Phase II Study of Dose-Adjusted EPOCH as Initial Therapy for Adults with High-Risk Acute Lymphoblastic Leukemia

Abstract

Treatments for adults with newly-diagnosed acute lymphoblastic leukemia (ALL) may be prohibitively toxic and/or resource-intense. To address this, we performed a phase II study of dose-adjusted etoposide, prednisone, vincristine, cyclophosphamide, and doxorubicin (DA-EPOCH). Imatinib or dasatinib was added for Ph+ disease; rituximab was added when CD20+. Fifty-three patients were evaluable: 28 with Ph+ disease, and 25 with Ph−. All patients had ≥1 high-risk clinical feature. Measurable residual disease-negativity by multiparameter flow cytometry within 4 cycles was achieved in 71% in patients with Ph+ ALL and 64% in Ph− ALL. Median overall survival (OS) was 49 months, with 2-year OS of 71%. Median relapse-free survival (RFS) in the 47 patients that attained morphologic remission was 24 months, with 2-year RFS of 57%. Early mortality was 2%. In summary, DA-EPOCH yields deep and durable remissions in adults with ALL comparable to some resource-intense strategies but with a low rate of treatment-related death.

Introduction

Pediatric-inspired approaches are considered a standard of care in young adults with Philadelphia chromosome negative (Ph−) acute lymphoblastic leukemia (ALL).¹ Unfortunately, toxicities and complexities limit their applicability outside of specialized centers. New agents, including ABL tyrosine kinase inhibitors (TKIs) for Ph+ disease and targeting of cell-surface antigens in B-cell ALL, have yielded new strategies, including some described as “chemotherapy-free.”^2-7 Prospective comparisons to more traditional approaches have yet to be performed, and follow-up in some cases remains short. As a result, standard therapy still relies on cytotoxic chemotherapy, particularly for Ph− disease in older adults and in resource-limited settings, where use of some novel agents is infeasible due to cost and/or complexity.

HyperCVAD (hyper-fractionated cyclophosphamide, vincristine, doxorubicin, and dexamethasone alternating with high-dose methotrexate and cytarabine) is among the most commonly used regimens in adults with ALL.⁸ It has been studied in combination with various TKIs and the anti-CD20 antibody rituximab,^9-11 making it a relatively versatile approach. However, because of its toxicity, incorporating additional drugs can be difficult. This has led to reduced dosing of effective novel agents¹⁰ with hyperCVAD or utilization of the lower-intensity backbone mini-hyperCVD (omitting anthracycline, dose-reduction of other agents).^5,12,13 Even so, non-relapse morbidity and mortality remains high with many of these regimens. Moreover, low-intensity therapy may yield poorer outcomes compared to standard hyperCVAD.^14,15

Dose-adjusted (DA)-EPOCH (etoposide, prednisone, vincristine, cyclophosphamide, and doxorubicin) is an effective regimen for high-grade lymphomas with a manageable safety profile,^16,17 but it has not been explored in ALL. Compared to other ALL regimens, DA-EPOCH generally is simpler to administer and less toxic, even in the elderly.¹⁸ Hence, DA-EPOCH is a regimen meriting further exploration in adults with ALL.

Historically, response to induction chemotherapy is a critical prognostic factor for adults with ALL. Though approximately 90% of patients achieve a morphologic complete remission (mCR), 5-year overall survival (OS) is still only about 50% for these patients.¹⁹ Today, measurable residual disease (MRD) is both an established tool for response assessment in routine care of ALL and a useful endpoint in the investigation of novel therapies.²⁰ Therefore, we performed a phase II trial to assess the activity of DA-EPOCH as initial therapy in adults with both Ph− and (with TKI) Ph+ ALL, as determined by the rate and time to MRD negativity (MRD−).

Methods

Study Participants

Patients aged 18 years or older with ALL or lymphoblastic lymphoma with detectable lymphoblasts in the blood or marrow were eligible, provided they were not appropriate for pediatric-inspired therapy. Our institutional approach to this decision, along with additional eligibility criteria, are summarized in the Supplementary Information. The study was performed in accordance with the Declaration of Helsinki and was registered in Clinicaltrials.gov (NCT03023046). All participants gave written informed consent using forms approved by the Fred Hutchinson/University of Washington Cancer Consortium Institutional Review Board (Cancer Consortium protocol 9770).

Procedures

ALL was diagnosed per routine clinical practice from blood and/or marrow testing. BCR::ABL1 RT-PCR was performed in Ph+ cases to identify the transcript isoform (i.e., p190 vs p210). Within the Ph− subgroup, cytogenetic risk was categorized according to NCCN criteria.²¹ IKZF1^plus signature was identified by chromosomal microarray as previously described.^21,22 When available, pre-treatment samples were archived for high-throughput sequencing (HTS)-based MRD detection.

DA-EPOCH was administered as previously described.²³ Rituximab was added if CD20 was present on leukemic blasts defined either quantitatively (20% or greater) or qualitatively (“present,” “high,” “intermediate,” or “low”; not “absent”). Central nervous system (CNS) prophylaxis with intrathecal chemotherapy was given in all patients via lumbar puncture; the precise schedule was left to the discretion of the treating physician. That said, methotrexate 12 mg intrathecally was recommended on days 1 and 5 of each cycle for either 8 or (in patients with an elevated WBC or LDH three times the upper limit of normal at diagnosis) 10 doses.¹⁶ Additional treatment details are provided in the Supplementary Material.

Patients were taken off study if they did not achieve mCR within 2 cycles, if treatment was delayed more than 3 weeks due to toxicity, for completion of protocol treatment (maximum 8 cycles), or if a treatment-related non-hematologic grade 4 adverse event (AE) occurred. Post-study interventions (including hematopoietic cell transplantation [HCT] or maintenance therapy) were offered at the discretion of the treating physician after at least 2 cycles to ensure adequate assessment of response and toxicity. Generally, at our center, HCT is recommended in first remission for patients with Ph+, KMT2A rearrangement, early thymic precursor subtype of T-cell ALL, and those who do not achieve MRD− by MFC within 3 months of starting chemotherapy; the same approach was used for patients on this study. Further, patients with Ph+ ALL who undergo HCT are generally offered maintenance TKI post-transplant, the details of which vary depending on individual patient circumstances. Absent transplantation, maintenance therapy was recommended: prednisone, vincristine, methotrexate, and mercaptopurine (POMP) for up to 2 years for patients with Ph− disease; and vincristine, prednisone, and daily TKI for 2 years for Ph+ disease similar to the MD Anderson experience after hyperCVAD, though TKI could be continued indefinitely.^24,25

Outcomes Measurement

Response was determined primarily with bone marrow aspirate (BMA) after cycle 1. BMA was repeated after cycle 2 if mCR was not attained after cycle 1, then at least after cycle 4 and at the end of treatment. Otherwise, the timing of BMA was left to the discretion of the treating physician. Bone marrow response was determined by morphology, multiparameter flow cytometry (MFC), and (for Ph+) BCR::ABL1 RT-PCR. MCR was defined as <5% blasts by morphology. MRD− was defined by <0.01% abnormal cells by MFC performed exclusively at the University of Washington Hematopathology Laboratory;²⁶ in cases where an abnormal population was detectable but below this threshold, an independent hematopathologist reviewed the data and assigned either MRD positive (MRD+) or ‘indeterminate’ MRD, the latter of which was considered MRD−. When MRD− by MFC (MFC−), HTS-based MRD testing was performed by Adaptive Biotechnologies (clonoSEQ).²⁷ For BCR::ABL1 RT-PCR, complete molecular response (CMR) was assigned in the absence of detectable p190 or p210.²⁸ For patients with EMD, an interim response assessment by cross-sectional imaging was performed once between the start of cycle 2 and 5 and again at the end of treatment. Results were interpreted by the treating investigator per NCCN criteria.²⁹

Adverse events were characterized according to the Common Terminology Criteria for Adverse Events (version 5.0),²⁷ with grade 3 or higher events (except asymptomatic grade 3 laboratory abnormalities) used to describe the toxicity profile. Survival was measured relative to the start of study treatment. Relapse was applied to patients who achieved mCR and then had either the re-emergence of morphologically-detectable disease (i.e., > 5% blasts) in the blood or marrow and/or a new site of proven extramedullary disease. Events were defined as any of the following: [1] unable to achieve MRD− by MFC, [2] relapse, [3] MRD recurrence by MFC after achieving MRD−, or [4] death from any cause.

Statistical Analysis

The primary objective of this phase II study was to examine the activity of DA-EPOCH as frontline therapy for adults with ALL. This was assessed using the rate of complete MRD response (i.e., MRD−) by MFC within four cycles of treatment in two independent subgroups: Ph+ disease and Ph−. Four cycles translates to approximately 12 weeks of treatment, similar to a “post-consolidation” timepoint that has been shown by other groups to have prognostic significance.^30,31 Our historical comparisons were our previously-published results with hyperCVAD with or without rituximab (when CD20+) or TKI (when Ph+).³² Here, we reported rates of MRD− by MFC within 90 days of treatment initiation (i.e., about 4 cycles) of 50% for Ph+ ALL and 59% for Ph− ALL. Additional details are found in the Supplementary Material, but we summarize the key points of this design below.

As the eligibility criteria enriched for patients with Ph+ ALL, enrollment into this subgroup was expected to be higher. We thus implemented a Simon two-stage minimax design to estimate the response rate more precisely from these patients. Consequently, DA-EPOCH + TKI would be considered potentially efficacious if at least 18 of 28 patients (64% or higher estimated response rate) achieved MRD− by MFC within 4 cycles. Alternatively, for the Ph− subgroup, we calculated a 90% confidence interval of the MRD− rate within 4 cycles in cohorts of pre-defined size. If the upper bound of this confidence interval was not below our historical results with hyperCVAD, we continued to enroll up to a maximum of 25 patients.

Reporting of secondary objectives was primarily descriptive. Binary outcomes were estimated with proportions and associated confidence intervals (CI), with comparisons made by Fisher’s exact test. Time-to-event outcomes were generated from Kaplan-Meier estimates and compared via log-rank tests. The data generated in this study are not publicly available due to institutional regulatory policies but are available upon reasonable request from the corresponding author.

Results

Patient and Treatment Characteristics

A total of 63 patients were screened, 54 were enrolled, and 53 were evaluable: 28 with Ph+ disease, and 25 with Ph− disease, reaching the maximum planned enrollment for each arm (Figure 1). Logistical and financial barriers were the primary reasons for screening failure. Enrollment occurred between March 2017 and February 2021.

Figure 1: Patient flow diagrams for patients who were screened, enrolled, and evaluable after receiving dose-adjusted EPOCH for newly-diagnosed ALL.

Patients with Philadelphia chromosome positive (Ph+) ALL are shown in panel A, and those with Ph− ALL are depicted in panel B. Numbers in parentheses following medians reflect the range of cycles given (i.e., minimum-maximum). Other abbreviations: CR, complete remission; F/U, follow-up; MAC, myeloablative conditioning; MRD, measurable residual disease; RIC, reduced-intensity conditioning; TKI, ABL tyrosine kinase inhibitor. Footnotes: * One patient with Ph+ ALL who achieved MRD− with dose-adjusted EPOCH + TKI received one cycle of hyperCVAD part B + TKI while waiting to undergo HCT. ^ Eight patients received vincristine + corticosteroid + TKI and one patient received corticosteroid + TKI.

Characteristics of the 53 evaluable patients (and the focus of the remainder of this report) are summarized in Table 1. Rituximab was given to 18 patients (34%), 6 of whom had Ph+ ALL. For patients with Ph+ ALL, dasatinib was the primary TKI used for 23 patients (82%) and imatinib for 5 (18%), for whom the choice was dictated by insurance authorization. For those with Ph− ALL, none of the patients had good-risk cytogenetics. Those with Ph− disease and poor-risk cytogenetics were classified as such due to the following abnormalities: low-hypodiploid/near-triploid (n = 8), t(v;14q32)/IGH (n = 3), complex (n = 2), KMT2A rearranged (n = 2), and alterations of IKZF1 (n = 1). Therefore, all patients in this study would be classified as having high-risk disease based on either age (91%), WBC at diagnosis (21%), or cytogenetics (83%).

Table 1:

Baseline characteristics of evaluable subjects treated with dose-adjusted EPOCH, segregated by Philadelphia chromosome status.

Characteristic	Ph+ (n = 28)		Ph− (n = 25)
	Number/ Median	%/ Range	Number/ Median	%/ Range
Age at Diagnosis
Median (Range), years	58	19-78	54	40-79
> 35 years	23	82%	25	100%
> 60 years	14	50%	7	28%
ECOG Performance Status
0-1	18	64%	16	64%
2-3	10	36%	9	36%
Sex
Female	12	43%	16	64%
Male	16	57%	9	36%
Race
Asian or Pacific Islander	0	0%	2	8%
Black or African American	1	4%	1	4%
Native American or Indigenous Peoples	0	0%	3	12%
White	27	96%	19	76%
Ethnicity
Hispanic or Latinx	2	7%	4	16%
Not Hispanic or Latinx	26	93%	21	84%
B Lineage	26	93%	23	92%
CD20+	6	23%	12	52%
Bone Marrow Involvementa
Median, Range (%)	90	0.2-98	86	24-98
>50%	13	87%	13	81%
High WBCb	8	29%	3	12%
Cytogenetics/FISH
Ph+	28	100%	N/A	N/A
p190	22	79%	N/A	N/A
p210	6	21%	N/A	N/A
Ph−	N/A	N/A	25	100%
Good Riskc	N/A	N/A	0	0%
Poor Riskc	N/A	N/A	16	64%
No Abnormalities/Unable to Assess	N/A	N/A	4	16%
Other	N/A	N/A	5	20%
IKZF1 plus	12	43%	1	4%
Measurable extramedullary disease	5	18%	3	12%

^aBone marrow examination was performed at screening in 31 patients (58%): 15 with Ph+ ALL and 16 with Ph− ALL. The data reported in this table represent only these patients. The remaining patients had sufficient circulating disease for testing to be performed on peripheral blood, in accordance with NCCN Guidelines (v4.2021).

^bDefined as > 30,000/μL for B-lineage and > 100,000/μL for T-lineage.

^cDefined according to NCCN Guidelines (v4.2021), with poor risk including the following [with the exception of the Philadelphia chromosome or t(9;22)(q34;q11.2)]: complex, low-hypodiploid/near-triploid, KMT2A rearranged, t(v;14q32)/IGH, intrachromosomal amplification of chromosome 21 (iAMP21), t(17;19), and alterations of IKZF1.

Abbreviations: FISH, fluorescence in situ hybridization; N/A, not applicable; Ph, Philadelphia chromosome; WBC, white blood cell count

Response Rates

The primary endpoint of MRD− by MFC within 4 cycles was achieved in 71% (95% CI, 51%-87%) in patients with Ph+ ALL and 64% (95% CI, 43%-82%) in Ph− ALL (Table 2). Fifteen patients (28%) were able to reach this depth of response after only one cycle (i.e., approximately 3 weeks after treatment initiation). Among these patients who achieved MRD− after 1 cycle, 6 relapses (40%) occurred vs 10 relapses (48%) in the 21 patients who reached MRD− by MFC later (p = 0.74). Regarding the execution of our Simon two-stage design for the Ph+ cohort, 11 of the first 15 evaluable patients (73%) reached MRD− by MFC within 4 cycles. This exceeded the first-stage threshold necessary to continue the study (i.e., 7 of 15 [47%]). At the conclusion of enrollment, 20 of 28 patients (71%) had responded to this depth, which also was greater than the pre-defined level for potential efficacy (i.e., 18 of 28 [64%]). For the Ph− arm, none of the response rates observed in the pre-defined patient cohorts was below the stopping-rule threshold, so enrollment proceeded to the maximum of 25 patients (data not shown). Further details of MRD response rates by BCR::ABL1 RT-PCR and HTS are shown in the Supplementary Material.

Table 2:

Response rates following DA-EPOCH using different methods of assessment.

		Ph+ (N=28)		Ph− (N=25)		Total
		Number	%	Number	%	Number	%
Morphologic Remission		27	96%	20	80%	47	89%
MRD−	Cycle 1	7	25%	8	32%	15	28%
	Cycle ≤ 4	20	71%	16	64%	36	68%
CMR	Cycle 1	1	4%	N/A	N/A	N/A	N/A
	Cycle ≤ 4	11	39%	N/A	N/A	N/A	N/A
HTS−a	Cycle 1	2	8%	2	9%	4	8%
	Cycle ≤ 4	9	38%	6	27%	13	28%

^aPercentages were calculated after excluding patients that were unable to have HTS testing performed for technical reasons (Ph+, n = 3; Ph−, n = 3).

Abbreviations: CMR, complete molecular response by BCR::ABL1 RT-PCR; HTS-, no detectable disease by high-throughput sequencing; MRD−, no detectable disease by multiparameter flow cytometry; Ph, Philadelphia chromosome;

Twenty-seven Ph+ cases (96%) achieved mCR, while 20 (80%) with Ph− achieved mCR. After one cycle, one patient with Ph+ ALL and three with Ph− ALL achieved mCR with incomplete count recovery, but all later had count recovery despite continuing treatment. Of 7 Ph− patients that did not achieve mCR after 1 cycle, 4 (57%) were removed from the study due to this inadequate response. Of the remaining three (43%) who proceeded to a second cycle, one patient did not achieve mCR and thus was removed from the study; the other two patients reached CR but remained MRD+. Of these five patients with Ph− ALL who never achieved mCR with DA-EPOCH, three had high-risk cytogenetics (two with hypodiploidy and one with IKZF1 alteration); one had a high WBC; and their ages ranged from 45-64. Alternatively, the one patient with Ph+ ALL to not achieve mCR opted to leave the study after one cycle due to toxicity; imatinib was the TKI used in this case.

In the eight patients with EMD, post-treatment response was assessed in six patients. In these patients, complete responses were seen in five (83%), and partial response was noted in one (17%), for an overall response rate at EMD sites of 100%. Of the two patients not evaluated by imaging post-treatment, one patient had EMD manifested only as splenomegaly but achieved MRD− by BMA, while the other came off study following one cycle of DA-EPOCH after not experiencing a response by BMA.

Post-Protocol Therapy

Details of treatments received following DA-EPOCH are shown in Figure 1. Among the 20 patients with Ph+ ALL who achieved MRD− with DA-EPOCH + TKI, 10 (50%) underwent HCT in first remission, 4 of whom (40%) later relapsed. Alternatively, nine patients (45%) were given TKI-based maintenance therapy, and four (44%) have since relapsed. The reasons for not referring these nine patients for HCT in first remission included the following: patient declined (n = 4), inadequate social support (n = 2), mutual decision between physician and patient (n = 2), and lack of available donor (n = 1). The one patient who declined further treatment later relapsed and died from ALL. As for the eight patients with persistent disease after DA-EPOCH + TKI, six transitioned to blinatumomab for treatment of MRD, and all (100%) remain alive and in remission at last contact.

Regarding the patients with Ph− ALL, 13 of 16 (81%) who achieved MRD− received POMP maintenance following DA-EPOCH, with 6 (46%) later relapsing. The remaining three patients (19%) who reached MRD− underwent HCT, one of whom (33%) relapsed: two of these patients had KMT2A-rearranged ALL, and the other had “secondary ALL” with a history of myeloma. The five patients who did not reach mCR were all given part B of hyperCVAD (high-dose methotrexate and cytarabine), and only one (20%) achieved mCR with this regimen; four (80%) of these patients have since died. Of the three remaining patients (i.e., those with persistent MRD after DA-EPOCH), two received blinatumomab, both reached MRD−, and neither relapsed (0%); however, one patient died of complications after subsequent HCT.

Overall, Event-Free, and Relapse-Free Survival

Kaplan-Meier curves depicting OS, EFS, and relapse-free survival (RFS) are shown in Figure 2. Among surviving patients, median follow-up was 26 months (range, 8.0-56 months). For all evaluable patients, median OS was 49 months, with the 2-year OS estimate of 71% (95% CI, 58%-86%). Median and 2-year EFS estimate for all evaluable patients were 15 months and 34% (95% CI, 23%-52%), respectively. RFS was examined in the 47 patients that reached mCR. Collectively, the median RFS was 24 months, and the 2-year RFS estimate was 57% (95% CI, 43%-75%). Twelve of the nineteen patients who relapsed had Ph+ disease: nine received dasatinib (39% given this drug), and three received imatinib (60% given this drug). Further, among these 19 relapses, 3 patients (6%) suffered isolated CNS relapse, with another 5 (9%) experiencing concurrent medullary and CNS relapse. All three patients who suffered an isolated CNS relapse had Ph+ ALL: two received dasatinib (9% given this drug), while one received imatinib (20% given this drug). No patients experienced a non-CNS isolated extramedullary relapse. Of those achieving MRD−, 12 patients (23%; 2 Ph− and 10 Ph+) underwent HCT in first remission (Figure 1). Relapse was less frequent (38%) in patients who reached HTS− compared to 62% of HTS+ patients, but this was not statistically significant (p = 0.43).

Figure 2: Kaplan-Meier curves of adults who received dose-adjusted EPOCH for newly-diagnosed ALL.

OS is shown in panel A, EFS in panel B, and RFS in panel C, segregated by Ph status. Associated medians and 2-year point estimates (with 95% CI) for each plot are included within each panel. Tick marks represent censored patients, determined by the date of last contact having not experienced an event. RFS was examined only in the 47 patients that obtained a complete remission with dose-adjusted EPOCH.

Impact of Adding Rituximab and Specific TKI to DA-EPOCH

While not randomly assigned, we explored how rituximab and the choice of imatinib vs dasatinib impacted rates of MRD−. Thirteen of 18 patients (72%; 5 of 6 with Ph+ ALL and 8 of 12 with Ph−) who received rituximab achieved MRD− by MFC within 4 cycles, as compared to 23 of 35 patients (65%; 15 of 22 with Ph+ ALL and 8 of 13 with Ph−) who did not receive rituximab (p = 0.76). As for the impact of TKI used, 18 of 23 patients (78%) given dasatinib reached MRD− by MFC within 4 cycles as opposed to 2 of 5 (40%) prescribed imatinib (p = 0.12). When compared by rates of CMR within 4 cycles, this was achieved in 9 patients (39%) on dasatinib vs 2 patients (40%) given imatinib (p = 1.0).

Identification of Risk Factors for Events or Mortality

To identify pre-treatment factors predictive of reaching MRD− within 4 cycles, EFS, and OS, univariate Cox models for selected characteristics were generated (Supplementary Table 3). None of these identified a factor that was statistically-significantly associated with these outcomes. Due to the relatively small number of events, multivariate analyses were not performed.

Adverse Events & Toxicity

Grade 3 or higher adverse events (excluding asymptomatic grade 3 laboratory abnormalities) attributable to DA-EPOCH from all courses given are enumerated in Table 3. These 72 events affected 38 patients, with 15 patients (28%) not experiencing any such events throughout their participation. Taken further, all grade 3 adverse events (including asymptomatic lab abnormalities and those not attributed to study therapy) are reported in Supplementary Table 4. Life-threatening and fatal toxicity was rare. The rate of treatment-related mortality was 2%. There were two grade 4 adverse events associated with septic shock that were attributed to DA-EPOCH, both of which resolved.

Table 3:

Non-hematologic Grade 3-5 adverse events attributed to dose-adjusted EPOCH.

CTCAE Category	Grade 3	Grade 4	Grade 5
Febrile Neutropenia	21
Sepsis	10	1
Mucositis Oral	8
Miscellaneous Infectiona	6
Upper Gastrointestinal Hemorrhage	2		1
Hypotension	3
Atrial fibrillation	2
Hypertension	2
Intracranial hemorrhage	2
Multiorgan Failure		1
Abdominal pain	1
Aspiration	1
Delirium	1
Diarrhea	1
Edema Limbs	1
Fibrinogen decreased	1
Hypoxia	1
Lower gastrointestinal hemorrhage	1
Myocardial Infarction	1
Oropharyngeal pain	1
Peripheral Motor Neuropathy	1
Sinus Bradycardia	1
Small Intestinal Obstruction	1

^aThis category includes the following events: enterocolitis (n = 2), endophthalmitis (n = 1), fungemia (n = 1), kidney infection (n = 1), and typhlitis (n = 1).

Note: blank cells denote zero instances of those events at those grades.

Individual hematologic toxicity events were not captured by study design. However, the degree of neutropenia and thrombocytopenia was used in the dose-adjustment paradigm for EPOCH (Supplementary Material). After cycle 1 (i.e., after sufficient cytoreduction that ALL was no longer the cause), only 15 patients (34%) had a grade 4 platelet count at any time, a level that would necessitate a dose level decrease. On the contrary, 28 patients (53%) were able to escalate to at least Dose Level 2 at some point, with 3 patients (6%) able to reach Dose Level 4.

Outcomes Among Patients Over Age 60

To highlight a particularly vulnerable population, 21 (40%) of the patients were over 60 years old at enrollment. As shown in Supplementary Table 5, these patients had proportionately more cases of Ph+ disease compared to their younger counterparts on this study (67% vs 44%, respectively; p = 0.16), and the proportion of Ph− ALL with poor-risk cytogenetics was lower (29% vs 78%, respectively; p = 0.06). Otherwise, baseline disease characteristics were relatively similar. Among these patients aged 60 and over, 19 (90%) achieved mCR, with 5 (24%) achieving MRD− after 1 cycle and 16 (76%) being MRD− within 4 cycles. The median OS was not reached, and the 2-year OS estimate was 73% (95% CI, 55%-97%). Median EFS was 17 months, and the 2-year EFS estimate 46% (95% CI, 28%-74%). Additional data from this subgroup are shown in the Supplementary Material.

Discussion

Here, we present the first known experience testing DA-EPOCH as initial therapy for adults with ALL. While the use of this regimen for lymphoma may be waning following recent data,³³ it remains a standard approach for high-grade B-cell lymphomas such as Burkitt lymphoma,^16,34 a disease with clinical features that are analogous to ALL. The continuous-infusion nature of drug administration with this regimen is perhaps what leads to the activity seen in these high-grade malignancies.³⁵ There were three primary takeaways from our study: DA-EPOCH induced deep and durable remissions, it was associated with few severe adverse events, and older adults fared comparably to younger patients.

There are numerous options now considered standard for adults with ALL.²¹ While many rely on complex schedules of multiagent chemotherapy, some newer options leverage the activity of antibody-based agents and (in the case of Ph+ disease) TKIs to reduce or even eliminate chemotherapy. Unfortunately, they may not be feasible outside of specialized centers comfortable with their unique toxicities, and drug costs are substantially greater. A study like ours cannot establish superiority over any of these strategies. That said, our results can potentially place DA-EPOCH among these options, particularly for patients in selected circumstances: advanced age, significant comorbidities, resource-limited settings, and/or treatment at centers less comfortable with complexities of traditional ALL regimens. Another appealing feature is the potential for outpatient administration with home-infusion support, something we were able to provide to selected patients after cycle 1 (data not shown).

Because MRD is a standard part of managing ALL, we used this as the primary endpoint in our study, an increasingly-utilized strategy in this disease.²⁰ By MFC, our results exceeded our predefined threshold for success in Ph+ ALL compared to historical results with hyperCVAD + TKI.³² In Ph− ALL, we observed similar rates of MRD− as hyperCVAD. These results also stand up to those reported with hyperCVAD from MD Anderson Cancer Center, both with the incorporation of rituximab and (when combined with mini-hyperCVD) inotuzumab ozogamicin.^9,13 For Ph+ ALL, however, the rates of MRD− with DA-EPOCH + TKI are slightly lower than those reported in some of the original studies of hyperCVAD + TKI.^10,11 That said, we saw similar rates of CMR as was seen after dasatinib and two cycles of blinatumomab in D-ALBA.⁶ The more recent combination of ponatinib and blinatumomab yielded higher rates of CMR,⁷ though this regimen is only feasible in highly-resourced centers. By focusing on MRD, we also could make several novel observations about its use. HTS is increasingly used for its higher degree of sensitivity.³⁶ However, ours is among the first trials reported to incorporate it in a prospective manner, demonstrating feasibility of an approach using this test specifically when it is more informative (i.e., MRD− by MFC).

DA-EPOCH arguably performed best in patients over age 60. Historically, older adults are less likely to achieve mCR and experience higher rates of induction complications (including mortality) and relapse.^14,15,37 Importantly, these trends toward worse outcomes in patients over age 60 were not observed with DA-EPOCH, underscoring its relative tolerability and effectiveness in older adults. While we lack the necessary follow-up to compare rates of long-term OS, DA-EPOCH yielded reassuring rates of mCR, MRD−, and early death relative to the oft-cited GMALL regimen for older adults.³⁸ Longer follow-up of older adults given mini-hyperCVD plus inotuzumab ozogamicin (with or without blinatumomab) has shown this resource-intense approach may only benefit patients under 70 whose disease lacks poor-risk cytogenetics, in part due to increased risk of death in remission.³⁹ While only a subgroup of our study, we did not observe these issues. Further, blinatumomab followed by POMP noted a 3-year OS of only 37%,⁴⁰ numerically lower than what we saw with DA-EPOCH.

There are several important limitations from our findings. First, the sample sizes are relatively small, limiting our conclusions. Second, though we treated high-risk patients on this trial, EFS and the rate of mCR specifically in Ph− ALL may appear relatively unsatisfactory. Importantly, events used to define EFS included persistent MRD. With the availability of blinatumomab in this scenario, the decision to transition to this therapy would contribute to this endpoint. However, as shown in Figure 1, this strategy frequently yielded durable remissions. While DA-EPOCH failed to achieve MRD−, it facilitated selection of patients more likely to benefit from alternative therapy, with sufficient cytoreduction to optimize the use of blinatumomab. As for the five patients with Ph− ALL who did not achieve mCR, all of them transitioned to part B of hyperCVAD (i.e., high-dose methotrexate and cytarabine), and only one was able to reach MRD+ mCR (Figure 1). This suggests that DA-EPOCH was not an ineffective regimen, but these patients had chemorefractory disease. Next, while we observed only three isolated CNS relapses, a legitimate criticism of DA-EPOCH is that it lacks adequate CNS activity. Notably, longer-term follow-up after dasatinib and blinatumomab from D-ALBA noted a similar rate of CNS relapse (i.e., 4 of 63).⁴¹ Future studies with this approach should monitor this. Lastly, since only 8% of patients had T-cell ALL, our results may not accurately reflect outcomes for this group.

In conclusion, DA-EPOCH yields frequent, deep, and durable remissions when used as frontline therapy for both Ph+ and Ph− ALL in adults with high-risk ALL. Our results support the use of this regimen as a standard approach, particularly if more complex or intense regimens are too toxic or infeasible. Future studies may identify characteristics of patients most likely to respond favorably to this lower-intensity approach, perhaps in the context of a multi-center trial with longer follow-up and better representation of patients with T-lineage disease. Such efforts could also attempt to optimize the approach to CNS prophylaxis, better describe how this regimen performs relative to more established treatments, and whether novel agents can be added to this backbone to improve outcomes. A study combining inotuzumab ozogamicin with DA-EPOCH is already underway,⁴² and other agents like venetoclax or asparaginase formulations would also be rational options to explore.