Molecular risk markers define risk of relapse in myeloid leukemia of Down syndrome beyond measurable residual disease

Key Points

• •Intensification of chemotherapy for patients with ML-DS and positive EOI-1 MRD neither improved event-free survival nor overall survival, but resulted in more febrile neutropenia and sepsis events.
• •Proposed molecular markers of relapse risk in ML-DS (mutation/deletion of CDKN2A, mutation of ZBTB7A, JAK2, or TP53) were prognostic in patients with ML-DS treated on both arms of Children’s Oncology Group AAML1531, suggesting molecular risk markers can be used to risk stratify therapy in ML-DS.

Visual Abstract

Abstract

Myeloid leukemia of Down syndrome (ML-DS) is a distinct form of pediatric acute myeloid leukemia (AML) that responds to reduced-intensity chemotherapy, as compared with non-DS AML that requires intensive chemotherapy and often stem cell transplant. While most patients with ML-DS have a favorable prognosis, outcomes for those with refractory or relapsed disease are dismal. Children’s Oncology Group study AAML1531 introduced the use of measurable residual disease by multiparameter flow cytometry at the end of the first course of induction therapy (EOI-1 MRD) for risk stratification of treatment intensity. Of 280 patients with ML-DS who were enrolled, 41 were classified as high risk (HR) due to positive EOI-1 MRD, and treated with intensified chemotherapy similar to that used for pediatric non-DS AML. Treatment intensification did not improve the 2-year event-free survival compared with patients who were MRD-positive treated with reduced-intensity therapy in the predecessor study AAML0431 (80.5% ± 12.4% vs 76%; P = .247) or overall survival (80.5% ± 12.4% vs 76.2% ± 18.6%; P = .819), but significantly increased the frequency of febrile neutropenia and sepsis events. While stratification of treatment intensity based on MRD was not beneficial, molecular markers of relapse risk proposed by the Japan Children’s Cancer Group for ML-DS (alterations of CDKN2A, ZBTB7A, JAK2, TP53) proved prognostic. Relapse risk was 50% in patients who were HR from AAML1531 with any high-risk molecular marker compared with 6.7% in those without. Similar relapse results were obtained in the MRD-negative AAML1531 group, suggesting molecular risk markers can predict outcome and thus be used to stratify therapy in ML-DS. This trial was registered at www.clinicaltrials.gov as #NCT02521493.

Introduction

Myeloid leukemia of Down syndrome (ML-DS) is a distinct form of pediatric acute myeloid leukemia (AML) that typically occurs in children aged ≤4 years, and is characterized by a predominantly megakaryoblastic phenotype, high prevalence of antecedent cytopenias, and somatic GATA1 mutations.¹ ML-DS is prenatally initiated by somatic GATA1 mutations as transient abnormal myelopoiesis, driven by distinct co-operating events,2, 3, 4 and highly sensitive to chemotherapy.^5,6 Over the last decade, ML-DS has been treated with DS-specific reduced-intensity chemotherapy regimens. While the majority of children with ML-DS have excellent outcomes (5-year event-free survival [EFS] ∼90%),7, 8, 9 a subset of 10% to 15% of patients do not survive, primarily due to relapsed or refractory disease (2-year overall survival [OS] 22%).8, 9, 10, 11, 12 Children’s Oncology Group (COG) study AAML1531 aimed to risk stratify treatment intensity in ML-DS for the first time using measurable residual disease by multiparameter flow cytometry after the first induction course. On the COG predecessor study AAML0431, patients who were positive for minimal residual disease (MRD) by multidimensional flow cytometry after their first course of therapy had a poorer outcome compared with those who were MRD negative.⁸ Therefore, on AAML1531, high-dose cytarabine was eliminated from treatment for those patients with ML-DS who were MRD-negative (<0.05%) at the end of Induction I (EOI-1; termed standard risk [SR]), while chemotherapy was intensified to a level currently used for non-DS AML for those patients with ML-DS who were found to be MRD-positive (>0.05%) at EOI-1 (termed high risk [HR]).

The results for the patients who were SR treated on AAML1531 were previously reported,¹³ and demonstrated that omission of high-dose cytarabine from the treatment of the SR group (based on negative EOI-1 MRD) resulted in decreased EFS, leading to the recommendation that high-dose cytarabine should be included in the treatment of this patient population. Moreover, we concluded that MRD was not a useful tool in identifying a cohort of patients with ML-DS for whom treatment intensity could be further reduced.

Here we report the outcomes of patients with ML-DS in the AAML1531 HR group, specifically the results of treatment intensification of conventional chemotherapy based on positive EOI-1 MRD. We also retrospectively determined the prevalence of mutations and deletions in the four genes (CDKN2A, JAK2, TP53, and ZBTB7A) recently identified as molecular high-risk markers in ML-DS,¹⁴ and compared their prognostic impact with that of MRD in the MRD-based HR and SR groups on AAML1531.

Patients and methods

Trial design

AAML1531 was a trial solely enrolling children with Down syndrome (constitutional or mosaic) and newly diagnosed AML aged between 90 days and 4 years. The study opened on 23 November 2015, and was closed on 1 October 2018, following an interim analysis that demonstrated lack of efficacy of SR treatment.¹³ The SR arm remained permanently closed to accrual, but AAML1531 reopened on 19 February 2019 to continue to enroll patients on the HR arm. The study closed to accrual on 15 April 2022. All patients received Induction I chemotherapy (thioguanine 50 mg/m² per dose, twice a day, days 1-4; cytarabine 200 mg/m² per 24 hours continuous infusion, days 1-4; daunorubicin 20 mg/m² on days 1-4 over 1-15 minutes) and a single dose of age-based intrathecal cytarabine (Figure 1). Gemtuzumab ozogamicin was not used in this or any cycle. For patients aged <36 months, dosing was based on weight. After count recovery, residual disease in the bone marrow was measured by multidimensional flow cytometry. Those with EOI-1 MRD <0.05% were classified as SR and were treated with reduced-intensity chemotherapy based on COG AAML0431, but with elimination of the second induction course of AAML0431 therapy, which consisted of high-dose cytarabine/asparaginase, in an effort to reduce infectious events. Patients with EOI-1 MRD ≥0.05% were classified as HR and had their subsequent treatment intensified to a level consistent with that of pediatric non-DS AML. Dosing in the HR arm was weight-based for all patients with the goal of toxicity reduction: Induction II (cytarabine 33 mg/kg per dose every 12 hours, days 1-4; mitoxantrone 0.4 mg/kg per day, days 3-6); Intensification I (cytarabine 33 mg/kg per dose every 12 hours, days 1-5; etoposide 5 mg/kg per day, days 1-5); Intensification II (cytarabine 100 mg/kg per dose every 12 hours, days 1, 2, 8, and 9; Escherichia coli asparaginase 200 IU/kg per dose or Erwinia chrysanthemi asparaginase 830 IU/kg per dose, days 2 and 9). Patients were not required to meet minimal organ function requirements prior to enrollment. The trial was approved by the central institutional review board of the National Cancer Institute, and institutional review boards of each enrolling center. Patients and their families provided informed consent or assent as appropriate. The trial was conducted in accordance with the Declaration of Helsinki (NCT02521493).

Figure 1.

Study design for AAML1531. SR and HR groups, and treatment intensity are defined by MRD at EOI-1. TAD, thioguanine 50 mg/m² per dose, twice a day, days 1-4; cytarabine 200 mg/m² per 24 hours continuous infusion, days 1-4; daunorubicin 20 mg/m² on days 1-4 over 1-15 minutes.

Cytogenetics

Analyses were performed by COG-approved local and reference cytogenetic laboratories in accordance with the American College of Medical Genetics and Genomics (ACMG) and College of American Pathologists (CAP) guidelines (https://www.acmg.net/ACMG/Medical-Genetics-Practice-Resources/Genetics_Lab_Standards/ACMG/Medical-Genetics-Practice-Resources/Genetics_Lab_Standards.aspx?hkey=0e473683-3910-420c-9efb-958707c59589 and https://www.cap.org/protocols-and-guidelines/cap-guidelines/current-cap-guidelines/initial-diagnostic-workup-of-acute-leukemia). A detailed summary of results designated according to the International System for Human Cytogenomic Nomenclature (ISCN),¹⁵ and images of representative karyograms and ancillary fluorescence in situ hybridization images were submitted by the laboratories for central review by 2 cytogeneticists. The complete karyotype ISCN designation was retained to enable analysis for recurring rearrangements or breakpoints, and for coding of balanced rearrangements. Each karyotype was scored for complexity defined as presence of ≥3 independent abnormalities, including ≥1 structural abnormalities.¹⁶

Flow cytometry

Bone marrow samples were submitted at the end of the first course of induction therapy to a single central reference laboratory (Hematologics Inc, Seattle, WA) and stained with a standardized panel of monoclonal antibodies (supplemental Data), designed to detect residual disease by using the difference from normal approach and adjusted for DS-specific, nonmalignant hematopoietic regeneration, following chemotherapy (ΔN).17, 18, 19, 20, 21, 22 Specimens were processed as previously described.^17,18 All data were reviewed by 2 independent analysts, who were blinded to patient information, and came to agreement on a patient’s residual disease status prior to issuing a report. A level of 0.05% of total nucleated cells was chosen as the clinical cut-off based on threshold of detection and regulatory agency requirements. Results were reported through a web-based platform (Rave EDC, iMedidata).

Mutational analysis

Whole-genome sequencing (n = 84) and transcriptome sequencing (n = 95) data were downloaded from the Kids First and INCLUDE data portal, CAVATICA, under project ID: Germline and Somatic Variants in Myeloid Malignancies in Children. Upon mapping the whole-genome sequencing data to hg38 using bwa aln,²³ single-nucleotide variants and indels were analyzed by Bambino et al,²⁴ followed by a Fendprocess filtering,²⁵ copy number variation by CONSERTING,²⁶ and structural variation by CREST.²⁷ Sequencing errors were computationally suppressed using CleanDeepSeq²⁸ and SequencErr.²⁹ All mutations were manually inspected in IGV.³⁰ For RNA-sequencing data, a custom script was developed to call variants. Taking GATA1 as an example, reads mapped (STAR³¹) to GATA1 region were extracted and remapped to GATA1 mRNA sequence NM_002049 using bwa aln.²³ Perfectly mapped reads were analyzed using SequencErr²⁹ to obtain allele counts to call variants, where sites with number of mutant reads ≥3 were considered by comparing allele fractions against all remaining samples, and sites with allele fractions >98% of remaining samples were manually reviewed along with their annotation (nonsilent) to determine validity. Reads with nonperfect mapping were called for indel³² or for structural variant³³ if the indel was longer, followed by manual review.

Statistical analysis

Data were current as of 30 June 2024, with a median follow-up of 4.91 years (range, 1.99-8.20) for patients who were HR. The significance of observed proportions was tested using Pearson χ² and Fisher exact test when data were sparse. The Kaplan-Meier method was used to estimate OS and EFS. OS was defined as time from study entry, EOI-1, or from relapse until death. EFS was defined as time from EOI-1 to failure by not achieving complete remission, by relapse, second malignancy, or death. Patients who experienced refractory disease with ≥5% bone marrow blasts after Induction II or who experienced a relapse or were not in complete remission by the EOI-III were defined as an induction failure. The cumulative incidence of relapse was obtained by methods that account for competing events, and was defined as the time from EOI-1 to relapse where deaths without a relapse were competing events. Gray test was used for comparisons. Patients lost to follow-up were censored at the date of last contact. Survival probabilities were reported with 95% confidence intervals (CI) calculated by the log-log transformation. This study was designed to compare patients from AAML1531 who were HR against a fixed 2-year EFS of 76%, which was observed for comparable patients with ML-DS (ie, MRD-positive at EOI-1) treated on AAML0431. Assuming a null EFS of 76% at 2 years, there was 95% power to detect an alternative EFS of 88% at 2 years with 1-sided testing at the 10% level of statistical significance if there were 41 patients who were HR who continued to Induction II. The study design included interim monitoring after ∼50% of the expected number of EFS events had been observed. Monitoring for insufficient EFS from the EOI-1 of the treatment for patients who were HR utilized monitoring based on the Lan-DeMets criterion with an α-spending function αt² (truncated at 3 standard deviations) and 10% type I error. A Woolson 1-sample log-rank test was used to compare the observed EFS with AAML0431 EFS for patients who were MRD-positive who continued to Induction II. EFS for AAML0431 was characterized by a separate cure model for patients who were HR: S(t) = 0.75 + 0.25∗exp(˗0.0037∗t), where t was measured in days. A P-value less than the boundary value of .025 for 50% information time would result in rejection of the hypothesis. The study specified comparing the Kaplan-Meier estimate of 2-year EFS (KM2) vs AAML0431 using a test statistic of ln(˗ln(KM2))˗ln(˗ln(0.76))/(Estimated standard deviation of KM2∗(ln(˗ln(KM2))) compared with a standard normal distribution.

Results

There were 280 patients (231 SR, 43 HR, 5 not assigned to a risk group, 1 ineligible), enrolled on AAML1531 as of 30 June 2025. In total, 41 patients had positive MRD at EOI-1 and were stratified into the HR group (2 patients who were HR came off study prior to Induction II; Figure 2). Patient characteristics are summarized in Table 1. The reference group consisted of patients who were MRD-positive in the predecessor study AAML0431, who were treated with DS-specific AML therapy of standard intensity (Figure 3).With the exception of a higher platelet count, characteristics of patients who were HR (AAML1531 Arm B, defined by positive flow cytometric MRD) with available sequencing data were not significantly different from other patients who were HR (Table 1). Similarly, characteristics of the combined SR (defined by negative MRD) and HR cohorts with available sequencing data were not significantly different, except for a lower proportion of CNS2b (supplemental Table 1). Intensification of chemotherapy for patients with ML-DS who were MRD-positive in the AAML1531 HR group did not result in an improvement of the 2-year EFS (80.5% [95% CI, 64.8%-89.7%]; P = .247) compared with 76% for the AAML0431 EOI-1 MRD-positive cohort. At 2 years, the OS was 82.9% (95% CI, 67.5%-91.5%) for the AAML1531 HR cohort compared with 76.2% for the AAML0431 EOI-1 MRD-positive cohort.

Figure 2.

CONSORT diagram. AAML1531 HR cohort. PR, partial response; RD, refractory disease.

Table 1.

Patient characteristics of the (MRD-positive) high-risk group of AAML1531

Characteristic	AAML1531 Patients who were HR						P value
	Patients who were HR (N = 41)		Without sequencing data (N = 22)		With sequencing data (N = 19)
	n	%	n	%	n	%
Sex
Male	22	53.7	10	45.5	12	63.2	.257
Female	19	46.3	12	54.5	7	36.8
Race
American Indian or Alaska Native	0	0.0	0	0.0	0	0.0	–
Asian	3	8.6	2	10.0	1	6.7	1.000
Black or African American	9	25.7	5	25.0	4	26.7	1.000
White	23	65.7	13	65.0	10	66.7	.678
Multiple races	0	0.0	0	0.0	0	0.0	–
Unknown	6		2		4
Ethnicity
Hispanic	3	8.8	2	10.5	1	6.7	1.000
Not Hispanic	31	91.2	17	89.5	14	93.3
Unknown	7		3		4
History of TAM
No	29	70.7	17	77.3	12	63.2	.322
Yes	12	29.3	5	22.7	7	36.8
Received prior treatment for TAM
No	11	91.7	4	80.0%	7	100.0%	.217
Yes	1	8.3	1	20.0	0	0.0
Morphology of myeloid leukemia associated with DS
AML	27	81.8	15	88.2	12	75.0	.398
At least MDS-EB2	2	6.1	1	5.9	1	6.3	1.000
MDS-EB2	2	6.1	0	0.0	2	12.5	.227
At least MDS-EB1	0	0.0	0	0.0	0	0.0	–
MDS-EB1	2	6.1	1	5.9	1	6.3	1.000
At least MDS-SLD	0	0.0	0	0.0	0	0.0	–
MDS-SLD	0	0.0	0	0.0	0	0.0	–
Not specified	8		5		3
CNS status
1	30	88.2	14	77.8	16	100.0	.105
2a	1	2.9	1	5.6	0	0.0	1.000
2b	1	2.9	1	5.6	0	0.0	1.000
2c	1	2.9	1	5.6	0	0.0	1.000
3a	0	0.0	0	0.0	0	0.0	–
3c	1	2.9	1	5.6	0	0.0	1.000
Unknown	7		4		3
Patient diagnosed with non-CNS extramedullary disease at study entry
No	39	95.1	20	90.9	19	100.0	.490
Yes: orbit site	2	4.9	2	9.1	0	0.0
Age in years from study entry, median (range)	1.58	(0.63-2.95)	1.57	(0.78-2.38)	1.82	(0.63-2.95)	.497
Bone marrow blast, median (range), %	31.4	(4-90)	31.4	(4-90)	31	(7-90)	.817
Peripheral WBC count, median (range), × 103/μL	7.7	(2.7-165)	7	(2.7-90.9)	8.1	(2.9-165)	.629
Peripheral platelet count, median (range), × 103/μL	66	(6-450)	58	(6-228)	80	(11-450)	.045
Peripheral blasts, median (range), %	17	(0-89)	20	(0-82)	10	(0-89)	.587
Hemoglobin, median (range), g/dL	9.5	(3.5-116)	8.75	(3.5-116)	9.9	(5.3-17.3)	.314

CNS, central nervous system; MDS-EB1, myelodysplastic syndrome with excess blasts-1; MDS-EB2, myelodysplastic syndrome with excess blasts-2; MDS-SLD, myelodysplastic syndrome with single lineage dysplasia; TAM, transient abnormal myelopoiesis; WBC, white blood cell.

Figure 3.

EFS and OS (at 2 years) of patients who were MRD-positive (EOI-1) and treated with intensified chemotherapy on Arm B of AAML1531 compared with patients who were MRD-positive and treated with standard intensity chemotherapy on the predecessor study AAML0431.

Among the HR group, there were 7 relapses (at a median of 172 days from EOI-1, range 41-617) and 1 death (during follow-up at 211 days from study entry) as a first event. All relapsed patients received further chemotherapy (cytarabine in 7, anthracycline in 2, and gemtuzumab ozogamicin in 5). Of the 7 patients who relapsed, 6 did not survive (2-year OS post-relapse 14.3% [95% CI, 0.7%-46.5%]; Figure 4). One patient was alive 10 months after allogeneic stem cell transplant.

Figure 4.

OS of (MRD-positive) AAML1531 patients who were HR after relapse.

Intensification of chemotherapy did not result in a survival benefit for patients with ML-DS who were MRD-positive, and increased infection-related morbidity (Table 2). The course-specific frequency of febrile neutropenia (grade ≥3) was significantly higher among patients treated with intensified therapy on the AAML1531 HR arm (Arm B) compared with the reduced-intensity therapy (no high-dose cytarabine) used for patients who were MRD-negative on the SR arm (Arm A): HR vs SR Induction II, 31.7% vs 3.7%, P < 0.01; Intensification I, 27.5% vs 6.1%, P = 0.001; Intensification II, 26.3% vs 8.6%, P = 0.008. Similarly, episodes of sepsis grade ≥3 were significantly more frequent among patients who were MRD-positive and treated with intensified chemotherapy on the AAML1531 HR arm (Arm B): HR vs SR Induction II, 9.8% vs 0%, P = 0.005; Intensification I, 5.0% vs 0%, P = 0.083; Intensification II, 0% vs 1.1%, P= 1.0. In addition, oral mucositis was more common in patients treated on the HR compared with the SR arm during Induction II (7.3% vs 0%, P = 0.005) and Intensification II (5.3% vs 0%, P = 0.026; supplemental Table 3).

Table 2.

Course-specific frequency of infection-related adverse events among patients with ML-DS who were MRD-positive treated with intensified chemotherapy (Arm B) compared with patients who were MRD-negative treated with reduced-intensity chemotherapy on the SR Arm (Arm A) of AAML1531

AAML1531 SR				AAML1531 HR				P value
Cycle	Number of patients	Febrile neutropenia grade ≥3		Cycle	Number of patients	Febrile neutropenia grade ≥3
	N	N	%		n	n	%
Induction II	108	4	3.7	Induction II	41	13	31.7	<.001
Induction III	101	7	6.9	Induction III
Intensification I	98	6	6.1	Intensification I	40	11	27.5	0.001
Intensification II	93	8	8.6	Intensification II	38	10	26.3	0.008

Cycle	Number of patients	Sepsis grade ≥3		Cycle	Number of patients	Sepsis grade ≥3		P value
	N	n	%		n	n	%
Induction II	108	0	0.0	Induction II	41	4	9.8	0.005
Induction III	101	1	1.0	Induction III
Intensification I	98	0	0.0	Intensification I	40	2	5.0	0.083
Intensification II	93	1	1.1	Intensification II	38	0	0.0	1.000

Induction I was common to both arms (and previously described)¹³; Induction III has no corresponding course on the HR arm.

We then retrospectively analyzed the prognostic impact of molecular high-risk lesions in AAML1531 SR and HR cohorts. This molecular high-risk group was recently defined by the presence of any of 4 lesions (mutation/deletion of CDKN2A, or mutation of ZBTB7A, JAK2, TP53).¹⁴ Out of 280 patients enrolled in AAML1531, results of targeted sequencing were available for 96 patients (34%), who had consented to an optional biology study (1 not risk stratified, 76/114 patients on SR Arm A, and 19/41 patients on HR Arm B). Of those, 16 displayed a molecular high-risk marker (11 on SR Arm A, 4 on HR Arm B, and 1 not risk stratified), amounting to 16.7% of patients analyzed (11/76, 14.4% in the SR arm, and 4/19, 21.1% in the HR arm; supplemental Table 1). Demographics, World Health Organization classification, and blast morphology were not significantly different between those patients who harbored a molecular high-risk marker vs those who did not. However, at diagnosis, patients with a molecular high-risk had a higher peripheral white blood cell count, peripheral blast count, bone marrow blast percentage, and lower hemoglobin than patients lacking a molecular high-risk marker. The 2-year EFS was 50% (95% CI, 24.5%-71%) for the 16 patients in the molecular high-risk group compared with 91.1% (95% CI, 82.3%-95.7%; P < .001) for the remaining 80 patients who did not harbor a high-risk molecular marker. OS at 2 years was 56.3% (95% CI, 29.5%-76.2%) for patients in the molecular high-risk group compared with 91% (95% CI, 82.1%-95.6%; P < .001) for those lacking any molecular high-risk marker. Within the MRD-defined SR and HR arms of AAML1531, survival analysis according to the presence or absence of the molecular high-risk profile showed that in the HR group, 2-year EFS was 50% for patients with a molecular high-risk marker compared with 93.3% for those without; in the SR Group, EFS was 54.5% for patients with a molecular high-risk marker compared with 90.6% for those lacking molecular high-risk markers (Figure 5). A similar pattern was observed for OS at 2 years: 50% for patients who were HR who had a molecular high-risk marker vs 93.3% for those without; and an OS of 63.6% for patients in the SR group who had a molecular high-risk marker vs 90.5% for those without. The risk of relapse was significantly different, with 6.7% and 9.4% in the HR and SR groups, respectively, if no molecular high-risk marker was present compared with 50% in HR and 45.5% in the SR group for patients with a molecular high-risk marker (supplemental Table 2). Cytogenetic findings were not statistically significantly different between patients who were HR (MRD-positive) and SR (MRD-negative) enrolled in AAML1531, aside from a trend towards a higher frequency of complex karyotypes in the former group. No cytogenetic features, including karyotype complexity, correlated with the presence of molecular high-risk markers.

Figure 5.

EFS and OS of patients with ML-DS enrolled in AAML1531 according to MRD vs presence of a molecular high-risk marker.

Discussion

Intensification of chemotherapy for patients with ML-DS with positive flow cytometric EOI-1 MRD did not improve survival, but resulted in increased toxicity. The course-specific proportion of patients experiencing grade ≥3 febrile neutropenia was significantly higher for patients receiving highly intensive non-DS AML-type chemotherapy on the HR Arm of AAML1531 (26.3%-31.7%) than for patients treated with reduced intensity on the SR arm (no high-dose cytarabine), but was comparable to the most intensive course of the predecessor ML-DS study AAML0431 (Induction II, 29.7%), which contained high-dose cytarabine.⁸ Among patients who lacked molecular high-risk markers, EFS and OS of those with positive EOI-MRD treated with intensified chemotherapy on the HR arm were comparable to those of patients who were MRD-negative treated on the SR arm (supplemental Table 2).

In marked contrast to the favorable outcome of primary ML-DS, the probability of survival for patients with relapsed or refractory disease remains dismal. OS of patients with relapsed or refractory ML-DS was 14% at 2 years in the MRD-positive HR group of AAML1531, similar to the MRD-negative SR group (Arm A, 16.7% at 1 year)^13,34 and results of an international retrospective study of relapsed or refractory ML-DS (3-year OS 22%).¹¹ Following the stepwise reduction of treatment intensity for ML-DS in clinical trials conducted by several study groups,7, 8, 9^,13,35,36 treatment-related mortality no longer is excessive (32%³⁷ vs 1.8% in AAML1531, and 1.4% in AAML0431⁸), and relapse has become the main cause of failure of contemporary ML-DS treatment. This fact highlights the lack of efficacy of conventional chemotherapy for relapsed/refractory ML-DS regardless of MRD level, and the need to prioritize the reduction of relapse events while maintaining low treatment-related mortality in the design of future trials for patients with ML-DS.

Our findings illustrate the need for better predictive markers in ML-DS that could also identify targetable lesions. We previously found that negative flow cytometric EOI-1 MRD did not identify a subset of patients with ML-DS for whom treatment intensity could be reduced (by omission of a course of high-dose cytarabine and asparaginase) without adversely affecting EFS and OS.¹³ As a result, the inclusion of a course of high-dose cytarabine (HD Ara-C)/E. coli asparaginase in the treatment of ML-DS is recommended for all patients with ML-DS. Patients who continued to enroll on AAML1531 who were found to have a MRD level <0.05% in the bone marrow at the end of the Induction course I, came off study and continued treatment as per the predecessor study AAML0431 (including a course of high-dose cytarabine/E. coli asparaginase as second course of induction therapy). We now report that intensification of conventional chemotherapy for those patients with ML-DS with positive EOI-1 MRD does not improve survival. These results contrast with the independent prognostic impact of EOI-1 MRD in pediatric non-DS AML,¹⁷ but are reminiscent of the lack of prognostic value of flow cytometric EOI-1 MRD in the most favorable cytogenetic and molecular subset of non-DS AML,^17,38,39 and demonstrate that in the current study flow cytometric EOI-1 MRD was unsuccessful as a predictive marker to select treatment intensity for ML-DS.

A recent retrospective analysis of 177 patients with ML-DS who had been uniformly treated on 2 studies by the Japan Children’s Cancer Group suggested that alterations in any of 4 genes (collectively present in 24.3% of patients) were prognostic in ML-DS, specifically deletion/mutation of CDKN2A, or mutation of ZBTB7A, JAK2, or TP53. EFS was 66.6% and OS 69.0% at 3 years in those with 1 of the 4 lesions, compared with 94.2% and 95.6%, respectively, for those patients lacking a molecular high-risk marker.¹⁴ Mutations of TP53 (3-year EFS 25.0%) and deletion/mutation of CDKN2A (3-year EFS 28.6%) had a particularly large adverse impact on the prognosis of ML-DS. We retrospectively analyzed outcomes of a subset of patients from AAML1531 for whom sequencing data were available (34%), and found that the presence of any of the 4 molecular high-risk markers in 15.8% of patients from AAML1531 who were analyzed correlated with EFS for both patients who were MRD-positive and MRD-negative (50% compared with 93.3% if no molecular risk marker present in patients who were HR), OS, and relapse risk (50.0% compared with 6.7% if no molecular risk marker present in patients who were HR; Figure 4 and supplemental Table 2). These results show that the molecular high-risk profile is prognostic independent of MRD level in patients with ML-DS. Our analysis validates the high-risk molecular markers of ML-DS in a second patient cohort and in the different treatment context of a COG clinical trial. Although our patient numbers were small, we note that among the patients who were MRD-positive who were stratified into a HR group on AAML1531 and treated with highly intensive non-DS AML chemotherapy, the presence of a high-risk molecular marker was associated with low EFS, OS, and high relapse risk, suggesting that patients with molecular high-risk ML-DS are unlikely to benefit from the intensification of conventional chemotherapy alone.

Future ML-DS clinical trials will be required to evaluate the new molecular high-risk markers prospectively as well as agents that may target the functional effects of underlying alterations, such as a JAK2 inhibitor for ML-DS with activating JAK2 mutations or a CDK4/6 inhibitor for patients with the highly unfavorable deletion/mutation of CDKN2A, in an effort to improve survival outcomes in this high-risk group.

Conflict-of-interest disclosures: L.E.B. was employed and had equity ownership in Hematologics, Inc at the time of study. M.L. has employment and ownership of Hematologics, Inc. The remaining authors declare no competing financial interests.