Improved CNS Control of Childhood Acute Lymphoblastic Leukemia Without Cranial Irradiation: St Jude Total Therapy Study 16

Abstract

PURPOSE

Despite contemporary treatment, up to 10% of children with acute lymphoblastic leukemia still experience relapse. We evaluated whether a higher dosage of PEG-asparaginase and early intensification of triple intrathecal therapy would improve systemic and CNS control.

PATIENTS AND METHODS

Between 2007 and 2017, 598 consecutive patients age 0 to 18 years received risk-directed chemotherapy without prophylactic cranial irradiation in the St Jude Total Therapy Study 16. Patients were randomly assigned to receive PEG-asparaginase 3,500 U/m² versus the conventional 2,500 U/m². Patients presenting features that were associated with increased risk of CNS relapse received two extra doses of intrathecal therapy during the first 2 weeks of remission induction.

RESULTS

The 5-year event-free survival and overall survival rates for the 598 patients were 88.2% (95% CI, 84.9% to 91.5%) and 94.1% (95% CI, 91.7% to 96.5%), respectively. Cumulative risk of any—isolated or combined—CNS relapse was 1.5% (95% CI, 0.5% to 2.5%). Higher doses of PEG-asparaginase did not affect treatment outcome. T-cell phenotype was the only independent risk factor for any CNS relapse (hazard ratio, 5.15; 95% CI, 1.3 to 20.6; P = . 021). Among 359 patients with features that were associated with increased risk for CNS relapse, the 5-year rate of any CNS relapse was significantly lower than that among 248 patients with the same features treated in the previous Total Therapy Study 15 (1.8% [95% CI, 0.4% to 3.3%] v 5.7% [95% CI, 2.8% to 8.6%]; P = .008). There were no significant differences in the cumulative risk of seizure or infection during induction between patients who did or did not receive the two extra doses of intrathecal treatment.

CONCLUSION

Higher doses of PEG-asparaginase failed to improve outcome, but additional intrathecal therapy during early induction seemed to contribute to improved CNS control without excessive toxicity for high-risk patients.

INTRODUCTION

Contemporary treatment of pediatric acute lymphoblastic leukemia (ALL) yields 5-year event-free survival rates of 80% or higher.¹ In addition to further advancing cure rates, current trials increasingly focus on improving the patients’ quality of life and preventing long-term sequelae. A major recent accomplishment has been the elimination of prophylactic cranial irradiation without jeopardizing leukemia control in the CNS. The St Jude Total Therapy Study 15 resulted in a 5-year event-free survival rate of 86%, with an isolated CNS relapse rate of 2.7% and a combined CNS relapse rate of 1.2%.² Corresponding rates for the Dutch Childhood Oncology Group Protocol ALL-9 were 81%, 2.6%, and 2.0%, respectively.³ Both studies, reported in 2009, featured intensive dexamethasone, vincristine, asparaginase, and triple intrathecal therapy—with methotrexate, hydrocortisone, and cytarabine—without prophylactic cranial irradiation, previously considered the standard treatment of patients with high-risk ALL.^2,3 Since then, three other major studies omitting prophylactic cranial irradiation have been reported, with 5-year rates of event-free survival rates ranging from 72% to 85%, isolated CNS relapse rates from 1.7% to 4.1%, and combined CNS relapse rates from approximately 1% to 1.8%.^4-6 These results are comparable to those of six other studies during the same period^7-12 that used prophylactic cranial irradiation in 0.6% to 42.8% of patients.¹³ These findings notwithstanding, 3.5% to 5.4% of patients not treated with prophylactic cranial irradiation required subsequent therapeutic irradiation because of CNS relapse.^2-6

In the Total Therapy Study 16, we sought to improve event-free survival and CNS control by refining risk-directed therapy and intensifying systemic and intrathecal chemotherapy.¹⁴ It has been proposed that the degree of asparagine depletion from the CSF by asparaginase is important for the treatment and prevention of CNS leukemia. Several studies have demonstrated that PEG-asparaginase 2,500 U/m² usually achieves complete depletion of asparagine from the blood but not necessarily from the CSF.^15-17 We therefore reasoned that a higher dose of PEG-asparaginase may be more effective in depleting asparagine from the blood and CSF and might improve systemic and CNS control. We also tested whether intensification of triple intrathecal chemotherapy during early remission induction in patients with presenting features that were associated with increased CNS relapse could improve CNS control. This strategy was prompted by our experience in Total Therapy Study 15,² where triple intrathecal treatment reduced the frequency of CNS relapse but did not seem to have reached maximum intensification, and was supported by its success in patients with advanced Burkitt lymphoma or Burkitt leukemia,¹⁸ lymphoid malignancies that are associated with a high risk of CNS relapse. Here, we present the outcome of these two therapeutic interventions and the overall results of this study.

PATIENTS AND METHODS

Participants

Between October 29, 2007, and March 26, 2017, 598 eligible patients younger than age 19 years with newly diagnosed ALL were enrolled in the Total Therapy Study 16 (Appendix, online only). The protocol was approved by the institutional review board, and written informed consent was obtained from the parents, guardians, or patients, with assent from the patients, as appropriate.

Diagnosis and Risk Classification

Diagnosis of ALL was based on immunophenotypic and genetic features of leukemic cells. Patients were classified as having low-risk, standard-risk (intermediate-risk), or high-risk leukemia on the basis of presenting characteristics and treatment response as determined by levels of minimal residual disease (MRD) measured by flow cytometry during remission induction and consolidation (Appendix and Fig 1). CNS status was defined as CNS-1, CNS-2, CNS-3, or traumatic lumbar puncture with blast cells (Appendix and Table 1).

FIG 1.

CONSORT diagram. MRD, minimal residual disease.

TABLE 1.

Treatment Outcome According to Selected Clinical and Biologic Characteristics*

Study Design and Outcomes

For the primary objective, 420 evaluable patients were planned to be stratified and randomly assigned (1:1) on the first day of continuation treatment to receive PEG-asparaginase at the conventional dose (2,500 U/m²) or at a higher dose (3,500 U/m²; Appendix). The primary end point was continuous complete remission. A secondary objective was to evaluate whether early intensification of triple intrathecal chemotherapy during early remission induction reduced the rate of isolated or any—isolated or combined—CNS relapse in patients at high risk for this complication. Because of the low rate of CNS relapse, a randomized study of any treatment intervention to improve CNS control was not feasible; therefore, data from patients with similar characteristics enrolled in the preceding Total Therapy Study 15 were used for comparison.² The study was originally powered for the primary objective only, but we performed a post hoc power calculation for the CNS secondary objective (Appendix).

Treatment

Treatment on Total Therapy Study 16 was based on that used in Study 15 with several modifications (Appendix). Remission induction consisted of prednisone, vincristine, daunorubicin, and PEG-asparaginase, followed by cyclophosphamide, cytarabine, and a thiopurine (Appendix Table A1, online only). Patients with Philadelphia chromosome–positive ALL received dasatinib 40 mg/m² twice daily from day 22 until the end of all treatment. Upon hematopoietic recovery—between days 43 and 46—consolidation therapy consisting of four courses of high-dose methotrexate, mercaptopurine, and triple intrathecal therapy was administered (Appendix Table A2, online only). All patients received antimetabolite-based continuation therapy for 120 weeks. During initial continuation therapy, low-risk patients received two reinduction cycles plus pulses of dexamethasone and vincristine, whereas standard-risk or high-risk patients received PEG-asparaginase every 2 weeks for 15 doses interrupted with pulses of doxorubicin plus vincristine plus dexamethasone and two reinduction cycles, followed by three rotating drug pairs (Appendix Table A3, online only).

Triple intrathecal chemotherapy was instilled at an age-appropriate dose with the number of doses varied according to presenting characteristics and CNS status that historically affected the risk of CNS relapse (Appendix Table A1). All patients received intrathecal therapy on days 1 and 15 of induction with additional doses on days 8 and 22 in patients with Philadelphia chromosome, KMT2A rearrangement, hypodiploidy < 44 chromosomes, or leukocyte count > 100,000/µL at presentation, and on days 4, 8, 11, and 22 for those with features associated with an increased risk of CNS relapse in Total Therapy Study 15, including T-cell ALL, the presence of TCF3-PBX1, CNS-2 or CNS-3 status, or traumatic lumbar puncture with blasts.² During continuation therapy, intrathecal therapy was also administered according to risk features (Appendix Table A3). Hence, patients with low-risk ALL received a total of 13 to 21 intrathecal treatments, and those with standard-risk ALL received 16 to 27 treatments.

Allogeneic hematopoietic stem-cell transplantation was an option for patients with high-risk leukemia. These patients received 1-2 courses of reintensification therapy before transplantation (Appendix Table A4, online only).

Statistical Analysis

We used the Kaplan-Meier method to estimate the probabilities of continuous complete remission, event-free survival, or overall survival with SE calculated according to Peto and Pike.¹⁹ Independent prognostic factors were identified using the Cox proportional hazards regression model. The cumulative incidence functions of isolated CNS relapse and any CNS relapse were estimated according to Kalbfleisch and Prentice²⁰ and compared with those in the Total Therapy Study 15 using Gray’s test²¹; death in remission and second neoplasm were regarded as competing events. All reported P values were two sided and not adjusted for multiple comparisons. Outcome data updated on May 1, 2019, were used in all analyses; 98% of survivors had been seen within 1 year. Median follow-up time for the 561 patients who were alive at the time of analysis was 6.26 years (interquartile range, 4.41 years; range, 1.15 to 11.43 years). All statistical analyses were based on intent-to-treat population and performed with SAS (version 9.4; SAS Institute, Cary, NC) and R version 3.3.0 (www.r-project.org).

RESULTS

Treatment Outcome

Of the 598 patients, including 12 infants, 590 (98.66%) entered complete morphologic remission (low risk, 99.62%; standard risk, 99.64%; high risk, 89.66%). There were 8 induction failures, 4 as a result of fatal infections and the other 4 refractory leukemia; 2 of these 4 patients were alive and in remission 2.2 and 4.5 years later, the latter after undergoing transplantation.

Allogeneic transplantation was performed in 24 patients at 1.58 to 30.35 months (median, 4.46 months) after remission induction (Appendix). Sixteen of the 24 patients were alive and in remission. There were 36 relapses including 26 hematologic, 7 isolated CNS, 1 isolated testicular, 1 combined CNS and ocular, and 1 combined hematologic and testicular. Six patients developed secondary malignancies (Appendix). There were 17 deaths in remission: 3 transplantation-related toxicities, 7 infections, 2 hepatic failures, 1 saddle pulmonary embolus, 2 respiratory failures, 1 systemic inflammatory response, and 1 cerebral hemorrhagic infarct.

The probability of 5-year event-free survival was 88.2% (95% CI, 84.9% to 91.5%), overall survival 94.1% (95% CI, 91.7% to 96.5%; Fig 2), and cumulative risk of any relapse 6.6% (95% CI, 4.4% to 8.7%), including 4.6% (95% CI, 2.8% to 6.5%) for an isolated hematologic relapse. The death-in-remission rate was 2.7% (95% CI, 1.4% to 4.0%).

FIG 2.

Kaplan-Meier and Kalbfleisch and Prentice analyses of outcomes in children with acute lymphoblastic leukemia. The 5-year and 10-year results are shown on the curves.

Impact of Presenting Features and Early Treatment Response

Table 1 lists the 5-year event-free survival and overall survival rates according to presenting features, risk assessment, and response to initial treatment. Because infants had a significantly poorer outcome than older patients, and patients with T-cell ALL had inferior outcomes compared with those with B-cell ALL, we analyzed prognostic factors separately in cases of T-cell ALL and B-cell ALL in noninfants (Table 2). Among patients with B-cell ALL, the only factors associated with inferior event-free survival and overall survival rates were MRD at the end of remission induction, black race, and KMT2A rearrangement, whereas ETV6-RUNX1 predicted a better outcome. Among patients with T-cell ALL, the only significant adverse predictor was the presence of any blasts in the CSF at diagnosis. Of note, the 15 patients with Philadelphia chromosome–positive ALL had an excellent 5-year event-free survival rate of 71.1% (95% CI, 42.9% to 99.3%), with 3 patients experiencing hematologic relapses, including the only patient who underwent transplantation for high MRD and subsequently died of relapse, and 1 who died of infection. The 5-year event-free survival rate for the 10 patients with early T-cell precursor ALL, one of whom died of infection during induction and another of relapse after transplantation, was 80.0% (95% CI, 53.5% to 100%).

TABLE 2.

Independent Risk Factors for Any Major Adverse Events (n = 66) or Deaths (n = 37) Among All Patients and Among Noninfant Patients With B-ALL or T-ALL

Treatment Toxicity

The cumulative risk of death from toxic effects during remission induction and continuation chemotherapy was 3.2% (95% CI, 1.8% to 4.6%). Table 3 summarizes the most notable toxic effects of treatment. Apart from allergic reactions to PEG-asparaginase, all toxic effects occurred more often among standard-risk and high-risk patients than among low-risk patients. Age greater than 10 years was associated with an increased risk of severe infection (11.5% [95% CI, 6.0% to 17.0%] v 5.9% [95% CI, 3.6% to 8.1%]; P = .027), osteonecrosis (31% [95% CI, 22.6% to 39.4%] v 2.2% [95% CI, 0.8% to 3.6%]; P < .001), hyperglycemia (27.3% [95% CI, 19.8% to 34.7%] v 7.2% [95% CI, 4.7% to 9.6%]; P < .001), and thrombosis (29.9% [95% CI, 22.1% to 37.7%] v 10.6% [95% CI, 7.7% to 13.6%]; P < .001) compared with findings in younger patients. Whereas none of the 12 infants experienced a reaction to PEG-asparaginase, hyperglycemia, or pancreatitis, they had a higher cumulative risk of grade 4 or 5 infections (58.3% [95% CI, 28.4% to 88.3%] v 7.2% [95% CI, 5.0% to 9.3%]; P < .001), seizures (16.7% [95% CI, 0.0% to 38.7%] v 6.0% [95% CI, 4.1% to 8.0%]; P = .08), and thrombosis (37.5% [95% CI, 5.9% to 69.1%] v 15.2% [95% CI, 12.2% to 18.2%]; P = .038) than did older children.

TABLE 3.

Selected Major Toxic Effects of Treatment

PEG-Asparaginase Randomization and CNS Relapse

Of the 598 evaluable patients, 414 patients were randomly assigned for the primary study objective to receive PEG-asparaginase 2,500 U/m² (n = 208) or 3,500 U/m² (n = 206). A total of 184 patients were not randomly assigned for various reasons (Fig 1). Although there were no significant differences in the presenting features between the 2 randomly assigned groups, the non-randomly assigned patients were more likely to have high-risk disease and less likely to have ETV6-RUNX1 ALL and CNS-3 status (Appendix Table A5, online only). There were no differences in the 5-year continuous complete remission rate (90.4% [95% CI, 85.7% to 95.1%] v 91.2% [95% CI, 86.7% to 95.7%]; P = .91), cumulative risk of isolated CNS relapse (1.1% [95% CI, 0.0% to 2.7%] v 0.5 [95% CI, 0.0% to 1.4%]; P = .98), and cumulative risk of any CNS relapse (1.1% [95% CI, 0.0% to 2.7%] v 1.0% [95% CI, 0.0% to 2.3%]; P = .55) between patients who were randomly assigned to receive 2,500 U/m² or 3,500 U/m². There was no difference in treatment outcome between any subsets of patients within the 2 randomly assigned groups (Appendix Table A6, online only).

Seven of 598 patients experienced isolated CNS relapse and 1 a combined CNS and ocular relapse. Six of these patients remained in second remission for 0.57 to 4.25 years, but 2 died of progressive disease (Appendix Table A7, online only). Five-year cumulative risk of isolated CNS relapse was 1.3% (95% CI, 0.3% to 2.2%) and of any CNS relapse 1.5% (95% CI, 0.5% to 2.5%; Fig 2). Whereas presenting leukocyte count > 100,000/µL and T-cell ALL were associated with an increased risk of isolated or any CNS relapse (Table 4), only T-cell ALL retained independent significance in the multivariable analysis for any CNS relapse (hazard ratio, 5.15 [95% CI, 1.3 to 20.6]; P = . 021) and for isolated CNS relapse (hazard ratio, 6.83 [95% CI, 1.5 to 30.5]; P = .012).

TABLE 4.

Cumulative Risk of Isolated or Any CNS Relapse According to Selected Clinical and Biologic Characteristics

There were 359 patients in the current study with presenting features that were associated with an increased risk of CNS relapse defined in Total Therapy Study 15,² and they had significantly lower CNS relapse rates than the corresponding 248 patients enrolled in Study 15 (5-year cumulative risk of any CNS relapse, 1.8% [95% CI, 0.4% to 3.3%] v 5.7% [95% CI, 2.8% to 8.6%]; P = .008; isolated CNS relapse, 1.5% [95% CI, 0.2% to 2.9%] v 4.0% [95% CI, 1.6% to 6.5%]; P = .049; Fig 3). There were no significant differences in the cumulative risk of grade 2 to 4 seizure (5.9% [95% CI, 3.2% to 8.7%] v 6.5% [95% CI, 3.7% to 9.4%]; P = .78) or grade 4 or 5 infections (8.1% [95% CI, 4.9% to 11.3%] v 8.4% [95% CI, 5.2% to 11.7%]; P = .95) between the 293 patients who received two to four intrathecal treatments and the 305 who received six intrathecal treatments during remission induction.

FIG 3.

Cumulative risk of (A) any CNS relapse and (B) isolated CNS relapse in the Total Therapy 15 and 16 Studies.

DISCUSSION

The results of this study suggest that early intensification of intrathecal therapy improves CNS control. Compared with historical controls treated in Total Therapy Study 15, the two extra doses of intrathecal therapy administered during the first 2 weeks of remission induction seemed to significantly reduce CNS relapse without excessive toxicity in patients with features that were previously associated with an increased risk of CNS relapse, such as black race, the presence of TCF3-PBX1, or any CNS involvement, including a CNS-2, CNS-3, or traumatic lumbar puncture with blasts status at diagnosis.² Despite the omission of prophylactic cranial irradiation in each of these subgroups, rates of isolated CNS or any CNS relapse were only 1.5% and 1.8%, respectively, significantly lower than the 4.0% and 5.7%, respectively, observed among historical controls treated in the Total Therapy Study 15.² The overall isolated CNS relapse rate of 1.3% and any CNS relapse rate of 1.5% in this study also compared favorably with those reported rates from studies conducted during the same time period, many of which still used prophylactic cranial irradiation for high-risk patients.^3-12 In this regard, the strategy of early intensification of intrathecal therapy for patients with CNS-2 or CNS-3 status has also be used in other clinical trials with improved results.^9,12

Our study enrolled consecutive patients without exclusion of any high-risk cases, such as infant or Philadelphia chromosome–positive ALL. Remarkably, none of the patients in some of the higher-risk categories experienced CNS relapse, including 34 patients with a traumatic lumbar puncture with blasts, 15 with Philadelphia chromosome, 16 with TCF3-PBX1, 12 with KMT2A-AFF1, and 12 with infant ALL. Only one of 21 patients with CNS-3 status at diagnosis experienced an isolated CNS relapse, but massive cerebral involvement did not preclude cure without cranial irradiation (Appendix Fig A1, online only). In contrast, these factors were still associated with increased CNS relapse in other contemporary clinical trials, even with the use of prophylactic cranial irradiation.^22-24 For example, in the Children’s Oncology Group (COG) AALLL0331 and AALL0232 trials, CNS-2, CNS-3, and traumatic lumbar puncture status were associated with increased CNS relapse and poorer event-free survival rates.²² In the 2 most recent clinical trials for Philadelphia chromosome–positive ALL, which included treatment with dasatinib, isolated or combined CNS relapse occurred in 7% of patients in one study²³ and 15.4% in the other.²⁴ Finally, in the Interfant-99 and Interfant-06 trials in which cranial irradiation was omitted, isolated CNS relapse occurred in 3% and 4.5%, and combined CNS relapse in 5% and 4.9% of infants, respectively.^25,26

It is likely that other treatment components of our trial also contributed to CNS control, not only for those with particularly high-risk features identified in Study 15, but also for all other patients. For example, high-dose methotrexate has been shown to improve CNS outcome in standard-risk or high-risk patients.^22,27 Dasatinib can cross the blood-brain barrier,²⁸ and our use of a dosage that was higher than that of other studies^23,24 (80 mg/m² per day v 60 mg/m² per day) may have resulted in higher drug levels in the CNS, preventing CNS relapse in our patients with Philadelphia chromosome–positive ALL. Dexamethasone and asparaginase treatment were used in all patients and intensified in those with standard- or high-risk ALL in this study. Both drugs have been shown to improve CNS control, and asparaginase potentiates the effect of dexamethasone.²⁹

There is no consensus regarding the optimal dose or number of doses of PEG-asparaginase that should be administered in patients with any ALL subtype. Two recent studies tested the effect of extra doses of PEG-asparaginase after remission induction.^30,31 In the COG AALL0331 study for standard-risk ALL, after remission induction with dexamethasone, vincristine, and 1 dose of PEG-asparaginase 2,500 U/m², patients were randomly assigned to receive consolidation treatment with or without 4 additional doses.³⁰ In the Nordic Society of Pediatric Hematology and Oncology (NOPHO) ALL2008 study for non–high-risk ALL, after remission induction and consolidation treatment, including 5 doses of PEG-asparaginase 1,000 U/m², patients who tolerated asparaginase were randomly assigned to receive an additional 10 doses administered every 2 weeks versus 3 doses every 6 weeks during early continuation treatment.³¹ Extra doses of PEG-asparaginase failed to improve outcomes for standard-risk or non–high-risk patients in both studies.^30,31 The optimal number of doses for high-risk ALL was uncertain. In our study, we found that increasing PEG-asparaginase to 3,500 U/m² had no excessive toxicity but did not improve systemic or CNS control for any subgroup of patients, which suggests that a threshold level of asparaginase exposure may have been reached at 2,500 U/m².

Our results corroborate the outcome of a 10-study meta-analysis that indicated that cranial radiotherapy has no important effect on the overall risk of relapse or survival in childhood ALL.³² In the trial reported here, an isolated CNS relapse occurred in 7 patients, 5 of whom remained alive in second remission for 2.3 to 6.2 years (median, 4.04 years). Of the 498 patients treated in our Total Therapy Study 15, isolated CNS relapse occurred in 11 patients, all of whom were alive in second remission for 5.8 to 16.0 years (median, 13.9 years) from diagnosis (Pui et al, unpublished data). Thus, among 916 patients with B-cell ALL and 180 with T-cell ALL treated in our 2 consecutive trials omitting prophylactic cranial irradiation, only 1 each with B-cell ALL and T-cell ALL died because of CNS relapse. With the recent advent of chimeric antigen receptor–modified T-cell therapy, it is possible that even patients who experience CNS relapse can be cured without the use of therapeutic cranial irradiation.^33,34 Indeed, the patient with a combined CNS and ocular relapse in this study was treated only with remission induction treatment, including intrathecal therapy followed by CD19 chimeric antigen receptor–modified T-cell therapy, and has remained in second remission for 25 months. The only risk factor for CNS relapse in this study was T-cell ALL with a cumulative risk of 4.3%. The recent improved overall outcome of T-cell ALL in the COG AALL0434 study suggests that nelarabine may be used to improve CNS control of this leukemia subtype.³⁵

Despite the reduction of CNS relapse and a corresponding increase in event-free survival, the overall survival rate reported here was similar to that of the Total Therapy Study 15.² This finding can be explained, in part, by the inclusion of infants in the current study, whose 5-year event-free survival was 50.0% (95% CI, 19.0% to 81.0%) and overall survival 58.3% (95% CI, 25.4% to 91.2%) compared with 89.0% (95% CI, 85.9% to 92.1%) and 94.9% (95% CI, 92.5% to 97.3%) for children older than 1 year. However, a more compelling explanation is that the intensity of conventional chemotherapy had reached its limit. Indeed, the cumulative risk of toxic death reported here was higher than in our earlier study (3.2% v 1.4%). Thus, additional improvement in outcome will need to rely more heavily on molecular therapeutic and cellular immunotherapy approaches.³⁶

Appendix

Patients and Methods

Participants.

Of the 600 consecutive patients younger than age 19 years with newly diagnosed acute lymphoblastic leukemia (ALL) enrolled in the Total Therapy Study 16 at St Jude Children's Research Hospital, 2 patients registered in the study were subsequently excluded because of revised diagnoses of acute bilineal leukemia and chronic myeloid leukemia in blast crisis.

Diagnosis and risk classification.

Diagnosis of ALL was based on immunophenotypic and genetic features of leukemic cells. Patients with B-cell ALL between age 1 and 10 years, with a leukocyte count of < 50,000/µL, a DNA index (the ratio of DNA content in leukemic cells to that in normal diploid G0/G1 cells) of ≥ 1.16, or the translocation t(12;21) (ETV6-RUNX1) were provisionally classified as having low-risk ALL. Infants with KMT2A rearrangement, patients with t(9;22) (BCR-ABL1), and those with early T-cell precursor ALL were considered to have high-risk ALL. The remaining patients were provisionally classified as having standard-risk (intermediate-risk) ALL.

Final risk classification was determined by the level of minimal residual disease (MRD) measured during remission and consolidation induction by flow cytometry. Provisional low-risk patients with an MRD level of ≥ 1% in bone marrow aspirate on day 15 of remission induction or 0.01% to 0.99% after the completion of 6 weeks of induction therapy were considered to have standard-risk ALL. An MRD level of ≥ 1% after the completion of induction therapy or persistent, re-emergent, or increased MRD during consolidation therapy indicated high-risk ALL, regardless of provisional risk classification.

CNS status was classified as follows: CNS-1 status (no detectable blast cells in a sample of cerebrospinal fluid), CNS-2 status (< 5 leukocytes/µL of CSF with blasts), CNS-3 status (≥ 5 leukocytes/µL of CSF with blasts or cranial palsy), and traumatic lumbar puncture with blast cells (≥ 10 erythrocytes/µL of CSF with blast cells).

Study design and outcomes.

For the primary objective, patients were randomly assigned (1:1) on the first day of continuation treatment to receive PEG-asparaginase at the conventional dose (2,500 U/m²) or at a higher dose (3,500 U/m²). Randomization was stratified by risk group (low v standard/high risk), immunophenotype (T v B), and MRD level on day 15 of induction (< 0.01% v ≥ 0.01%) among patients with T-cell ALL. Patients with Down syndrome were not randomly assigned and were treated with the conventional dosage. The number of patients with presenting features that were associated with a high risk of CNS relapse—black race, any CNS involvement, T-cell ALL, and TCF3-PBX1—was equally distributed between the 2 randomized arms (149 in the conventional dose and 151 in the higher dose arm). Neither investigators, nor patients were masked to treatment assignment. The accrual goal for the primary objective was 420 evaluable patients, which would provide 80% statistical power (α = .05; 2-sided log-rank test) to detect an improvement of 8.5% in the 5-year continuous complete remission rate. Two interim analyses were planned, and the final analysis was performed 3 years after the last patient was randomly assigned.

The primary objective of the randomization of dosage of PEG-asparaginase had 2 end points. The primary end point was continuous complete remission assessed from the randomization date to the first adverse event, including relapse at any site, death during remission, or second malignant neoplasm. The secondary end point was overall survival time from the randomization date to death from any cause.

A secondary objective was to evaluate whether early intensification of triple intrathecal chemotherapy during early remission induction would reduce the rate of isolated or any (isolated or combined) CNS relapse in patients who were at high risk for this complication compared with the cohort in the preceding Total Therapy Study 15.² Because it was difficult to determine a priori the number of patients (out of the total enrollment for the primary objective) who would satisfy the high-risk criteria for CNS relapse to receive early intensification of triple intrathecal chemotherapy during early remission induction, power analysis was not performed for this objective. Post hoc power analysis on the basis of realized sample size (n1 = 359 [Study 16] and n2 = 248 [Study 15]) and the observed 5-year any-CNS relapse rates (0.018 [Study 16] and 0.057 [Study 15]) showed that the log-rank test has the post hoc power 0.869 at the .05 significant level. The post hoc power for isolated CNS relapse was 0.682. Despite this relatively low power, the comparison nonetheless showed borderline significance (P = .049).

Event-free survival time was calculated from the date of diagnosis to any first adverse event, including death during remission induction and induction failure as a result of drug resistance. Overall survival was measured as the time from diagnosis to death from any cause. Patients without an event were censored at the last day they were known to be alive.

Treatment.

Total Therapy Study 16 was similar to that used in Study 15 with several modifications (Tables A1, A2, A3, and A4). Intramuscular native Escherichia coli asparaginase was replaced by intravenous PEG-asparaginase, which was intensified in standard- and high-risk patients. Mercaptopurine during days 22 to 35 of remission induction was replaced with thioguanine for patients with wild-type thiopurine methyltransferase activity. Cyclophosphamide was intensified to 4 fractionated doses (300 mg/m² every 12 hours for four doses) instead of 1 g/m² for 1 dose for standard- or high-risk patients with poor early treatment response—that is, 5% or more leukemia cells on day 15 of remission induction. Two extra doses of triple intrathecal therapy on days 4 and 11 of remission induction were administered to patients with increased risk of CNS relapse. Dasatinib instead of imatinib was administered for patients with Philadelphia chromosome–positive ALL. To improve quality of life, dexamethasone dose was decreased to 6 mg/m² per day for all patients after week 69 of continuation treatment, and the duration of continuation treatment was shortened from 146 weeks to 120 weeks for boys.

Remission induction and consolidation.

Remission induction began with prednisone, vincristine, daunorubicin, and PEG-asparaginase (Table A1). Patients with an MRD level of ≥ 1% after 2 weeks of induction were administered an additional dose of PEG-asparaginase on day 15. Subsequent induction therapy between days 22 and 35 consisted of cyclophosphamide, thioguanine (or mercaptopurine in thiopurine methyltransferase intermediate or poor metabolizers), and cytarabine for all patients, except infants with KMT2A fusion, who were treated with a 5-day course of clofarabine, cyclophosphamide, and etoposide. For patients with Philadelphia chromosome–positive ALL, dasatinib (40 mg/m² twice daily) was started on day 22 and continued until the end of all treatment. Upon hematopoietic recovery—between days 43 and 46—MRD was assessed and consolidation therapy consisting of 4 courses of high-dose methotrexate, mercaptopurine, and triple intrathecal therapy was administered (Table A2).

Continuation therapy.

Continuation therapy for low-risk patients consisted of daily mercaptopurine and weekly methotrexate interrupted with pulses of mercaptopurine, dexamethasone, and vincristine, in addition to 2 reinduction cycles with dexamethasone, vincristine, and PEG-asparaginase with 1 dose of doxorubicin added to the first reinduction (Table A3). Patients with standard-risk or high-risk disease received PEG-asparaginase every 2 weeks for a total of 15 doses and daily mercaptopurine with pulses of doxorubicin plus vincristine plus dexamethasone (Table A3). Patients also received 2 reinduction cycles that consisted of dexamethasone, vincristine, and PEG-asparaginase, together with doxorubicin in the first reinduction and high-dose cytarabine in the second reinduction. They then received 3 rotating drug pairs (mercaptopurine plus methotrexate, cyclophosphamide plus cytarabine, and dexamethasone plus vincristine). Doses of mercaptopurine and methotrexate were adjusted according to the tolerance, the phenotype and genotype of thiopurine methyltransferase, and erythrocyte thioguanine metabolites. Total scheduled doses of anthracyclines were limited to 80 mg/m² and 230 mg/m², and cyclophosphamide to 1 g/m² and 4.8 g/m² in low-risk disease and standard-risk patients, respectively. Continuation treatment lasted 120 weeks for all patients.

CNS-directed therapy.

Triple intrathecal chemotherapy was instilled at an age-appropriate dose immediately after the collection of CSF with a diagnostic lumbar puncture and was repeated on day 15 (Table A1). Additional intrathecal therapy was administered on days 8 and 22 of remission induction in patients with Philadelphia chromosome, KMT2A rearrangement, hypodiploidy less than 44 chromosomes, or leukocyte count > 100,000/µL at presentation, and on days 4, 8, 11, and 22 for patients with features associated with an increased risk of CNS relapse in Total Therapy Study 15, including T-cell ALL, presence of TCF3-PBX1, CNS-3 status, CNS-2 status, or traumatic lumbar puncture with blasts. Leucovorin rescue (5 mg/m² per dose, max 5 mg) was administered orally 24 and 30 hours after each intrathecal treatment during remission induction. Intrathecal therapy was administered every 2 weeks during consolidation therapy and according to risk features thereafter—every 4 or 8 weeks in low-risk patients and every 4 weeks in standard-risk or high-risk patients during the first year of continuation therapy (Table A3). Patients who were at high risk of CNS relapse continued to receive intrathecal therapy until week 97 of continuation therapy. Patients with low-risk ALL received a total of 13 to 21 intrathecal treatments, and those with standard-risk ALL, 16 to 27 treatments.

Allogeneic hematopoietic stem-cell transplantation.

Transplantation was an option for patients with high-risk leukemia whose treatment before transplantation was identical to that of patients with standard-risk disease. To maximize the reduction of MRD before transplantation, we administered reintensification therapy to all patients and 1 or 2 additional courses of clofarabine, etoposide, cyclophosphamide, and dexamethasone to those who experienced suboptimal responses to the initial reintensification (Table A4).

Results

Treatment outcome.

The 598 evaluable patients included 12 infants, 9 of whom had KMT2A rearrangements. Median age at diagnosis was 6.04 years (range, 0.12 to 18.89 years), and median leukocyte count was 13,900/µL (range, 200 to 905,300/µL). On the basis of MRD measurements (available for 596 patients), we reclassified the risk status of 94 patients: 70 patients from provisional low risk to standard risk and 24 from provisional standard risk to high risk (Fig 1).

Of 598 patients, 590 (98.66%) entered complete morphologic remission (low risk, 99.62%; standard risk, 99.64%; high risk, 89.66%). There were 8 induction failures—4 because of fatal infections and the other 4 refractory leukemia; 2 of these 4 patients were alive and in remission 2.2 and 4.5 years later (the latter after transplantation). Twenty-four patients, including 3 with induction failure, underwent allogeneic stem-cell transplantation (4 from matched-sibling donors, 15 from matched-unrelated donors, 4 from haploidentical donors, and 1 cord blood) at 1.58 to 30.35 months (median, 4.46 months) after remission induction. Transplantation was performed in these patients because of an MRD level of ≥ 1% at the end of induction therapy (n = 10), persistent MRD at week 16 after remission induction (n = 2), increasing MRD (n = 6), and early T-cell precursor ALL (n = 6). Among the latter, 3 patients had an MRD level < 0.01%, one 0.01%, and one 3.82% at the end of induction, as well as 1 patient with increasing MRD to 0.077% during consolidation. Sixteen (66.7%) of 24 patients who underwent transplantation, including 1 nonresponder, were alive and in remission. Two nonresponders died of progressive disease and cardiopulmonary failure, 3 died of complications, 2 patients with relapsed disease died of progressive disease, and 1 patient who developed a second malignancy died as a result of respiratory failure.

There were 36 cases of relapses, including 26 hematologic, 7 isolated CNS, 1 isolated testicular, 1 combined CNS and ocular, and 1 combined hematologic and testicular. Six patients developed secondary malignancies: T-lymphoblastic lymphoma after the completion of therapy for B-cell ALL, renal-cell carcinoma and squamous-cell carcinoma after transplantation, low-grade cerebellar glioma, myelodysplastic syndrome with germline ETV6 variant, glioblastoma multiforme with constitutional mismatch repair deficiency syndrome, and acute myeloid leukemia with monosomy 7. There were 17 deaths in remission: 3 transplantation-related toxicities, 7 infections, 2 hepatic failures, 1 saddle pulmonary embolus, 2 respiratory failures, 1 systemic inflammatory response, and 1 cerebral hemorrhagic infarct.

FIG A1.

T1 postcontrast sagittal and coronal images of an 8-month-old girl with KMT2A-rearranged ALL presenting with massive cerebral involvement, scalp lesions and leukocyte count of 31,100/µL. Cytospin on CSF showed 0 RBC/µL and 93 WBC/µL with 99% blasts. Images show an extensive parieto-occipital dural-based tumor involving superior sagittal sinus, bilateral brain parenchyma, and erosion through the skull. Superior sagittal sinus was displaced but not occluded on venogram. There was a smaller scalp lesion (arrow) involving the dura, but not the brain. Lesions resolved completely after 3 months of chemotherapy. She has remained in continuous first remission for 11.3 years.

TABLE A1.

Remission Induction Therapy

TABLE A2.

Consolidation Therapy: Total Therapy Study 16 and Study 15

TABLE A3.

Continuation/Reinduction Therapy

TABLE A4.

Reintensification Therapy

TABLE A5.

Comparisons of Presenting Features Between Non-Randomly Assigned and Randomly Assigned Patients and Between the Two Randomly Assigned Groups

TABLE A6.

Comparisons of Continuous Complete Remission Duration of Patients With Selected Clinical and Biologic Features Between the 2 Randomly Assigned Groups

TABLE A7.

Clinical and Biologic Features of the 7 Patients With Isolated and 1 With Combined CNS Relapse

AUTHOR CONTRIBUTIONS

Conception and design: Sima Jeha, Jeffrey E. Rubnitz, Raul C. Ribeiro, Jun J. Yang, William E. Evans, Ching-Hon Pui

Financial support: Ching-Hon Pui

Administrative support: William E. Evans, Ching-Hon Pui

Provision of study material or patients: Sima Jeha, John Choi, John T. Sandlund, Hiroto Inaba, Jeffrey E. Rubnitz, Raul C. Ribeiro, Ching-Hon Pui

Collection and assembly of data: Sima Jeha, John Choi, Cheng Cheng, Elaine Coustan-Smith, Dario Campana, Hiroto Inaba, Tanja A. Gruber, Susana C. Raimondi, Jun J. Yang, Charles G. Mullighan, William E. Evans, Ching-Hon Pui

Data analysis and interpretation: Sima Jeha, Deqing Pei, John Choi, Cheng Cheng, John T. Sandlund, Elaine Coustan-Smith, Dario Campana, Jeffrey E. Rubnitz, Raul C. Ribeiro, Raja B. Khan, Jun J. Yang, Charles G. Mullighan, James R. Downing, William E. Evans, Mary V. Relling, Ching-Hon Pui

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

Improved CNS Control of Childhood Acute Lymphoblastic Leukemia Without Cranial Irradiation: St Jude Total Therapy Study 16

The following represents disclosure information provided by authors of this manuscript. All relationships are considered compensated unless otherwise noted. 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.

Open Payments is a public database containing information reported by companies about payments made to US-licensed physicians (Open Payments).

Elaine Coustan-Smith

Stock and Other Ownership Interests: Unum Therapeutics (I), Nkarta Therapeutics (I), Medisix Therapeutics (I)

Honoraria: GlaxoSmithKline (I)

Consulting or Advisory Role: Medisix Therapeutics (I), Unum Therapeutics (I), Nkarta Therapeutics (I)

Patents, Royalties, Other Intellectual Property: Juno Therapeutics (I), Celgene (I), Unum Therapeutics (I), Nkarta Therapeutics (I), Medisix Therapeutics (I), Patent and patent application

Travel, Accommodations, Expenses: GlaxoSmithKline (I), Celgene (I), Nkarta Therapeutics (I)

Dario Campana

Stock and Other Ownership Interests: Nkarta Therapeutics, Medisix Therapeutics, Unum Therapeutics

Honoraria: GlaxoSmithKline

Consulting or Advisory Role: Unum Therapeutics, Nkarta Therapeutics, Medisix Therapeutics

Patents, Royalties, Other Intellectual Property: Royalties from Juno Therapeutics (a Celgene company) for license of chimeric antigen receptors; patents and patent applications in the area of cell therapy of cancer not licensed or licensed to Unum Therapeutics, Nkarta Therapeutics, or Medisix Therapeutics; patent applications in the area of leukemia immunophenotyping and minimal residual disease monitoring

Travel, Accommodations, Expenses: GlaxoSmithKline, Nkarta Therapeutics

Hiroto Inaba

Research Funding: Servier

Tanja A. Gruber

Stock and Other Ownership Interests: Bristol-Myers Squibb

Charles G. Mullighan

Honoraria: Incyte, Amgen, Pfizer, Illumina

Consulting or Advisory Role: Incyte, Illumina, Pfizer

Speakers' Bureau: Amgen, Pfizer, Loxo, AbbVie, Pfizer

Patents, Royalties, Other Intellectual Property: Inventor on a pending patent application related to gene expression signatures for detection of underlying Philadelphia chromosome–like events and therapeutic targeting in leukemia (PCT/US2012/069228)

Travel, Accommodations, Expenses: Incyte, Amgen, Pfizer, Illumina

Mary V. Relling

Research Funding: Servier

Ching-Hon Pui

Consulting or Advisory Role: Adaptive Biotechnologies

No other potential conflicts of interest were reported.