Randomized assessment of delayed intensification and two methods for parenteral methotrexate delivery in childhood B-ALL: Children’s Oncology Group Studies P9904 and P9905

Abstract

The delayed intensification (DI) enhanced outcome for patients with acute lymphoblastic leukemia (ALL) treated on BFM 76/79 and CCG 105 after a prednisone-based induction. Childrens Oncology Group protocols P9904/9905 evaluated DI via a post-induction randomization for eligible National Cancer Institute (NCI) standard (SR) and high-risk (HR) patients. A second randomization compared intravenous methotrexate (IV MTX) as a 24- (1 g/m²) vs. 4-h (2 g/m²) infusion. NCI SR patients received a dexamethasone-based three-drug and NCI HR/CNS 3 SR patients a prednisone-based four-drug induction. End induction MRD (minimal residual disease) was obtained but did not impact treatment. DI improved the 10-year continuous complete remission (CCR) rate; 75.5 ± 2.5% vs. 81.8 ± 2.2% p = 0.002, whereas MTX administration did not; 4-h 80.8 ± 1.9%; 24-h 81.4 ± 1.9% (p = 0.7780). Overall survival (OS) at 10 years did not differ with DI: 91.4 ± 1.6% vs. 90.9 ± 1.7% (p = 0.25) without but was higher with the 24-h MTX infusion; 4-h 91.1 ± 1.4%; 24-h 93.9 ± 1.2% (p = 0.0209). MRD predicted outcome; 10-year CCR 87.7 ± 2.2 and 82.1 ± 2.5% when MRD was <0.01% with/without DI (p = 0.007) and 54.3 ± 8% and 44 ± 8% for patients with MRD ≥ 0.01% with/without DI (p = 0.11). DI improved CCR for patients with B-ALL with and without end induction MRD.

Introduction

Substantial progress has been made in the treatment of childhood acute lymphoblastic leukemia (ALL) [1–3] since remissions were described in 1948 [4]. Critical components of current combination chemotherapeutic regimens include treatment intensification following remission induction. Reviews of randomized clinical trials have described the efficacy of several post-induction intensification therapies [1, 5–8] with the most common being the BFM (Berlin-Frankfurt-Muenster) protocol II that includes a re-induction/re-consolidation phase also referred to as delayed intensification (DI) [9]. The Pediatric Oncology Group (POG) initially obtained similar results using multiple courses of intravenous methotrexate (MTX) at 1 g/m²/24-h, with leucovorin rescue, for post-induction intensification [10, 11]. This report describes the outcome of a randomized comparison of two different approaches to parenteral MTX delivery and the impact of a DI in the context of the MTX infusions and a dexamethasone-based induction for National Cancer Institute (NCI) standard risk (SR) patients.

This trial analyzed, but did not act upon, the prognostic value of minimal residual disease (MRD), in peripheral blood on day 8 and bone marrow on day 29, determined by flow cytometry performed at a single central reference laboratory [12].

Material and methods

Patients

Patients with B-ALL, aged 1 to 21 years, were enrolled on the classification/induction protocol, COG P9900. Patients with standard risk (SR) features according to the NCI/Rome criteria (age 1–9.99 years and white blood cell count (WBC) < 50,000/microliter) received a three-drug, dexamethasone-based induction. Patients with NCI high-risk (HR) features (age ≥ 10 years and/or WBC ≥ 50,000/microliter) initially received a four-drug, dexamethasone-based induction but prednisone replaced dexamethasone, after the first 34 patients were treated, because of excessive toxicity (Table 1). Patients with central nervous system (CNS) 3 status (≥5 WBC/microliter on cytospin of cerebrospinal fluid with blasts present) or testicular leukemia received the four-drug induction. Patients with an M2 marrow (5–25% blasts) on day 29 of induction received two additional weeks of the same therapy. Written informed consent was obtained according to Institutional Review Board and FDA guidelines. At the end of induction, patients who achieved an M1 marrow (<5% blasts) by day 29 or 43, were eligible to participate in post-induction therapeutic studies. NCI SR patients with favorable genetics (trisomies of 4/10 or t(ETV6/RUNX1) were classified as lower risk and enrolled on 9904. NCI SR patients without and NCI HR with favorable genetics or pts who did not meet refined NCI high-risk age and WBC criteria [13] were eligible for 9905 (Table 2). Patients with CNS 3 disease, t(9;22), t(4;11), or hypodiploidy were excluded from both studies. This paper describes P9904 and P9905; the results of P9906 have been published [14]. The studies are registered at http://clinicaltrials.org as NTC00005585 and NCT00005596.

Table 1.

Treatment and randomization

Three-drug induction		Days
Dexamethasone	6 mg/m2/day po	1–28
Vincristine	1.5 mg/m2 IV	1,8,15,22
l-asparaginasea	10,000 IU/m2 IM	2,5,8,12,15,19
PEG asparaginaseb	2500 IU/m2	Day 4, 5, or 6
Cytarabineb	IT by age	Day 1
MTX	IT by age	1a, 8, 15c, 22c
Four-drug induction		Days
Dexamethasonea	6 mg/m2/day po	1–28
Prednisoneb	40 mg/m2/day po	1–28
Vincristine	1.5 mg/m2 IV	1,8,15,22
Daunomycin	30 mg/m2 IV	8,15,22
l-asparaginase	10,000 IU/m2 IM	2,5,8,12,15,19
MTX	IT by age	1,8,15c, 22c
Extended induction-only for patients with M2 marrows on day 29
Prednisone	40 mg/m2/day po	29–42
Vincristine	1.5 mg/m2 IV	29, 36
Daunomycin	30 mg/m2 IV	29, 36
PEG asparaginase	2500 IU/m2 IM	29, 30, or 31
Randomized consolidation or delayed intensification
Regimens A and B		Weeks
Methotrexate	1g/m2/24-h (Reg A)d 2 g/m2/4-h (Reg B)d	7, 10, 13, 16, 19, 22
Vincristine	1.5 mg/m2 IV	8, 9, 17, 18
Dexamethasone	6 mg/m2 × 7 days	8, 17
6-Mercaptopurine	50 mg/m2 nightly	5–24
Regimens C and D		Weeks
Methotrexate	1 g/m2/24-h (Reg C)d 2 g/m2/4-h (Reg D)d	7, 10, 13, 24, 27, 30
Vincristine	1.5 mg/m2 IV	8, 9, 16, 17, 18, 28, 29
Dexamethasone	6mg/m2 × 7 days	8, 16-18, 28
Daunomycin	30 mg/m2 IV	16, 17, 18
PEG asparaginase	2500IU/m2 IM	16
	1 g/m2 IV	20
Cyclophosphamide
Cytarabine	75 mg/m2 IV or SQ × 4 days/week	20, 21
6-Thioguanine	60 mg/m2 nightly	20–21
Intensive continuation (4, 12-week cycles; P9905 only)
Methotrexate	25 mg/m2/dose q 6 h × 4 doses	1,3,5,7,9,11
6-Mercaptopurine	75 mg/m2 nightly	Continuously
Methotrexate	Dose by age IT	12
Vincristine	1.5 mg/m2 IV	12
Dexamethasone	6 mg/m2 × 7 days	12
Continuation (to 2.5 years continuous complete remission)
Methotrexate	20 mg/m2/dose po weekly	Reg A/B: weeks 73–130 Reg C/D: weeks 81–130
6-Mercaptopurine	75 mg/m2 nightly	Continuously
Methotrexate	Dose by age IT	Every 12 weeks
Vincristine	1.5 mg/m2 IV	Day 1 and 8 of every 12th week;
Dexamethasone	6 mg/m2 × 7 days	Every 12 weeks

MTX methotrexate

Italicized entries indicate the drugs delivered during the randomized components of the trial

^aPrior to amendment

^bAfter amendment

^cFor patients with WBC < 5/μl and blasts in the CSF

^dRandomized assignment: all patients participated in the MTX randomization; 9904 patients with trisomies 4/10 or TCF3-PBX1 were excluded from the DI randomization

Table 2.

9900 risk group classification for B-ALL

Risk group	Characteristics
Low	NCI SR, in complete remission (CR), favorable genetics (trisomies 4 & 10 or ETV 6-RUNX1)
Standard risk	NCI SR, in CR, NO favorable or unfavorable genetics; OR NCI SR with CNS 3/testicular disease OR NCI HR, in CR, AND Favorable genetics or Age/WBC not in Shuster Algorithma NCI SR or NCI HR with TCF3-PBX1 translocation
High risk	NCI SR, in CR, with unfavorable genetics (KMT2A/MLL) NCI SR with CNS 3/testicular disease OR NCI HR, in CR, and Age/WBC in Shuster algorithm
Very high risk	NCI SR or NCI HR with: BCR-ABL1 OR induction failure OR DNA Index < 0.81
aShuster’s algorithm
Boys (age in years)			Girls (age in years)
8			12
9			13
10			14
11			15
≥12			≥16

Central reference laboratories

Blast immunophenotyping was performed locally and confirmed at the Johns Hopkins Hospital central reference laboratory (MJB). DNA index; fluorescence in situ hybridization (FISH) for trisomies 4 and 10; and PCR testing for TCF3-PBX1, BCR-ABL1, KMT2A (MLL)-AF4, and ETV6-RUNX1 fusion transcripts were performed at the University of New Mexico central reference laboratory (CLW) [12]. Additional KMT2A gene rearrangements were detected by Fluorescence in situ hybridization (FISH) using a break apart probe strategy. Chromosome analyses were performed at certified local cytogenetics laboratories, with karyotype results centrally reviewed. All reference laboratory information was available for day 29 risk group assignments. Peripheral blood samples on day 8 and bone marrow samples on day 29 of induction were sent to the Johns Hopkins reference laboratory for MRD determination. The results of this testing were not made available to treating institutions and did not affect therapy [12].

Treatment assignment and randomization

Patients on P9904/P9905 were risk classified at the end of induction on P9900 (Table 2). Treatment assignment and randomizations are given in the CONSORT diagram (Fig. 1). COG P9904/P9905 trial had a pseudo-factorial design where all patients were randomized to intravenous MTX as a 24-h infusion of 1 g/m² vs. a 4-h infusion of 2 g/m². The 1 g/m² infusion had been incorporated in previous POG trials [15] and the 2 g/m² infusion had been used at St. Jude Children’s Research Hospital [16]. The shorter infusion was created to facilitate outpatient drug delivery. The dose and schedule of leucovorin rescue was the same in both regimens. A subset of the patients then had a second randomization to determine whether a DI phase would improve the outcome of patients who received intensification with six courses of intravenous MTX. NCI SR patients with trisomies of chromosomes 4 and 10 were only randomized to the four or 24-h MTX infusions based on their excellent outcome on earlier POG trials [17, 18]. Patients enrolled on P9905 participated in both randomizations except for patients with TCF3-PBX1 or high-risk patients as defined by the Shuster Pragmatic High-Risk Group who had a favorable genetic lesion (trisomies 4 and 10 or ETV6-RUNX1). The poor outcome of these two groups on prior POG trials [15, 19] led to non-random assignment to receive DI.

Fig. 1.

CONSORT diagram

Treatment

The initial three- and four-drug inductions were modified because of unacceptable toxicity (Table 1). Prednisone (40 mg/m²/day × 28 days) replaced the dexamethasone (6 mg/m²/day × 28 days) in the four-drug induction (6/19/2000). The initial dexamethasone-based three-drug induction included E coli asparaginase (6000 IU/m² × 6) and intrathecal (IT) MTX on day 1. Post amendment, IT cytarabine was given on day 1 and a single dose of pegaspargase was given on either day 4, 5, or 6 (11/27/2002).

All patients on P9904 and P9905 who achieved an M1 marrow at day 29 or 43 of induction received six courses of IV MTX, randomized as a 24-h-infusion of 1 g/m² or 4-h infusion of 2 g/m². Leucovorin rescue began at hour 42 (10 mg/m² q6h × 3); the dose at hour 54 was not given in the absence of toxicity if the 48 h MTX level was <0.2 μM. IT MTX was given every 3 weeks with the IV MTX. The six courses of IV/IT MTX were interrupted at week 16 for patients randomized or assigned to receive the DI (Table 1).

Following completion of post-induction intensification, patients enrolled on P9905 received a continuation phase, which included divided-dose oral MTX with leucovorin rescue every other week for 24 courses, and nightly oral 6-MP. Intrathecal MTX was given during week 12 of each cycle with a 7-day pulse of dexamethasone and vincristine given on days 1 and 8 of the pulse. Patients on P9904 began this continuation phase immediately after consolidation while the P9905 patients entered this phase after the continuation described above. Total duration of therapy was 130 weeks for all patients.

Definitions of response/toxicity

Complete remission (CR) was defined as fewer than 5% marrow blasts with no evidence of extramedullary disease. Duration of complete remission was the time between demonstration of complete remission and relapse at any site. Bone marrow relapse required >25% lymphoblasts and CNS relapse required ≥5 WBC/microliter cerebrospinal fluid with identifiable blasts. Toxicity reporting followed CTCAE version 2 guidelines.

Statistical considerations

Patients risk classified as low or high risk after completing induction therapy on P9900, were eligible to enroll on P9900 or P9905. Comparisons presented here are for the randomized cohorts described above. The method of per-muted blocks was used for the randomizations. The primary endpoint was continuous complete remission (CCR), defined as time from enrollment on P9904/P9905 to first event (relapse, second malignancy, or remission death) or date of last contact (non-events). The MTX randomization was powered (80%) to compare 4-year CCR (82.5% vs. 87.5%), log-rank test, 5% two-sided alpha, with a total sample size of 1604, with 4-years minimum follow-up. The DI randomization was powered (80%) to compare 4-year CCR (80% vs. 85%), log-rank test, 5% one-sided alpha, with a total sample size of 1410, with 4-years minimum follow-up. Overall survival (OS) was defined as time from enrollment to death or date of last contact for patients who were alive. Survival rates were estimated using the Kaplan-Meier method and standard errors of Peto et al. [20, 21] Results are expressed as estimate ± standard error. Survival curves were compared using the log-rank test. Cumulative incidence rates for isolated CNS relapse and marrow relapse were computed using the cumulative incidence function for competing risks, and comparisons were conducted using the K-sample test. [22]. Multivariate analysis of risk factors for outcome was performed using the Cox regression model [23]. The Chi-square or Fisher’s Exact test were used for comparison of proportions between groups. A p-value < 0.05 was considered as significant for all comparisons. The COG statistical office was responsible to COG member institutions for patient enrollment, confirmation of eligibility, randomized assignment, and oversight of data collection and analysis.

All analyses were performed using SAS software (version 9.4; SAS Institute, Cary, NC). Graphics were generated using R version 2.13.1 (http://www.r-pproject.org).

Results

Patient characteristics

Two-thousand six-hundred and fifty-five eligible patients with B-ALL enrolled on the classification study P9900 between 12/13/1999 and 2/28/2005, received a three- or four-drug induction based on NCI risk group (standard vs. high), and were successfully risk stratified as eligible for post-induction therapy on P9904, P9905, or P9906 (Fig. 1). The CR rate on P9900 was 98.8% for NCI SR patients without CNS 3 or overt testicular disease (1768/1790) and 96.2% for NCI HR patients and NCI SR with CNS 3 or overt testicular disease (1034/1075). Of these, 1076/1352 (79.9%) and 838/988 (84.8%) of the patients classified as low and high risk, elected to enroll and participate in the post-induction randomizations on P9904 and P9905, respectively. Eight patients enrolled on P9904 and 21 on P9905 were found to be ineligible; in addition, one on P9904 and five on P9905 were found to be inevaluable. The DI randomization included 1396 patients; 431 patients on P9904 who were SR with ETV6-RUNX1 fusion and 965 patients on P9905 who were either NCI SR without favorable cytogenetics or NCI HR patients who did not meet the more restrictive criteria of high risk according to the Shuster age/WBC/gender criteria, with or without favorable genetics. All 1879 eligible patients, including the P9904 patients with favorable trisomies, entered the MTX randomization. Patient characteristics are presented by regimen (Table 3A, B).

Table 3A.

Patient characteristics for DI vs. No DI

Patient characteristics	Not randomized		Randomized
	With DI (N = 86)a	Without DI (N = 397)b	With DI (N = 689)	Without DI (N = 707)
Gender
Male	40	205	341	373
Female	46	192	348	334
Race
White	46	319	544	547
Black or African American	16	19	49	50
Native Hawaiian or other Pacific Islander	0	4	3	9
Asian	5	15	23	26
American Indian or Alaska native	2	5	4	5
Other	5	10	15	24
Unknown	12	25	51	46
Ethnicity
Hispanic or Latino	22	80	139	137
Not Hispanic or Latino	59	297	537	552
Unknown	4	21	13	18
CNS status
CNS 1	76	367	639	646
CNS 2	10	30	50	61
CNS 3	0	0	0	0
Testicular
No	41	218	375	407
Yes	0	0	0	0
\NA	45	179	314	300
Age at diagnosis
<10 years	63	397	587	598
10+ years	23	0	102	109
Initial WBC
<50,000/μl	67	397	604	623
≥50,000/μl	19	0	85	84
NCI risk group
Standard risk	46	396	509	516
High risk	40	0	180	191
Congenital abnormality
Unknown	0	0	1	0
None	47	292	479	469
Down syndrome	2	2	22	23
Other	4	7	8	12
Cytogenetic features
t(1;19)(q23;p13.3)	73	0	2	0
ETV6-RUNX1	7	0	260	260
Trisomies 4 & 10	7	397	66	73

CNS central nervous system, DI delayed intensification, NCI National Cancer Institute, WBC white blood cell

^aPatients with TCF3-PBX1 or high-risk patients as defined by the Shuster Pragmatic High-Risk Group who had a favorable genetic lesion (trisomy 4 and 10 or ETV6-RUNX1) fared poorly on prior studies so were non-randomly assigned to the DI. The poor outcome for these two groups on prior POG trials [12, 18]

^bNCI SR patients with trisomies of 4 and 10 fared well during prior trials and were assigned to receive therapy without a DI phase [17]

Table 3B.

Patient characteristics for 24 h MTX vs. 4 h MTX

Patient characteristics	24 h MTX (N = 939)	4 h MTX (N = 940)
Gender
Male	476	483
Female	463	457
Race
White	714	743
Black or African American	72	62
Native Hawaiian or other Pacific Islander	8	8
Asian	41	28
American Indian or Alaska native	9	7
Other	27	27
Unknown	68	65
Ethnicity
Hispanic or Latino	188	190
Not Hispanic or Latino	726	719
Unknown	25	31
CNS status
CNS 1	859	869
CNS 2	80	71
CNS 3	0	0
Testicular
No	517	524
Yes	0	0
NA	422	416
Age at diagnosis
<10 years	825	820
10+ years	114	120
Initial WBC
<50,000/μl	846	845
≥50,000/μl	93	95
NCI risk group
Standard risk	736	731
High risk	202	209
Congenital abnormality
Unknown	0	1
None	640	648
Down syndrome	20	29
Other	9	22
Cytogenetic features
t(1;19)(q23;p13.3)	38	36
ETV6-RUNX1	261	267
Trisomies 4 & 10	269	274

Impact of the delayed intensification

There was a statistically significant improvement in CCR for patients randomized to receive a DI phase, though there was no difference in overall survival. The 10-year CCR for patients randomized to receive DI was 81.8 ± 2.2% vs. 75.5. ± 2.5% for those treated without DI (p = 0.0016). Overall survival for those randomized to receive DI was 91.4 ± 1.6% vs. 90.9 ± 1.7% at 10 years, p = 0.2518. Among the NCI SR patients, there was a significant difference in CCR (p = 0.0014) (Table 4), favoring the DI. While the trial was not powered to look at subsets, NCI SR patients with the ETV6-RUNX1 fusion and P9905 NCI HR patients with and without favorable cytogenetics showed trends for improvements with DI that were in the same direction as the overall results (Table 4). Both the proportion of relapses occurring within three years of diagnosis and the distribution of sites of relapse were similar between the DI and no DI regimens. Isolated marrow (51.9 and 53.3% with no DI, DI), combined marrow (20 and 22.8% with no DI, DI) and isolated CNS (24.4 and 15.8% with no DI, DI) accounted for ~95% of relapses (p = 0.44). Day 29 MRD was a powerful predictor of outcome (Fig. 2a, b), with a significant difference in outcome, among the MRD-negative patients, with and without the administration of a DI (p = 0.0073).

Table 4.

Summary of outcomes by cohort

	No DI		DI
Cohort	10 year CCR	N	10 year CCR	N	One-sided p-value*
Overall (P9904/P9905 NCI SR and HR)	75.5 ±2.5%	707	81.8 ±2.2%	689	0.0016
P9904/P9905 NCI standard riska	80.0 ± 2.7%	505	86.6 ±2.3%	499	0.0014
P9904 NCI standard risk with ETV6/RUNX1	86.9 ± 3.7%	216	88.0±3.3%	215	0.3411
P9905 NCI standard risk	75.2 ±3.6%	300	85.2 ±3.0%	294	0.0007
P9905 NCI high risk	63.1 ± 5.7%	191	68.5 ±5.5%	180	0.1560

CCR continuous complete remission, DI delayed intensification, HR high risk, NCI National Cancer Institute, SR standard risk

^*p-values for the log-rank test

^aNCI std risk pts with trisomies of chromosomes 4 and 10 or t(1;19) were excluded from the DI randomization. Those with the trisomies did not receive the DI and those with t(1;19) received the DI

Fig. 2.

a Overall continuous complete remission (CCR) based on delayed intensification (DI) and end of induction minimal residual disease status (MRD). b CCR among NCI standard risk patients based on DI and MRD status

Impact of methotrexate infusion

The CCR rates at 10 years, for the two MTX infusions [4-h-80.8 ± 1.9% (n = 940); 24-h-81.4 ± 1.9% (n = 939) (p = 0.7780), were overlapping, though there was a significant difference in survival, favoring the 24-h infusion [4-h-91.1 ± 1.4% (n = 940); 24-h-93.9 ± 1.2% (n = 939) (p = 0.0209). Relapse events more likely to respond to salvage therapy, i.e., isolated extramedullary and late relapses, were not more common among those receiving MTX over 24-h. There were no statistically significant differences in acute toxicities associated with the 4- or 24-h infusions.

Neither analyses of the NCI SR or NCI HR subgroups nor cytogenetic or MRD subsets revealed significant differences in CCR associated with the MTX randomization (data not shown). The accompanying trial, ACCL01P3, was designed to assess parental care giving demands, based on the length of the MTX infusion. Differences in parental demands could not be clearly assessed since 67% (56/84) patients randomized to the 4-h infusion received their therapy in the hospital. The most common reason cited was the lack of an outpatient facility able to deliver the IV MTX and the 8 h of post-MTX IV hydration recommended in the protocol [24].

Relationship between intensification and minimal residual disease

MRD is prognostic irrespective of whether or not patients received the DI (Fig. 2a, b). Randomized patients with ≥ 0.01% MRD at end induction had a 10-year CCR of 44 ± 8% without the DI and 54.3 ± 8% with the DI (p = 0.1119) while those who were MRD negative, had a 10-year CCR of 82 ± 2.5% without the DI vs. 87.7 ± 2.2% with the DI (p = 0.0073). DI improved the 10-year CCR rate in both the MRD-positive and -negative subgroups though the difference only reached statistical significance in the MRD-negative subset, 85.9 ± 2.7% vs. 91.4 ± 2.1% (p = 0.0068), without and with the DI.

Toxicity

P9900-induction

There were two deaths among the first 34 patients enrolled on the original dexamethasone-based four-drug induction and 11 (0.9%) deaths among the first 1263 patients treated with the three-drug induction. The excessive toxicity led to the treatment modifications described above. Following these changes, there were two deaths from infection among 527 patients (0.38%).

Post-induction therapy

The DI phase was associated with more infectious complications with 210/748 (28%) of the DI courses (Reg C + D) associated with grade 2–4 infections vs. 161/1084 (14.9%) of the IV MTX courses (Reg A + B). (p < 0.0001). There were no significant differences in toxicities associated with the two MTX infusion schedules.

There were two deaths in remission, both secondary to Gram-negative sepsis; one during DI (Reg C) and one during maintenance (Reg B). CNS ischemia and seizures were rare during consolidation, occurring in fewer than 0.6% of the patients treated on any regimen.

Second malignancies

Fourteen patients developed a second malignancy (SMN) or lymphoproliferative disorder; nine of whom developed acute myelogenous leukemia (AML) or myelodysplasia (MDS) (five with monosomy seven. Four developed a B-cell lymphoma/lymphoproliferative disorder (three EBV-positive); one patient a peripheral T-cell lymphoma. Five of the 14 (three with MDS/AML) received a DI. Ten of the 14 SMN developed during therapy, two were at the end of therapy and two patients developed MDS 41 and 72 months from diagnosis. There was no difference in the cumulative incidence rates (CI) between the randomized No DI and DI groups (10-year CI of 5.8 ± 3% vs. 5.9 ± 3%), p = 0.97.

Discussion

Previous POG ALL trials focused on anti-metabolite regimens for intensification while a multi-drug re-induction/reconsolidation or DI was used by many cooperative groups (reviews in refs. [1, 3, 6]). Originally described by Rheim et al. [9], the use of a DI (Protocol II) was based on two models of resistance. The Goldie-Coldman model [25] postulates that cells acquire chemoresistance early from the accumulation of mutational events and that the chances of doing so are related to tumor burden. The Norton-Simon hypothesis [26] proposes that when disease burden has been significantly reduced, the remaining cells are relatively resistant to therapy. Therefore, an aggressive re-induction and late intensification, incorporating agents with incomplete cross-resistance, may be beneficial. The DI phase in our study included dexamethasone instead of prednisone, doxorubicin instead of daunomycin. and 6-thioguanine replaced 6-mercaptopurine.

This trial asked whether anti-metabolite and delayed intensification strategies for intensification targeted similar subgroups of tumor cells in clinically and biologically defined subtypes of patients or whether the addition of a DI would improve the outcome of children treated with parenteral MTX intensification by reducing tumor burden further in the population or by targeting additional sub-populations of cells. Our results demonstrate that the addition of a DI phase to treatment that included six courses of IV MTX improved CCR without excessive short-term toxicity. Overall survival (OS) at 10 years did not differ with the DI. This likely reflects superior salvage among patients who are younger at initial diagnosis as 85% of our patients were <10 at initial diagnosis [27–29].

Parenteral MTX has also been incorporated into multiple therapies for children with ALL, with its addition associated with a decrease in bone marrow and testicular relapses and an improvement in EFS [5, 30, 31]. Lower escalating dose MTX was superior to oral MTX in patients with NCI SR ALL on CCG 1991 and the augmented BFM regimen [32], developed by the CCG, has significantly enhanced outcome for all patients with higher risk ALL, regardless of age or immunophenotype. Methotrexate exposure in the augmented regimen is significantly enhanced vs. the standard BFM regimen, both because of a higher dose/m² of systemic MTX, prior to continuation therapy (60 vs. 980 to 1380 mg/m²), and because the augmented regimen uses parenteral drug; avoiding loss of effect because of the poor bioavailability of oral MTX. Recently, the COG AALL0232 HR B-ALL trial demonstrated an advantage for high dose as opposed to lower escalating dose MTX [33]. The current trial did not demonstrate a difference in acute toxicity or CCR associated with the delivery of six courses of parenteral MTX infused over either 4 h (2 g/m²) or 24-h (1 g/m²) though there was a difference in survival favoring the 24 h infusion.

Both the original four-drug induction using dexamethasone and an anthracycline and the original three-drug induction were associated with excessive toxicity. Coin-cidently with the two deaths and two “near misses” on the four-drug, dexamethasone-based induction (34 patients enrolled), a paper describing [34] excessive mortality associated with a dexamethasone-based induction, at the Dana Farber Cancer Institute, was published, supporting the change from dexamethasone to prednisone for the NCI high-risk patients. All of the individual components of the three-drug, dexamethasone-based induction had been safely utilized in previous protocols though the specific combination of these standard agents (dexamethasone, vincristine, asparaginase, and IT MTX) had not been used until the pilot P9705 wherein there were no deaths among the small number of patients enrolled (n = 59). The two changes made in the three-drug induction regimen resulted in a decrease in induction mortality, though the difference did not reach statistical significance.

This COG 9904/9905 trial demonstrated a CCR advantage for the incorporation of a delayed intensification phase and an OS advantage for the 24 vs. 4-h MTX infusion. Previously, CCG 1922 [35] randomized NCI SR patients to a dexamethasone or prednisone-based induction with a single DI phase, whereas the CCG 1952 SR ALL trial [36] included two DI phases in the context of a prednisone-based induction. Comparing these two trials, patients receiving a dexamethasone-based induction with one DI had an event free survival comparable to that for standard risk patients receiving two DI’s after a prednisone-based induction (85% in both trials). CCG 1991 then demonstrated an advantage for parenteral vs. oral MTX during the interim maintenance phase of therapy. In that study, NCI SR patients receiving a dexamethasone-based induction and parenteral MTX had a better outcome with the addition of the DI phase.

AIEOP-BFM ALL 2000 [37] randomized patients without unfavorable cytogenetic features who were MRD negative on days 33 and 78 to receive a standard vs. reduced intensity DI. Although there was an overall diminution in disease-free survival associated with the reduced intensity therapy, there was no significant difference for the young patients with an ETV6-RUNX1 translocation. This trial supports their findings with no difference in outcome among NCI SR patients with the ETV6/RUNX1 fusion. These lower risk NCI SR patients, with favorable cytogenetics features, who are MRD negative in peripheral blood at both day 8 and in marrow on day 29, have been shown in previous P9904 analyses to have a 5-year CCR of 97 +/− 1% [38]. This subset likely can be cured without a DI, with therapy devoid of anthracyclines and alkylating agents; an observation that has significant implications for ALL therapy here and in low- and middle-income countries. Conversely, the addition of DI to the therapeutic regimen did not produce a statistically significant improvement in outcome for MRD-positive patients, though the trend favored the administration of the DI. The absence of a significant difference reflects a lack of power in subset analysis, and may be secondary to the differences between the DI in this study and those incorporated in BFM and CCG protocols [9, 32]. Dexamethasone was given at a dose of 6, not 10 mg/m²/day and daunomycin was used instead of doxorubicin. More importantly, current trials incorporate analyses that identify recurrent, prognostically significant genetic abnormalities as well as previously unidentified leukemias with high-risk lesions, including Philadelphia chromosome-like disease [39–41]; offering targeted therapies that may improve outcome [42]. Future trials will incorporate novel approaches to therapy for patients with persistent disease at the end of consolidation [43].