilla, R., Ross Pinkerton, C.R., Treleaven, J.G., Matutes, E., Powles, R.L., Catovsky, D. (1998) Detection of minimal residual disease in B-lineage acute lymphoblastic leukaemia by quantitative flow cytometry. British Journal of Haematology, 101, 158164.
  • Freeman, W.M., Walker, S.J., Vrana, K.E. (1999) Quantitative RT-PCR: pitfalls and potential. Biotechniques, 26, 112125.
  • Gajjar, A., Ribeiro, R., Hancock, M.L., Rivera, G.K., Mahmoud, H., Sandlund, J.T., Crist, W.M., Pui, C.H. (1995) Persistence of circulating blasts after 1 week of multiagent chemotherapy confers a poor prognosis in childhood acute lymphoblastic leukemia. Blood, 86, 12921295.
  • Gaynon, P.S., Bleyer, W.A., Steinherz, P.G., Finklestein, J.Z., Littman, P.S., Miller, D.R., Reaman, G., Sather, H.N., Hammond, G.D. (1988) Modified BFM therapy for children with previously untreated acute lymphoblastic leukemia and unfavorable prognostic features. American Journal of Pediatric Hematology and Oncology, 10, 4250.
  • Gaynon, P.S., Bleyer, W.A., Steinherz, P.G., Finklestein, J.Z., Littman, P., Miller, D.R., Reaman, G., Sather, H., Hammond, G.D. (1990) Day 7 marrow response and outcome for children with acute lymphoblastic leukemia and unfavorable presenting features. Medical Pediatric Oncology, 18, 273279.
  • Goulden, N.J., Knechtli, C.J., Garland, R.J., Langlands, K., Hancock, J.P., Potter, M.N., Steward, C.G., Oakhill, A. (1998) Minimal residual disease analysis for the prediction of relapse in children with standard-risk acute lymphoblastic leukaemia. British Journal of Haematology, 100, 235244.DOI: 10.1046/j.1365-2141.1998.00574.x
  • Hann, I.M., Morris Jones, P.H., Evans, D.I. (1977) Discrepancy of bone marrow aspirations in acute lymphoblastic leukaemia in relapse. Lancet, i, 12151216.
  • Harms, D.O. & Janka-Schaub, G.E. (2000) Cooperative study group for childhood acute lymphoblastic leukemia (COALL): long-term follow-up of trials 82, 85, 89 and 92. Leukemia, 14, 22342239.
  • Harrell, F.E., Lee, K.L., Matchar, D.B., Reichert, T.A. (1985) Regression models for prognostic prediction: advantages, problems, and suggested solutions. Cancer Treatment Reports, 69, 10711077.
  • Henderson, S.J. & Wood, S.K. (1986) Factors producing error in the total leucocyte coutn as measured on the Coulter Counter S plus IVD. Clinical Laboratory Haematology, 8, 341346.
  • Jacobs, P. (1977) Discrepant bone marrow aspirations in leukaemia. Lancet, ii, 355356.
  • Jacquillat, C., Weil, M., Gemon, M.F., Auclerc, G., Loisel, J.P., Delobel, J., Flandrin, G., Schaison, G., Izrael, V., Bussel, A., Dresch, C., Weisberger, C., Rain, D., Tanzer, J., Najean, Y., Seligmann, M., Boiron, M., Bernard, J. (1973) Combination therapy in 130 patients with acute lymphoblastic leukemia (Protocol 06 LA 66-Paris). Cancer Research, 33, 32783284.
  • Jacquy, C., Delepaut, B., Van Daele, S. (1997) A prospective study of minimal residual disease in childhood B-lineage leukaemia: MRD level at the end of induction is a strong predictive factor of relapse. British Journal of Haematology, 98, 140146.
  • Kamps, W.A., Veerman, A.J.P., Van Wering, E.R., Van Weerden, J.F., Slater, R., Van der Does-Van den berg, A. (2000) Long-term follow-up of Dutch Childhood Leukemia Study Group (DCLSG) protocols for children with acute lymphoblastic leukemia. 1984–91. Leukemia, 14, 22402246.
  • Kawarabayashi, K., Tsuda, I., Tatsumi, N. (1987) Leukemic blasts detected by the technicon H*1 hematology analyser. American Journal of Clinical Pathology, 88, 624627.
  • Landman-Parker, J., Aubin, J., Delabesse, E., Tabone, M.D., Adam, M., Millien, C., Leboeff, D., Buzyn-Veil, A., Dollfus, C., Leverger, G., MacIntyre, E.A. (1996) Simplified strategies for minimal residual disease detection in B-cell precurssor acute lymphoblastic leukaemia. British Journal of Haematology, 95, 281290.
  • Lilleyman, J.S., Gibon, B.E.S., Stevens, R.F., Will, A.M., Hann, I.M., Richards, S.M., Hill, F.G.H. (1997) Clearance of bone marrow infiltration after 1 week of therapy for childhood lymphoblastic leukaemia: Clinical importance and the effect of daunorubicin. British Journal of Haematology, 97, 603606.
  • Loyer, M.W. & Hamilton, M.A. (1984) Interval estimation of the density of organismes using a serial-dilution experiment. Biometrics, 40, 907916.
  • Maartens, A.C.M., Schuttz, F.W., Hagenbeck, A. (1987) Nonhomogeneous distribution of leukemia in the bone marrow during minimal residual disease. Blood, 70, 10731078.
  • Miller, D.R., Leikin, S., Albo, V., Sather, H., Karon, M., Hammond, D. (1983) Prognostic factors and therapy in acute lymphoblastic leukemia of childhood: CCG-141. Cancer, 51, 10411049.
  • Miller, D.R., Coccia, P.F., Bleyer, W.A., Lukens, J.N., Siegel, S.E., Sather, H.N., Hammond, G.D. (1989) Early response to induction therapy as a predictor of disease-free survival and late recurrence of childhood acute lymphoblastic leukemia: a report from the Childrens Cancer Study Group. Journal of Clinical Oncology, 7, 18071815.
  • Nelson, L., Charache, S., Wingfield, S., Keyser, E. (1989) Laboratory evaluation of differential white blood cell coutn information from Coulter S-Plus IV and tehnicon H-1 in patient populations requiring rapid ‘turnaround’ time. American Journal of Clinical Pathology, 91, 563569.
  • Ouspenskaia, M.V., Johnston, D.A., Roberts, W.M., Estrov, Z., Zipf, T.F. (1995) Acurate quantification of residual B-Precursor acute lymphoblastic leukemia by limiting dilution and a Pcr-based detection system: a description of the method and the principles involved. Leukemia, 9, 321328.
  • Panzer-Grümayer, E.R., Schneider, M., Panzer, S., Fasching, K., Gadner, H. (2000) Rapid molecular response during early induction chemotherapy predicts a good outcome in childhood acute lymphoblastic leukemia. Blood, 95, 790794.
  • Pongers-Willemse, M.J., Seriu, T., Stolz, F., D'Aniello, E., Gameiro, P., Pisa, P., Gonzalez, M., Bartram, C.R., Panzer-Grumayer, E.R., Biondi, A., San Miguel, J.F., Van Dongen, J.J. (1999) Primers and protocols for standardized detection of minimal residual disease in acute lymphoblastic leukemia using immunoglobulin and T cell receptor gene rearrangements and TAL1 deletions as PCR targets: report of the BIOMED-1 CONCERTED ACTION: investigation of minimal residual disease in acute leukemia. Leukemia, 13, 110118.
  • Pui, C.H., Boyett, J.M., Rivera, G.K., Hancock, M.L., Sandlund, J.T., Ribeiro, R.C., Rubnitz, J.E., Behm, F.G., Raimondi, S.C., Gajjar, A., Razzouk, B., Campana, D., Kun, L.E., Relling, M.V., Evans, W.E. (2000) Long-term results of total therapy studies 11, 12 and 13A for childhood acute lymphoblastic leukemia at St Jude Children's research hospital. Leukemia, 14, 22862294.
  • Rautonen, J., Hovi, L., Siimes, M.A. (1988) Slow disappearance of peripheral blast cells: an indenpendent risk factor indicating poor prognosis in children with acute lymphoblastic leukemia. Blood, 71, 989991.
  • Reiter, A., Schrappe, M., Ludwig, W.D., Hiddemann, W., Sauter, S., Henze, G., Zimmermann, M., Lampert, F., Havers, W., Niethammer, D., Odenwald, E., Ritter, J., Mann, G., Welte, K., Gadner, H., Riehm, H. (1994) Chemotherapy of 998 unselected childhood acute lymphoblastic leukemia patients. Results and conclusions of the multicenter trail ALL-BFM 86. Blood, 84, 31223133.
  • Riehm, H., Reiter, A., Schrappe, M., Berthold, F., Dopfer, R., Gerein, V., Ludwig, R., Ritter, J., Stollman, B., Henze, G. (1986) The in vivo response on corticosteroid therapy as an additional prognostic factor in childhood acute lymphoblastic leukemia (therapy study ALL-BFM-83). Klinishe Pädiatrie, 199, 151160.
  • Rivera, G.K., Raimondi, S.C., Hancock, M.L., Behm, F.G., Pui, C.H., Abromowitch, M., Mirro, J., Ochs, J.S., Look, A.T., Williams, D.K., Murphy, S., Dahl, G.V., Kalwinsky, D.K., Evans, W.E., Kun, L.E., Simone, J.V., Crist, W.M. (1991) Improved outcome in childhood acute lymphoblastic leukaemia with reinforced early treatment and rotational combination chemotherapy. Lancet, 337, 6166.
  • Roberts, W.M., Estrov, Z., Ouspenskaia, M.V., Johnston, D.A., McLain, K., Zipf, T.F. (1997) Measurement of residual leukemia during remission in childhood leukemia. New England Journal of Medecine, 336, 317323.
  • Rumbke, C.L. (1985) The imprecision of the ratio of two percentages observed in differential blood cell counts: a warning. Blood Cells, 11, 137.
  • Schrappe, M., Reiter, A., Riehm, H. (1996) Cytoreduction and prognosis in childhood acute lymphoblastic leukemia. Journal of Clinical Oncology, 14, 24032405.
  • Schrappe, M., Arico, M., Harboot, J., Biondi, A., Zimmermann, M., Conter, V., Reiter, A., Valsecchi, M.G., Gadner, H., Basso, G., Bartram, C.R., Lampert, F., Riehm, H., Masera, G. (1998) Philadelphia chromosome-positive (Ph+) childhood acute lymphoblastic leukemia: good initial steroid response allows early prediction of a favorable treatment outcome. Blood, 92, 27302741.
  • Schrappe, M., Reiter, A., Zimmerman, M., Harbott, J., Ludwig, W.-D., Henze, G., Gadner, H., Odenwald, E., Riehm, H. (2000) Long-term results of four consecutive trials in childhood ALL performed by the ALL-BFM study group from 1981to 1993. Leukemia, 14, 22052222.
  • Simon, R. & Altman, D.G. (1994) Statistical aspects of prognostic factor studies in oncology. British Journal of Cancer, 69, 979985.
  • Smith, M., Arthur, D., Camitta, B., Caroll, A.J., Crist, W., Gaynon, P., Gelber, R., Heerema, N., Korn, E.L., Link, M., Murphy, S., Pui, C.H., Pullen, J., Reaman, G., Sallan, S.E., Sather, H., Shuster, J., Simon, R., Trigg, M., Tubergen, D., Uckun, F., Ungerleider, R. (1996) Uniform approach to risk-classification and treatment assignment for children with ALL. Journal of Clinical Oncology, 14, 1824.
  • Steenbergen, E.J., Verhagen, O.J., Van Leeuwen, E.F., Van Den Berg, H., Behrendt, H., Von Slater, R.M., Dem Borne, A.E., Van Der Schoot, C.E. (1995) Prolonged persistence of PCR-detectable minimal residual disease after diagnosis or first relapse predicts poor outcome in childhood B-precursor acute lymphoblastic leukemia. Leukemia, 9, 17261734.
  • Steinherz, P.G., Gaynon, P.S., Breneman, J.C., Cherlow, J.M., Grossman, N.J., Kersey, J.H., Johnstone, H.S., Sather, H.N., Trigg, M.E., Chappell, R., Hammond, D., Bleyer, W.A. (1996) Cytoreduction and prognosis in acute lymphoblastic leukemia – The importance of early marrow response: report from the childrens cancer group. Journal of Clinical Oncology, 14, 389398.
  • Thyss, A., Suciu, S., Bertrand, Y., Mazingue, F., Robert, A., Vilmer, E., Mechinaud, F., Benoit, Y., Brock, P., Ferster, A., Lutz, P., Boutard, P., Marguerite, G., Plouvier, E., Michel, G., Plantaz, D., Munzer, M., Rialland, X., Chantraine, J.M., Norton, L., Solbu, G., Philippe, N., Otten, J. (1997) Systemic effect of intrathecal methotrexate during the initial phase of treatment of childhood acute lymphoblastic leukemia. Journal of Clinical Oncology, 15, 18241830.
  • Tivey, H. (1952) Prognosis for survival in the leukemias of childhood. Pediatrics, 10, 4859.
  • Toyoda, Y., Manabe, A., Tsuchida, M., Hanada, R., Ikuta, K., Okimoto, Y., Ohara, A., Mori, T., Ishimoto, K., Sato, T., Kaneko, T., Maeda, M., Koike, K., Shitara, K., Hoshi, Y., Hosoya, R., Tsunematsu, Y., Bessho, F., Nakazawa, S., Saito, T. (2000) Six months of maintenance chemotherapy after intensified treatment for acute lymphoblastic leukemia in childhood. Journal of Clinical Oncology, 18, 15081515.
  • Tsuchida, M., Ikuta, K., Hanada, R., Toyoda, Y., Saito, T., Isoyama, K., Sugita, K., Toyoda, Y., Manabe, A., Koike, K., Kinoshita, A., Maeda, M., Ishimoto, K., Sato, T., Okimoto, Y., Kaneko, T., Kajiwara, M., Sotomatsu, M., Hayashi, Y., Yabe, H., Hosoya, R., Hoshi, Y., Ohira, M., Bessho, F., Tsunematsu, Y., Tsukimoto, I., Nakazawa, S. (2000) Long-term follow-up of childhood acute lymphoblastic leukemia in Tokyo Children's Cancer Study Group 1981–95. Leukemia, 14, 22952306.
  • Tubergen, D.G., Gilchrist, G.S., O'Brien, R.T., Coccia, P.F., Sather, H.N., Waskerwitz, M.J. (1993) Improved outcome with delayed intensification for children with acute lymphoblastic leukemia and intermediate presenting features: a children's cancer group phase III. Journal of Clinical Oncology, 11, 527537.
  • Van Dongen, J.J.M., MacIntyre, E.A., Gabert, J.A., Delabesse, E., Rossi, V., Saglio, G., Gottardi, E., Rambaldi, A., Dotti, G., Griesinger, F., Parreira, A., Gameiro, P., Gonzalez Diaz, M., Malec, M., Langerak, A.W., San Miguel, J.F., Biondi, A. (1999) Standardized RT-PCR analysis of fusion gene transcripts from chromosome aberration in acute leukemia for detection of minimal residual disease. Leukemia, 13, 19011928.
  • Van Dongen, J.J.M., Seriu, T., Panzer-Grumayer, E.R., Biondi, A., Pongers-Willemse, M.J., Corral, L., Stolz, F., Schrappe, M., Masera, G., Kamps, W.A., Gadner, H., Van Wering, E.R., Ludwig, W.D., Basso, G., De Bruijn, M.A.C., Cazzaniga, G., Hettinger, K., Van der Does Van Der Berg, A., Hop, W.C.J., Riehm, H., Bartram, C.R. (1998) Prognostic value of minimal residual disease in acute lymphoblastic leukaemia in childhood. Lancet, 352, 17311738.DOI: 10.1016/S0140-6736(98)04058-6
  • Van Wering, E.R., Beishuizen, A., Roeffen, E.T., Van der Linden-Schrever, B.E., Verhoeven, M.A., Hahlen, K., Hooijkaas, H., Van Dongen, J.J. (1995) Immunophenotypic changes between diagnosis and relapse in childhood acute lymphoblastic leukemia. Leukemia, 9, 15231533.
  • Vilmer, E., Suciu, S., Ferster, A., Bertrand, Y., Cavé, H., Thyss, A., Benoit, Y., Dastugue, N., Fournier, M., Souillet, G., Manel, A.M., Robert, A., Nelken, B., Millot, F., Lutz, P., Rialland, X., Mechinaud, F., Boutard, P., Behar, C., Chantraine, J.M., Plouvier, E., Laureys, G., Brock, P., Uyttebroeck, A., Margueritte, G., Plantaz, D., Norton, L., Francotte, N., Gyselinck, J., Waterkeyn, C., Solbu, G., Philippe, N., Otten, J. (2000) Long term results of three randomized trials (58831, 58832, 58881) in childhood acute lymphoblastic leukemia: a CLCG-EORTC report. Leukemia, 14, 22572266.
  • WassermanR., N., Y., J.H., B.A., S., R.B., B., G. (1992) Residual disease at the end of the induction therapy as a predictor of relapse during therapy in childhood B-lineage acute lymphoblastic leukemia. Journal of Clinical Oncology, 10, 18791885.