Impact of Initial CSF Findings on Outcome Among Patients With National Cancer Institute Standard- and High-Risk B-Cell Acute Lymphoblastic Leukemia: A Report From the Children’s Oncology Group

Abstract

Purpose

To determine the prognostic significance of blasts, and of white and red blood cells, in CSF samples at diagnosis of acute lymphoblastic leukemia (ALL), a uniform CSF and risk group classification schema was incorporated into Children’s Oncology Group B-cell ALL (B-ALL) clinical trials.

Methods

CSF status was designated as follows: CNS1, no blasts; CNS2a to 2c, < 5 WBCs/μL and blasts with/without ≥ 10 RBCs/μL or ≥ 5 WBCs/μL plus blasts, with WBCs ≥ 5 times the number of RBCs; CNS3a to 3c, ≥ 5 WBCs/μL plus blasts with/without ≥ 10 RBCs/μL or clinical signs of CNS disease. CNS2 status did not affect therapy; patients with CNS3 status received two extra intrathecal treatments during induction and augmented postinduction therapy with 18 Gy of cranial radiation.

Results

Among 8,379 evaluable patients enrolled from 2004 to 2010, 7,395 (88.3%) had CNS1 status; 857 (10.2%), CNS2; and 127 (1.5%), CNS3. The 5-year event-free and overall survival rates were, respectively, 85% and 92.7% for CNS1, 76% and 86.8% for CNS2, and 76% and 82.1% for CNS3 (P < .001). In multivariable analysis that included age, race/ethnicity, initial WBC, and day-29 minimal residual disease < 0.1%, CSF blast, regardless of cell count, was an independent adverse predictor of outcome for patients with standard- or high-risk disease according to National Cancer Institute criteria. The EFS difference reflected a significant difference in the incidence of CNS, not marrow, relapse in patients with CNS1 versus CNS2 and/or CNS3 status.

Conclusion

Low levels of CNS leukemia, regardless of RBCs, predict inferior outcome and higher rates of CNS relapse. These data suggest that additional augmentation of CNS-directed therapy is warranted for CNS2 disease.

INTRODUCTION

Prevention of CNS relapse dramatically increases cure rates for children with acute lymphoblastic therapy (ALL).^1,2 Event-free survival (EFS) and overall survival (OS) have continued to improve through the application of detailed, risk-based classification systems and large randomized, therapeutic trials. In parallel, the intensity of systemic and intrathecal (IT) therapy has increased, and the use of cranial irradiation has largely been eliminated.^3,4

Before Children’s Oncology Group (COG) ALL trials opened, an analysis of clinical, biologic, and early-response data from > 6,000 patients enrolled in legacy Pediatric Oncology Group and Children’s Cancer Group (CCG) trials from January 1986 to November 1999 linked the presence of blasts in the diagnostic CSF sample with an inferior EFS,⁵ as did reports from St Jude Children’s Research Hospital (SJCRH),⁶ the Berlin-Frankfurt-Münster (BFM) cooperative group,⁷ and a Dutch⁸ cooperative group. The best approach to therapy for those with CNS2 involvement was unclear, though, and the data did not reflect the incorporation of current risk stratification schemas or optimal therapies (Table 1). Specifically, minimal residual disease (MRD) had not been included in treatment assignment; dexamethasone was not universally administered during induction in patients with standard-risk (SR) disease according to National Cancer Institute (NCI) criteria, ⁹ and augmented postinduction therapies were not administered to all patients with high-risk (HR) disease according to NCI criteria.¹⁰ Also, the effect of a traumatic lumbar puncture with blasts (TLP+) versus a low blast count without red cells was unclear.¹¹ Accordingly, the first COG trials did not alter therapy for patients with CNS2 disease. Patients with CNS3 disease received augmented treatment with two extra IT doses during induction and augmented postinduction therapies with cranial irradiation. A uniform CSF classification system, which was based on the presence of lymphoblasts and the numbers of red and white cells in the diagnostic CSF sample, was incorporated. These data were captured to allow future assessment of the impact of CNS status on outcome. This report describes the relationship between diagnostic CSF findings and outcome for 8,379 patients with NCI-categorized SR and HR B-cell ALL (B-ALL) treated during COG trials AALL0331 and AALL0232. These data document the adverse prognostic effect of initial CNS2 and CNS3 status regardless of the presence of red cells.

Table 1.

Definitions

PATIENTS AND METHODS

From December 2003 to January 2011, patients age 1 to 30 years were enrolled in the COG ALL classification trial, AALL03B1. Eligible patients with B-ALL subsequently were enrolled in therapeutic trials AALL0331¹² (NCT001103285, open from April 2005 to May 2010) NCI-categorized SR disease or in AALL0232¹³ (open from December 2003 to January 2011) for NCI-categorized HR disease (Tables 1 and 2; Appendix Table A1, online only). Early response was assessed by bone marrow morphology determined locally on days 8, 15, and 29 of induction and through centralized flow-based MRD levels on day 29.^14,15 Rapid early responders (RERs) had an M1 marrow (< 5% blasts) by day 15 of induction therapy and had day-29 MRD levels < 0.1%. Slow early responders (SERs) had an M2 (5% to 25% blasts) or M3 (> 25% blasts) marrow at day 15 of induction and/or had MRD levels > 0.1% on day 29. Patients with very-high-risk features (Table 1) were removed from protocol therapy after induction. This analysis included 8,379 eligible and evaluable patients from AALL0232 (n = 3,019) and AALL0331 (n = 5,360). Local CSF cell counts and Wright-stained cytomorphology interpretations were used by participating institutions to determine CNS status.

Table 2.

Patient Characteristics From COG Protocols AALL0331 and AALL0232

Treatment and Definitions

CSF classification and details of the treatment protocols are listed in Table 1 and in Figs 1A and 1B. Additional information is available elsewhere (AALL0331 [NCT001103285] and AALL0232).¹³ In AALL0331, only patients with CNS1 disease status were allocated to the postinduction low-risk arm. Patients with CNS3 disease were nonrandomly assigned to the HR arms of AALL0232 and AALL0331. Patients in both trials who were RERs received one interim-maintenance (IM) and one delayed-intensification (DI) phase; those who were SERs received two IM and DI phases. Patients with NCI-categorized SR, CNS3 disease enrolled in AALL0331 received a three-drug, dexamethasone-based induction with IT methotrexate on days 1,8,15, 22, and 29 followed by full augmented BFM therapy with two IM and two DI courses before maintenance; both IM phases included escalated-dose intravenous (IV) methotrexate without leucovorin rescue plus pegaspargase. Cranial radiation (18 Gy) was administered during the second DI phase. Patients with CNS3 disease enrolled in AALL0232 initially were nonrandomly assigned to receive dexamethasone (10 mg/m²/d on days 1 through 14) during induction and high-dose methotrexate plus leucovorin rescue during the first IM phase. During the trial, the corticosteroid assignment was changed to prednisone (60 mg/m²/d on days 1 through 28) for those patients age 10 years or older because of excessive osteonecrosis; younger patients continued to receive dexamethasone.^13,16 Like the patients with NCI-categorized SR, CNS3 disease, those with HR, CNS3 disease received weekly IT methotrexate during induction, two IM and two DI phases, and 18 Gy of cranial radiation delivered during the second DI phase. Therapy was not modified for CNS2 disease.

Fig 1.

Treatment schemas for Children’s Oncology Group protocols AALL0331 and AALL0232. (A) AALL0331: Patients with standard-risk–low disease had trisomies of chromosomes 4,10, and 17 or had an ETV6-RUNX1 fusion (ie, FAV genetics) with a day-15 (or day-8) M1 marrow and a day-29 minimal residual disease (MRD) level less than 0.1%; CNS1, no testicular disease, no very-high-risk (VHR) features. These patients were randomly assigned to standard therapy or standard therapy plus PEG-asparaginase (PEG). Patients with standard-risk–average disease did not have FAV genetics unless the status was CNS2 at diagnosis, and they had a day-15 (or day-8) M1 marrow and a day-29 MRD level less than 0.1%; no CNS3 or testicular disease; and no VHR features. These patients were randomly assigned according to a 2 × 2 factorial design to standard therapy with or without intensified consolidation (ie, intensive) or with or without intensified postconsolidation therapy with augmented interim maintenance (IM; ie, Aug IM1) and delayed intensification (DI; ie, Aug2 DI) until June 2008, when results of Children’s Cancer Group (CCG) trial CCG1991 demonstrated the superiority of escalated-dose intravenous (IV) methotrexate (MTX) compared with oral MTX. At that point, all patients received IV MTX, and the intensified postconsolidation comparison was closed. Patients with standard-risk–high disease had CNS3 disease at diagnosis, were pretreated with corticosteroids, had an MLL rearrangement, and were rapid early responders (RERs), or were slow early responders (SERs) without VHR features. After induction, these patients received intensive consolidation, two augmented IM phases, and two augmented DI phases before the maintenance phase. Patients with CNS3 status received cranial radiation during the second DI phase. The first IM phase was changed to high-dose MTX (HDMTX) in May 2011, when HDMTX proved superior to escalated-dose MTX in AALL0232. After induction, patients with t(9;22)(q34;q11); fewer than 44 chromosomes; and/or a DNA index less than 0.81, ≥ 1% MRD, or an M2 marrow after extended induction, MLL with SER status, and/or an M3 marrow on day 29 were not eligible to continue therapy in AALL0331 or AALL0232. Extended induction was administered to patients with an M2 marrow or a MRD level greater than 1% at the end of induction. Patients with an M1 marrow and less than 1% MRD at the end of extended induction continued with therapy as SERs. (B) AALL0232: As published,¹³ all patients except those with CNS3 or testicular disease at diagnosis or corticosteroid pretreatment were randomly assigned in a 2 × 2 factorial design to dexamethasone (Dex) or to prednisone during induction and to escalated-dose MTX or HDMTX during IM1 until June 2008, when the corticosteroid random assignment was restricted to patients younger than age 10 years at diagnosis because of excessive avascular necrosis in older patients who received dexamethasone. The MTX random assignment closed when HDMTX demonstrated greater efficacy (January 2011). Patients with either CNS3 or testicular disease at diagnosis, corticosteroid pretreatment, an MLL rearrangement, and RER status and patients who were SERs received two IM and two DI phases. Patients with CNS3 disease received 18 Gy cranial radiation, and SERs received 12 Gy. Extended induction was given to patients with an M2 marrow or greater than 1% MRD at the end of induction. Patients with an M1 marrow and less than 1% MRD at the end of extended induction continued to receive therapy as SERs. Ara-C, cytarabine; Cyclo, cyclophosphamide; D, day; Dauno, daunomycin; Doxo, doxorubicin; IT, intrathecal; TG, thioguanine; VCR, vincristine.

Bone marrow relapse was defined as an M3 marrow any time after complete remission was attained. CNS relapse was defined as positive cytomorphology and ≥ 5 WBCs/μL in the CNS or positive cytomorphology with CSF levels of 0 to 4 WBCs/μL on two successive occasions at least 1 month apart. CNS relapse was considered isolated if the marrow was M1 at relapse and was considered combined if the marrow was M2 or M3.

Study Design and Statistical Analysis

Patients with locally defined CNS1; CNS2a, 2b, or 2c; and CNS3a, 3b, or 3c disease enrolled in ALL03331 and AALL0232 were included. EFS was defined as the time from study entry to first event (induction failure, induction death, relapse, second malignancy, remission death) or date of last follow-up for event-free patients. OS was defined as time from study entry to death or date of last follow-up. Survival rates were estimated with the Kaplan-Meier method; standard errors were calculated according to Peto et al.^17,18 Survival curves were compared with the log-rank test. Cumulative incidence rates (CIRs) were computed with the cumulative incidence function for competing risks, and comparisons were made using the K-sample test.¹⁹ Multivariable analysis (Cox proportional hazards model) was used to identify independent prognostic factors. Comparison of proportions between groups used the χ² test. P < .05 was considered statistically significant. All analyses were performed with SAS software (version 9.4; SAS Institute, Cary, NC). Graphics were generated with R version 2.13.1 (http://www.r-pproject.org).

RESULTS

Demographics

CNS1 status was more common among patients with NCI-categorized SR disease and those with MRD < 0.1% on day 29. White and black patients were more likely to have CNS1 disease than others; Hispanics had more CNS2 disease than non-Hispanics (Table 2). CNS2a status was more common than CNS2b or 2c disease among patients with NCI-categorized SR disease, patients who were white, those who were younger than age 10 years, and those with < 50,000/μL WBCs (Appendix Table A1).

EFS and OS

The 5-year EFS and OS rates for the 7,214 patients with NCI-categorized SR and HR disease without blasts in the diagnostic CSF sample (ie, CNS1) were significantly better than for others: 85% (standard error [SE], 0.6%) and 92.7% (SE, 0.4%) versus 76% (SE, 2%) and 86.8% (SE, 1.6%) for CNS2 (n = 836) and 76% (SE, 5.0%) and 82.1% (SE, 4.7%) for CNS3 (n = 124; P < .001 for EFS and OS; Figs 2A and 2B). When analyzed by NCI risk group, patients with NCI-categorized SR or HR CNS1 disease had statistically significant superior outcomes (Table 3). Multivariable analysis with Cox proportional hazards regression demonstrated that age, WBC at diagnosis, race/ethnicity, MRD at induction on day 29, and CNS1 versus CNS2 or CNS3 status were independent predictors of EFS and OS (Appendix Tables A2 and A3, online only).

Fig 2.

The 5-year (A) event-free survival (EFS) and (B) overall survival (OS) for patients with standard- and high-risk disease according to National Cancer Institute criteria (combined risks) by CNS status at diagnosis. The 5-year EFS and OS rates for patients with CNS1 status were significantly better than for others: 85% (SE, 0.6%) and 92.7% (SE, 0.4%), respectively, versus 76% (SE, 2%) and 86.8% (SE, 1.6%), respectively, for CNS2 and 76% (SE, 5%) and 82.1% (SE, 4.7%), respectively, for CNS3 (P < .001 for EFS and OS).

Table 3.

EFS and CIRs of Relapse by CNS Status at Diagnosis

Among the CNS2 categories, there was no significant difference in the impact on EFS or OS, or in the CIR of isolated or combined CNS relapse when CNS2a status with blasts without RBCs was compared to TLP+ status, despite a significantly different distribution of CNS2a versus 2b or 2c status (patients with CNS2a status were more likely to be white and to have NCI-categorized SR disease; Appendix Tables A1 and A4, online only; Figs 3A and 3B).

Fig 3.

The 5-year (A) event-free survival (EFS) and (B) overall survival (OS) for patients with standard- and high-risk disease according to National Cancer Institute criteria with blasts with and without red blood cells in the diagnostic spinal fluid. The 5-year EFS and OS rates were not significantly different for patients with and without a traumatic lumbar puncture at diagnosis: 74.5% (SE, 2.6%) and 86.8% (SE, 2.0%), respectively, for CNS2a versus 79.8% (SE, 3.3%) and 86.9% (SE, 2.8%), respectively, for CNS2b and 72.2% (SE, 6.0%) and 86.2% (SE, 4.7%), respectively, for CNS2c.

Overall, the inferior outcome associated with blasts in the diagnostic CSF sample was attributable largely to an increase in the risk of CNS relapse (Appendix Table A4; Figs 3A and 3B, and Figs 4A-4C). The differences in the risk of isolated marrow relapse were not statistically significant among patients with CNS1, CNS2, or CNS3 disease or in a comparison of patients with CNS1 or CNS2 status (5.7% [SE, 0.3%] v 6.5% [SE, 0.9%], respectively; P = .1811; Fig 4C). Overall, patients with CNS1, CNS2, and CNS3 disease had cumulative incidences of isolated CNS relapse of 2.0 (SE, 0.2%), 5.6% (SE, 0.8%), and 5.1% (SE, 2.0%), respectively (P < .001) and had cumulative incidences of combined CNS relapse of 2.8% (SE, 0.2%), 7.7% (SE, 0.97%), and 5.1% (SE, 2.0%), respectively (P < .001). The cumulative incidences of marrow relapse were 5.7% (SE, 0.3%), 6.5% (SE, 0.9%), and 9.3% (SE, 2.9%), respectively (P = .09; Table 3). Sixty of 836 patients with CNS2 disease experienced CNS relapse (isolated plus combined); 40 patients had ≥ 5 WBCs/μL in the CSF with blasts, 20 had < 5 WBCs/μL in the CSF with blasts detected by morphology in two consecutive CSF samples that were obtained 4 weeks apart. These subsets were too small to analyze separately.

Fig 4.

(A) The cumulative incidence of relapse (CIR) that involved the CNS among patients with standard- and high-risk disease according to National Cancer Institute criteria by CNS status at diagnosis. (B) The CIR rates for combined CNS relapses were 2.8% (SE, 0.2%), 7.7% (SE, 0.97%), and 5.1% (SE, 2.0%; P < .001), for patients with CNS1, CNS2, and CNS3 statuses, respectively. For isolated CNS relapse, the values were 2.0% (SE, 0.2%), 5.6% (SE, 1.8%), and 5.1% (SE, 2.0%; P < .001), for patients with CNS1, CNS2, and CNS3 statuses, respectively. (C) The cumulative incidence of isolated bone marrow relapse among patients with standard- and high-risk disease by CNS status at diagnosis. The cumulative incidence rates of isolated marrow relapse among patients with CNS1, CNS2, and CNS3 statuses at diagnosis were 5.7% (SE, 0.3%), 6.5% (SE, 0.9%), and 9.3% (SE, 2.9%), respectively (P = .0874).

Subset Analyses: Potential Impact of Specific Therapies on Outcome

In the CCG 1991 trial, patients with NCI-categorized SR disease were randomly assigned to receive oral or escalated-dose IV methotrexate during two IM phases of therapy. When analysis demonstrated a superior outcome for those treated with IV methotrexate,²⁰ AALL0331 was amended to incorporate IV methotrexate during IM for all patients. For the 1,354 patients with CNS1 or CNS2 status who received IV methotrexate after this amendment (Appendix Table A4), there continued to be a significant difference in the cumulative incidence of combined CNS relapse, which was 2.1% (SE, 0.44%) for CNS1 (n = 1,250) and 5.8% (SE, 2.3%) for CNS2 (n = 104; P = .017). The CIR of isolated CNS relapse was no longer significant. The difference in EFS rates at 4 years also remained significant: 93.5% (SE, 1.2%) and 88% (SE, 5.4%) for CNS1 and CNS2 disease, respectively. Importantly, the small subset (n = 81) of otherwise low-risk patients with NCI-categorized SR disease who had favorable blast cell cytogenetics, peripheral-blood MRD less than 1% on induction day 8, and day-29 marrow MRD < 0.1% had a projected 5-year EFS rate of 95.3% (SE, 3.5%) despite CNS2 status, compared with a projected 5-year EFS rate of 95.2% (SE, 0.95%) for the low-risk patients with CNS1 disease.

In the AALL0232 trial of HR ALL,¹³ patients younger than age 10 years were randomly assigned to receive dexamethasone or prednisone during induction and high-dose methotrexate with leucovorin rescue versus escalated-dose methotrexate with pegaspargase during the first IM phase. Ultimately, the dexamethasone and high-dose methotrexate arm proved superior for patients younger than age 10 years who received this therapy. There was neither a significant difference in outcome for those with CNS2 status, with 5-year EFS rates of 84.0% (SE, 3.5%) for the 216 patients with CNS1 disease and 91.2% (SE, 5.9%) for the 46 patients with CNS2 disease (P = .519), nor a difference in the CIR of isolated or combined CNS relapse (Appendix Table A4). In AALL0232, 518 patients with CNS1 disease and 95 patients with CNS2 disease who had an SER received 12 Gy of cranial radiation and a second IM and DI phase. Among these patients, CNS2 continued to be associated with a significant increase in the risk for isolated, though not combined, CNS relapse (Appendix Table A4). The difference in EFS rate (52.3% [SE, 3%] v 44.4% [SE, 6.6%]; P = .11) was not significant; however, in this small subset, the trend favored those with CNS1 status.

DISCUSSION

This is, to our knowledge, the largest analysis of the impact of diagnostic CNS status in patients with ALL ever reported; it provides enhanced power to examine the impact of CNS status and the relative impact of TLP. Patients received risk-based stratification for postinduction therapy on the basis of clinical variables, lymphoblast genetics at diagnosis, and early response to therapy. CNS-directed therapy was augmented only for CNS3 disease or for patients who were SERs and who had NCI-categorized HR disease. These two groups were treated with full augmented BFM therapy and received cranial radiation doses of 12 Gy (SERs) or 18 Gy (CNS3). Lymphoblasts in the diagnostic CSF sample were predictive of an inferior outcome with a statistically significant increase in the cumulative incidence of CNS relapse among the NCI-categorized SR and HR populations. Both CNS2 and CNS3 statuses at diagnosis were associated with an inferior outcome; importantly, the outcome of patients with TLP and blasts in their diagnostic CSF sample was equivalent to those with blasts without red cells.

Though CNS2 and CNS3 statuses were associated with an increase in the relative risk of relapse in the legacy CCG trials, the applicability of those data to the COG AALL0232 and AALL0331 trials were unclear when these trials were designed. Specifically, CNS2 disease was associated with a lower 5-year EFS rate (67% v 78.4% with CNS1 and 76.2% with CNS3),⁵ but neither blast cell cytogenetics nor MRD had been used for risk stratification in the legacy trials. As detailed by Levinsen et al,²¹ patients with CNS involvement at diagnosis are likely to have additional high-risk features that, if identified at diagnosis, would alter treatment allocation and outcome. Thus, it was hypothesized when COG trials AALL0232 and AALL033 were designed that the prognostic significance of CNS2 status might be abrogated through the universal delivery of dexamethasone for NCI-categorized SR disease and augmented postinduction therapies for patients with HR disease, in the context of a COG classification that included blast cell cytogenetics and end-induction MRD for treatment assignment. This report confirms that CNS disease at diagnosis is associated with other higher-risk features: specifically, older age, a higher initial white blood cell count, and end induction MRD positivity. It also demonstrates the continued negative impact of CSF blasts in the diagnostic CSF sample, regardless of whether the lumbar puncture was traumatic and despite the treatment refinements used in COG trials AALL0232 and AALL0331. The decrease in the 5-year EFS rate for patients with CNS2 status in these trials was largely attributable to a statistically significant increase in CNS relapse.

CNS2 status was associated with an inferior outcome on an early treatment protocol that delivered only six doses of IT therapy but that included cranial irradiation for 81% of the patients with CNS2 status.²² Subsequently, the BFM,⁷ Dana-Farber Cancer Institute Consortium (DFCI),²³ and SJCRH⁶ groups described the outcomes for patients with CNS1, CNS2, or CNS3 status; the BFM and SJCRH groups also have evaluated the impact of TLP with blasts. CNS2 status was not associated with an inferior outcome in any of these trials, nor was it prognostic in the NOPHO 2000 trial (K. Schmiegelow, personal communication, September 2015).

However, TLP+ continued to be associated with inferior EFS in the BFM-ALL95²⁴ and the Total Therapy Study XV³ trials, and CNS3 disease remained a harbinger of poor outcome in all but the DFCI 00-01 trial.^25,26 The improvement in outcome for the patients who had CNS2 status compared with those who had CNS1 status who were treated in the BFM, DFCI, NOPHO, and SJCRH trials has been largely attributable to use of more intensive IT therapy during induction, though the type of IT therapy used and the schedule of administration vary. There are, however, other factors that may have affected the outcome for patients with CNS2 status in these trials. Specifically, patients with both B-ALL and T-cell ALL were included; also, in the BFM-ALL95,²⁴ DFCI,²⁶ and NOPHO²⁷ trials, those patients with CNS2 status who, by other criteria, also had HR or very-high-risk disease received cranial irradiation. The St Jude Total XV trial³ used chemotherapy only without cranial irradiation, and it involved an increase in the number of doses of IT triple therapy administered to patients with CNS involvement during induction and postinduction therapy. Patients with CNS2 status did as well as those with CNS1 status, though TLP+ predicted a worse outcome.

The current COG trials for newly diagnosed B-ALL (AALL0932 [NCT01190930] and AALL1131 [NCT01406756]) incorporate additional refinements in risk classification that may affect CNS relapse. The flow-based MRD cutoff for more intensive therapy is now 0.01%, not 0.1%; patients with SR disease who have neutral genetics and day-8 peripheral blood MRD ≥ 1% are now treated with therapies used for HR disease after induction, as are those with intrachromosomal amplification of chromosome 21. All patients with NCI-categorized SR disease now receive IV methotrexate during two IM phases, and all patients with NCI-categorized HR disease receive high-dose methotrexate during the first IM phase; those younger than age 10 years receive dexamethasone during induction. Subset analyses presented here suggest that dexamethasone and high-dose methotrexate for those younger than age 10 years may ameliorate the impact of CNS2 status. The impact of the IV methotrexate for patients with NCI-categorized SR disease is less clear; there is a persistent difference in combined CNS relapse in that group. The current HR trial for B-ALL, AALL1131, includes a postinduction random assignment between IT methotrexate and IT triple therapy for those without very-high-risk features. These modifications may improve the outcome for the CNS2 population, and the random assignment question may establish which form of IT therapy is superior. Other regimens have commonly increased the intensity of IT therapy delivery during the induction phase for those with CNS2 status. BFM-ALL95,²⁴ DFCI,²⁶ NOPHO (K. Schmiegelow, personal communication, September 2015), and St Jude¹¹ protocols have abrogated the negative impact of CNS2a status with weekly to twice-weekly IT therapy during induction until blasts clear. Pending the outcomes of the changes/random assignments described in this discussion, current COG trials supplement standard IT dosing with twice-weekly IT cytarabine until three samples are without blasts for patients with CNS2 disease. Though more intensive therapies have overcome the prognostic significance of CNS2 disease, only DFCI 00-01 has negated the negative prognosis associated with CNS3 status; however, this is based on the outcome of only 17 patients for both T-cell ALL and B-ALL.²⁶ The DFCI therapy provided intensive asparaginase therapy to all patients as well as cranial irradiation for those with HR disease. The current COG random assignment between IT methotrexate and IT triple therapy may provide additional information, but the creation of a more efficacious therapy for the CNS3 population, which ideally avoids the use cranial irradiation, remains a challenge.

This evaluation of > 8,000 children and adolescents with B-ALL demonstrates that, in the absence of specific therapy, blasts in the diagnostic spinal fluid have a negative prognostic impact regardless of the presence of red cells. Though the small subset of patients with NCI-categorized SR disease who have favorable blast cell genetics and a good response at the end of induction fare well, CNS involvement at any level is an independent predictor of outcome for other patients and warrants consideration of differential CNS-directed therapy to improve outcomes.

Appendix

Table A1.

Patient Characteristics From COG Protocols AALL0331 and AALL0232

Table A2.

Cox Multivariable Analysis of Likelihood Estimates of Event-Free Survival

Table A3.

Cox Multivariable Analysis of Likelihood Estimates for Overall Survival

Table A4.

Summary Table of Cumulative Incidence Rates of Relapse

AUTHOR CONTRIBUTIONS

Conception and design: Naomi Winick, Meenakshi Devidas, Kelly Maloney, Eric Larsen, Mignon L. Loh, Elizabeth Raetz, Stephen P. Hunger, William L. Carroll

Provision of study materials or patients: Naomi Winick, Kelly Maloney, Eric Larsen, Brent Wood, Mignon L. Loh, Elizabeth Raetz, Stephen P. Hunger, William L. Carroll

Collection and assembly of data: Meenakshi Devidas, Kelly Maloney, Eric Larsen, Michael J. Borowitz, Andrew Carroll, Julie M. Gastier-Foster, Nyla A. Heerema, Cheryl Willman, Brent Wood, Stephen P. Hunger, William L. Carroll

Data analysis and interpretation: Naomi Winick, Meenakshi Devidas, Si Chen, Leonard Mattano, Michael J. Borowitz, Julie M. Gastier-Foster, Cheryl Willman, Brent Wood, Mignon L. Loh, Elizabeth Raetz, Stephen P. Hunger, William L. Carroll

Manuscript writing: All authors

Final approval of manuscript: All authors

Accountable for all aspects of the work: All authors

AUTHORS' DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST

Impact of Initial CSF Findings on Outcome Among Patients With National Cancer Institute Standard- and High-Risk B-Cell Acute Lymphoblastic Leukemia: A Report From the Children’s Oncology Group

The following represents disclosure information provided by authors of this manuscript. All relationships are considered compensated. Relationships are self-held unless noted. I = Immediate Family Member, Inst = My Institution. Relationships may not relate to the subject matter of this manuscript. For more information about ASCO's conflict of interest policy, please refer to www.asco.org/rwc or ascopubs.org/jco/site/ifc.

Naomi Winick

No relationship to disclose

Meenakshi Devidas

Honoraria: Pfizer, Novartis

Si Chen

No relationship to disclose

Kelly Maloney

No relationship to disclose

Eric Larsen

No relationship to disclose

Leonard Mattano

Employment: Pfizer (I)

Stock or Other Ownership: Pfizer, Pfizer (I)

Consulting or Advisory Role: Pfizer, Mylan, Novartis

Michael J. Borowitz

Consulting or Advisory Role: HTG Molecular Diagnostics

Research Funding: Amgen, Becton Dickinson

Travel, Accommodations, Expenses: Beckman Coulter

Andrew Carroll

No relationship to disclose

Julie M. Gastier-Foster

Research Funding: Bristol-Myers Squibb

Nyla A. Heerema

Honoraria: Abbott Vysis

Cheryl Willman

No relationship to disclose

Brent Wood

Honoraria: Amgen, Seattle Genetics

Research Funding: Amgen (Inst), Seattle Genetics (Inst), Juno Therapeutics (Inst), Stemline (Inst)

Travel, Accommodations, Expenses: Amgen, Seattle Genetics

Mignon L. Loh

No relationship to disclose

Elizabeth Raetz

No relationship to disclose

Stephen P. Hunger

Stock or Other Ownership: Express Scripts, Amgen, Merck (I), Amgen (I), Pfizer (I)

Honoraria: Jazz Pharmaceuticals, Spectrum Pharmaceuticals, Erytech Pharma

Patents, Royalties, Other Intellectual Property: Co-inventor on US patent No. 8,568,974 B2: Identification of novel subgroups of high-risk pediatric precursor-B acute lymphoblastic leukemia, outcome correlations and diagnostic and therapeutic methods related to same. It has not been licensed, and there is no income.

William L. Carroll

No relationship to disclose