Entospletinib in Combination with Induction Chemotherapy in Previously Untreated Acute Myeloid Leukemia: Response and Predictive Significance of HOXA9 and MEIS1 Expression

Acute myeloid leukemia (AML) is a biologically heterogeneous hematologic malignancy characterized by a reduction of normal hematopoietic cell production and proliferation of leukemic blasts in the blood and bone marrow (BM). Despite an improved understanding of mechanisms of leukemogenesis, primary refractoriness to chemotherapy and frequent relapses result in poor long-term disease-free survival and overall survival (OS) in the majority of patients (1). A key focus of pharmacologic innovation in AML treatment has been to target molecular mutations present within leukemic blasts (2).

Spleen tyrosine kinase (SYK) is a nonreceptor tyrosine kinase involved in cellular proliferation, differentiation, and survival that is expressed broadly in most hematopoietic cells (3). The loss of SYK expression in AML cell lines is associated with morphologic evidence of differentiation and expression of mature myeloid cell surface markers, suggesting that SYK plays a role in counteracting the differentiation of leukemic cells (4). SYK protein expression appears to be modulated by HOXA9 and MEIS1, homeodomain-containing transcription factors that are overexpressed in approximately 30%–40% of AML cases and correlate with a poor prognosis (5–8). Recent data have shown that overexpression of HOXA9 and MEIS1 leads to an upregulation of SYK protein and increased SYK activity (9). Furthermore, transformation of myeloid progenitors with HOXA9 and MEIS1 in preclinical models induced addiction to SYK signaling. Accordingly, pharmacologic inhibition or knockdown of SYK significantly reduced tumor burden and prolonged survival in AML mouse models (9). SYK signaling occurs via stimulation of β-integrin and Fcγ receptors resulting in activation of STAT3 and STAT5 transcription factors and promotion of leukemic cell proliferation (10). Constitutive activation and direct phosphorylation of the FLT3 receptor by SYK has also been reported (4, 11).

Entospletinib is an orally bioavailable, selective inhibitor of SYK that binds to the ATP pocket of the active site and disrupts the kinase activity of the enzyme. Kinase selectivity profiling has shown a more than 14-fold selectivity of entospletinib for SYK versus other kinases, as compared with the less selective SYK inhibitor, fostamatinib (12). Therapeutic activity of entospletinib has been evaluated in patients with B-cell malignancies, where it was found to be well-tolerated, demonstrating only modest single-agent activity as compared with other B-cell receptor signaling agents (13–15).

Given the role of SYK signaling in leukemic cell proliferation and differentiation, we chose to explore the activity of entospletinib in combination with 7+3 induction chemotherapy. Herein, we report results of a phase Ib/II study of patients with previously untreated AML who received entospletinib (with a 14-day monotherapy lead-in to evaluate the effects on myeloid differentiation and response) and intensive chemotherapy. We also describe a biomarker analysis exploring the hypothesis that entospletinib may be more effective in patients with high baseline HOXA9 and MEIS1 mRNA expression.

This was an international multicenter phase Ib/II study (NCT02343939) conducted from July 2015 to February 2018 and consisted of two parts: phase Ib, entospletinib dose escalation (200 and 400 mg twice daily) + 7+3 in part 1 and phase II, entospletinib (400 mg twice daily) + 7+3 dose expansion. The study was conducted in accordance with the Declaration of Helsinki, Good Clinical Practice guidelines, and relevant regulatory laws. The study protocol was approved by each center's institutional review board. All patients provided written informed consent.

Patients aged ≥18 years with previously untreated AML by World Health Organization criteria (16), Eastern Cooperative Oncology Group (ECOG) performance status ≤2, left ventricular ejection fraction ≥ 45%, and life expectancy ≥3 months were eligible. Exclusion criteria included a diagnosis of acute promyelocytic leukemia, known active central nervous system or leptomeningeal leukemic involvement, history of active nonmyeloid malignancies or allogeneic stem cell transplant (ASCT), uncontrolled systemic infections, known active hepatitis C, hepatitis B, cirrhosis or ongoing liver injury from any cause, drug-induced pneumonitis, or inflammatory bowel disease. Use of proton pump inhibitors, moderate CYP2C9, and strong CYP3A and CYP2C9 inducers was not allowed because of expected reduction in entospletinib absorption. All patients underwent a baseline BM aspiration and biopsy and were risk stratified according to the European Leukemia Network (ELN) 2010 classification (17).

Patients received entospletinib monotherapy every 12 hours as a lead-in for 14 days (cycle 0). However, induction chemotherapy could be started earlier based on medical need, as determined by the investigator. Entospletinib was continued daily in combination with 7+3 (cytarabine 100 mg/m²/day, days 1–7 plus daunorubicin 60 mg/m²/day, days 1–3) for up to two induction cycles (cycles 1 and 2). Hydroxyurea was permitted during cycle 0 for rising white blood cell (WBC) count. The phase Ib portion of the study consisted only of induction (no consolidation on study). In the phase II portion of the study, patients who achieved complete remission (CR) with incomplete blood count recovery (CRi) received postremission chemotherapy in combination with entospletinib, followed by entospletinib maintenance (≤12 cycles). Patients were removed from study after induction at any time for allogeneic stem cell transplant (ASCT) at the discretion of the treating physician. The study design is outlined in Supplementary Fig. S1. Postremission therapy consisted of age-adjusted high-dose cytarabine (HiDAC) chemotherapy (3 g/m² HiDAC i.v. every 12 hours on days 1, 3, and 5 for patients aged <60 years or 1 g/m² cytarabine i.v. daily on days 1–5 for patients aged ≥60 years) in combination with 400 mg entospletinib (every 12 hours on days 1–28 of each 28-day cycle). Patients who maintained a CR/CRi after three cycles of postremission chemotherapy, and did not proceed to ASCT, were offered entospletinib maintenance (400 mg every 12 hours on days 1–28 of each 28-day cycle, up to 12 cycles).

Dose escalation followed a standard 3+3 design (dose level 1, 200 mg and dose level 2, 400 mg). Dose-limiting toxicities (DLT) were assessed during entospletinib monotherapy (cycle 0) and during induction (cycles 1 and 2). Patients who did not complete at least 21 days of entospletinib or missed any doses of 7+3 for reasons other than toxicity, were not evaluable for the DLT assessment and were replaced. However, if entospletinib was discontinued because of toxicity, this was DLT. Furthermore, grade 4 nonhematologic toxicities attributable to entospletinib except for alopecia, nausea, and vomiting controllable with antiemetic therapy, line-associated venous thrombosis, infection (infection-related toxicities such as fever/sepsis), and fatigue were considered DLTs. The phase II expansion dose was 400 mg entospletinib twice daily based on tolerability in this trial and outside pharmacokinetics studies suggesting this was the optimal dose.

A BM aspiration was performed at the end of cycle 0 to assess the effect of entospletinib monotherapy. All patients proceeded to induction chemotherapy at the end of cycle 0 regardless of the marrow result. After first induction, patients underwent a BM biopsy on cycle 1, day 14. Those with residual disease, received a second induction with the same schema as cycle 1. If CR or CRi was not achieved by the end of cycle 2, this was considered a treatment failure and the patient was removed from the study. BM aspirate samples were collected for disease assessment and biomarker research at the end of every two cycles of postremission chemotherapy and at the end of every four cycles of maintenance. Assessments of clinical response were made according to the revised International Working Group criteria (18). Cytogenetic and molecular mutation testing were done at baseline and repeated at subsequent BM examinations.

BM mononuclear cells (BM-MNCs) from BM aspirates obtained at baseline were analyzed for mRNA expression of HOXA9 and MEIS1 using a custom assay. Specifically, RNA was extracted from BM-MNCs using the miRNeasy Kit (Qiagen Ltd.), following the manufacturer’s instructions. The probe sets for HOXA9 and MEIS1 were custom designed and added to the nCounter PanCancer pathway panel. The NanoString nCounter System (NanoString Technologies, Inc.) was used to measure the gene expression profiles with an input of 100 ng of total RNA. Expression data were first normalized using the NanoStringNorm R package with 18 housekeeping genes. HOXA9 and MEIS1 expression levels were then normalized to expression from pooled healthy BM-MNCs (n = 20). Median value of the average normalized HOXA9 and MEIS1 expression was used as cutoff to define HOXA9:MEIS1 high or low expression groups. Clinical response, event-free survival (EFS), and OS were compared between HOXA9 and MEIS1 expression groups.

The mutational status of NPM1 and FLT3 (ITD and TKD), and KMT2A/mixed lineage leukemia (MLL) gene rearrangements was determined by clinically validated assays in the hospital laboratories of patients’ respective institutions. The variant allele frequency cutoff for variant calling was set to 0.10. Mutations were evaluated for potential associations with outcomes and HOXA9 and MEIS1 expression.

In the phase Ib portion of the study (induction only), the primary endpoint was determination of the recommended phase II dose of entospletinib in combination with chemotherapy. In the phase II portion of the study (induction and postremission), the primary endpoint was composite CR rate (proportion of patients who achieved CR or CRi) at induction completion. Secondary endpoints included the occurrence of adverse events (AEs), EFS (defined as time from the start of the study therapy until the date of treatment failure, AML relapse, or death from any cause, whichever occurred first), and OS (defined as the interval from the start of study therapy to death from any cause). The study also included relapse-free survival (RFS), defined as time from the date of attaining CR/CRi until the date of AML relapse or death from any cause, whichever occurred first, as an exploratory endpoint. The planned sample size was up to 14 patients in the phase Ib portion [based on two planned dose levels (200 and 400 mg) with up to six subjects per level and 10% are unevaluable] and approximately 40 additional patients in the phase II cohort. This sample size ensured a narrow confidence interval (CI; ∼7%–14% distance from the point estimates), based on the CR rate of standard chemotherapy (7+3), which was reported as 52% in a Cancer and Leukemia Group B study of more than 1,000 subjects (19).

Patients who received ≥1 dose of study treatment were included in the efficacy and safety analyses. Descriptive summary statistics were computed for patient characteristics, categorical efficacy endpoints (with corresponding 95% CIs), and safety variables. Kaplan–Meier estimates were used for EFS, OS, and RFS, and their 95% CIs.

Fifty-three patients (n = 12, phase Ib and n = 41, phase II) with previously untreated, de novo (n = 39) or secondary (n = 14) AML were enrolled (58% male, median age 60 years; Table 1; Supplementary Table S1) in this study. The majority of patients (n = 30) were intermediate II or adverse risk per ELN 2010 criteria. No patients with core binding factor AML were enrolled. All patients had been deemed fit for intensive chemotherapy. Patient disposition is shown in Fig. 1. Thirty-seven (70%) patients achieved a remission (CR, n = 27 and CRi, n = 10). Sixteen patients (30%) did not achieve remission after the protocol-specified two cycles of induction. Of the 41 patients enrolled in the phase II portion, where postremission therapy was part of the trial, 15 patients (37%) received one to three cycles of HiDAC, of whom six (15%) continued to receive maintenance entospletinib monotherapy. Twenty-two (42%) patients went to ASCT after achieving CR/CRi on study, of which 15 underwent ASCT immediately after induction and an additional seven in the phase II portion underwent stem cell transplant (SCT) after one to two cycles of HiDAC as postremission therapy while waiting for a donor. The median duration (range) of entospletinib exposure was 7.1 (0.9–72.9) weeks overall; 6.2 (0.9–10.0) weeks in phase Ib and 9.3 (1.1–72.9) weeks in phase II. Sixteen (30%) patients did not receive the full 14-day lead-in due to concern for progression of disease based on rising WBC count or other clinical symptoms, or patient request (ranged from 5 to 13 days of lead-in). Thirteen (25%) patients required hydroxyurea due to rising WBC counts [four (8%) prior to starting entospletinib and nine (17%) while on entospletinib].

There were no DLTs in the phase Ib dose-escalation cohort. The phase II dose was established as 400 mg twice daily on the basis of the phase Ib data and additional sponsor experience indicated that dose proportional pharmacokinetics were lost at higher doses (20).

Entospletinib alone or in combination with chemotherapy was well-tolerated. Most of the AEs that occurred on treatment were consistent with those expected following treatment with 7+3 (Table 2). Common treatment-emergent (TE) hematologic AEs/laboratory abnormalities with severity grade ≥3 by Common Terminology Criteria for Adverse Events (version 4.03) included febrile neutropenia (n = 44, 83%), leukopenia (n = 49, 92%), thrombocytopenia (n = 41, 77%), anemia (n = 28, 53%), and neutropenia (n = 19, 36%). Gastrointestinal AEs were mostly grade 1 or 2, with diarrhea being the most common grade 3 gastrointestinal TEAE (10%). The most common nonhematologic TEAEs/laboratory abnormalities with severity grade ≥3 included lung infection (n = 11, 21%), device-related infection (n = 9, 17%), hypoxia (n = 9, 17%), rash (n = 7, 13%), and hyperbilirubinemia (n = 6, 11%). Although occurring in <15% of patients, grade ≥3 rash and hyperbilirubinemia were unique AEs attributable to entospletinib, which led to discontinuation of one patient each, respectively. Without regard to attribution, any rash was observed in 23 patients (43%), although grade 3 rash occurred in only seven patients (13%, none higher than grade 3). The time course of rash eruption varied among patients, with some developing rash during the lead-in and others during treatment with entospletinib + 7+3. In general, the rash was characterized as an erythematous, diffuse morbilliform rash that could be pruritic. Withholding entospletinib improved the rash to grade 1 within 10 days, allowing drug to be restarted; steroids and additional supportive care were used as deemed appropriate by the treating physician. The rash recurred in one of four patients when rechallenged and entospletinib had to be discontinued. Hyperbilirubinemia was predominantly indirect, consistent with the known effect that entospletinib has on inhibition of UGT1A1, leading to reversible increases in unconjugated bilirubin values. Serious TEAEs were reported in 23 (43%) patients and were considered related to entospletinib in seven (13%) patients; these included four patients with febrile neutropenia and one patient each with cognitive disorder, dyspnea, pneumonitis, and maculopapular rash. One patient developed a grade 3 lung infection that clinically was consistent with pneumonia; however, the investigator was unable to definitively rule out the possibility of pneumonitis. This event was considered related to entospletinib. The patient responded to antibiotics, antifungals, and steroids with resolution of symptoms at the time of count recovery.

Overall, 18 (34%) patients required entospletinib dose interruptions or reduction due to AEs; the TEAEs leading to dose interruptions or reductions occurring in ≥3 patients were febrile neutropenia (grade 3–4), and hyperbilirubinemia and maculopapular rash (both grade 2–3). Nine (17%) patients discontinued study drug due to AEs, 1 each for angioedema, increased blood bilirubin, cerebrovascular accident, cognitive disorder, dyspnea, gastric hemorrhage, homicidal ideation, maculopapular rash, and sepsis.

There was no TEAE leading to death. Two deaths (4%, both due to disease progression) occurred within 30 days after last dosing date. The 30-day induction mortality rate was 0%.

Entospletinib + 7+3 resulted in a CR rate of 51% with a CR with incomplete blood count recovery (CRi) rate of 19% and composite CR rate (CR + CRi) of 70% (Table 3). Of the 10 patients with CRi at induction completion, minimal residual disease (MRD) assessment by flow cytometry was available for seven patients, of whom four were MRD positive and three MRD negative, demonstrating delayed count recovery from myelotoxicity rather than suboptimal response. Fourteen (26%) patients had secondary AML, and the composite CR rate in this group was 64%.

We identified three AML subsets where the composite CR rate was noted to be higher than that of the entire group: FLT3-ITD (n = 6; CR, 83%), NPM1 (n = 15; CR, 87%), and patients with KMT2A gene rearrangements (n = 10; CR, 90%; Table 3). Responses occurred across all KMT2A rearrangements, including t(6:11) (Supplementary Table S1). Only one of 14 patients with secondary AML had an MLL rearrangement. One patient with t(9;11) achieved a morphologic and cytogenetic CR with incomplete count recovery after cycle 0 (before chemotherapy); the patient subsequently continued on study with induction chemotherapy and ultimately received ASCT in CR1.

After a median follow-up of 26.2 months, the median OS was 37.1 [95% CI, 16.8–not available (NA)] months (Supplementary Fig. S2). The median (95% CI) EFS and RFS were 9.0 (2.3–NA) months and 14.8 (7.7–NA) months. No significant differences were observed between patients with CR (n = 27) and CRi (n = 10).

Baseline BM-MNC samples were available from 34 patients for HOXA9 and MEIS1 expression analysis. The composite CR rate was similar between the full patient set and the HOXA9:MEIS1 available set (70% vs. 71%; Table 3) and the distributions of the ELN risk groups were not significantly different between these two sets. Overall, there were no significant differences in HOXA9 and MEIS1 expression between patients who achieved a CR/CRi (n = 24) and non-CR patients (n = 10; P = 0.72 for HOXA9; P = 0.79 for MEIS1, Student t test). Among patients with high HOXA9 and MEIS1 expression, 76% (13/17) achieved a CR/CRi with entospletinib + 7+3, compared with 65% (11/17) of the patients with low HOXA9 and MEIS1 expression (Table 3). There were no differences in HOXA9 and MEIS1 expression between patients with de novo and secondary AML.

Analysis of OS data suggested that patients with high baseline HOXA9 and MEIS1 expression had significantly better OS (HR, 0.32; 95% CI, 0.100–0.997; P = 0.038, log-rank test; Fig. 2). However, it should be noted that the groups were not balanced for cytogenetic or molecular risk. Significantly higher HOXA9 and MEIS1 expression was observed in patients with AML with KMT2A gene rearrangements (n = 6) and NPM1 mutations (n = 10, 3 with concomitant FLT3-ITD; P < 0.05) as compared with respective wild-type groups (n = 28 for KMT2A wild-type and n = 24 for NPM1 wild-type; Fig. 3; Supplementary Fig. 3).

This is the first report of the small-molecule SYK inhibitor, entospletinib, given in combination with standard induction chemotherapy in patients with AML. Incorporation of a monotherapy lead-in as part of the trial design was feasible and allowed for preliminary assessment of single-agent activity as well as tolerability. Notably, entospletinib monotherapy led to a morphologic and cytogenetic remission in one patient with t(9;11) AML. This is the first report of a patient with KMT2A-AML responding to entospletinib monotherapy and the first signal of clinical activity of this drug within this cytogenetic subgroup, consistent with preclinical data suggesting efficacy in this subset (9).

Clinical responses were observed broadly in both de novo and secondary AML, across ELN risk groups, and in select molecular subsets. The composite CR rate in this study was 70%, comparable with what we would expect from 7+3 induction chemotherapy alone in an AML study of all risk types. Although none of the other patients with KMT2A-rearranged AML achieved a morphologic or cytogenetic response with monotherapy alone, the composite CR rate for this group was 90%. In general, entospletinib was well-tolerated with no 30-day induction mortality in this trial. Other AEs observed were consistent with what are commonly observed following 7+3 chemotherapy with the development of cytopenias, febrile neutropenia, and infections. Although there were no DLTs observed during the dose escalation, unique toxicities attributable to entospletinib that required a dose adjustment were transaminitis and indirect hyperbilirubinemia; especially notable was rash. The erythematous morbilliform rash that developed in patients was diffuse, pruritic, and tended to resolve in 7–10 days by withholding entospletinib. However, some patients did develop the rash again upon rechallenge, supporting this is related to entospletinib exposure.

Given the observed response to monotherapy in a patient with t(9;11) AML, we sought to determine whether there were additional molecular or cytogenetic subsets that may be highly sensitive to SYK inhibition with entospletinib. We identified three AML subsets where the composite CR rate was noted to be higher than that of the entire group: FLT3-ITD (n = 6; CR, 83%), NPM1 (n = 15; CR, 87%), and patients with KMT2A gene rearrangements (n = 10; CR, 90%). Notably, all three subsets are associated with high HOXA9 and MEIS1 expression, which was confirmed in this study (Fig. 3). Both HOXA9 and MEIS1 are critical to leukemic cell survival, and high coexpression of HOXA9 and MEIS1 results in increased SYK protein levels in AML (9). Furthermore, high expression of HOXA9 alone or in combination with MEIS1 is a poor prognostic factor in patients with AML treated with standard-of-care therapies (6). In our study, improved OS in the high HOXA9 and MEIS1 expression population, patients with KMT2A gene rearrangements, and NPM1 mutations, was consistent with preclinical findings, suggesting that AML subtypes with increased HOXA9 and MEIS1 expression are addicted to SYK signaling and may be more sensitive to entospletinib treatment. However, HOXA9 and MEIS1 expression data were unavailable for nearly one-third of the patients, which may confound interpretation of biomarker analyses. Given the small sample size, the notable, but preliminary data on the predictive utility of HOXA9 and MEIS1 expression should be evaluated further in a larger study. Finally, although it was not possible to directly measure via IHC the degree of SYK inhibition, chemokines downstream of SYK such as CCL3 and CCL4 were decreased following entospletinib therapy. However, both the baseline level and the level of decrease were not significantly different between HOXA9/MEIS1 high and low expression patients. One possible explanation for this is that peripheral chemokine levels are not sensitive enough to reflect the chemokine levels in BM between the HOXA9/MEIS expression groups.

Patients with KMT2A/MLL gene rearrangements historically have a wide range of CR rates depending on the translocation partner and corresponding genetic fusion. Despite reported CR rates ranging from 47% to 87.5% (21, 22), patients with KMT2A have low survival rates if treated without an ASCT. In our study, patients with KMT2A gene rearrangements achieved a CR of 90% (9/10) with entospletinib in combination with standard chemotherapy; the combination was well-tolerated and did not appear to result in toxicities that would preclude transplantation (indeed, six of these patients went on to receive ASCT), thus making SYK inhibition with entospletinib an acceptable induction chemotherapy option. Furthermore, based on the one patient with KMT2A-rearranged AML with a morphologic and cytogenetic remission on monotherapy, the clinical activity of entospletinib should be explored further in these patients.

Therapeutic innovation in AML requires that we develop drugs and choose treatment regimens that integrate disease-specific molecular and cytogenetic information to maximize response while minimizing toxicity. On the basis of the results of this study, gene expression patterns may also be targeted, similar to our approach to molecular mutations, to inform treatment selection for patients.

A.R. Walker reports other from Gilead (clinical trial support) during the conduct of the study. J.S. Blachly reports grants and non-financial support from Gilead (clinical trial and correlative science) during the conduct of the study, personal fees from AbbVie (consulting), AstraZeneca (consulting), KITE Pharma (consulting), and Innate Pharma (consulting) outside the submitted work, and a patent for Leukemia Diagnostic Device pending (holder O.S. U., inventor J.S. Blachly). B. Bhatnagar reports other from Karyopharm Therapeutics (research support), Cell Therapeutics Inc. (research support), Novartis (advisory board honorarium), Astellas (advisory board honorarium), Kite (advisory board honorarium), Pfizer (advisory board honorarium), and Cell Therapeutics (advisory board honorarium) outside the submitted work. A.S. Mims reports personal fees from Jazz Pharmaceuticals (data safety and monitoring committee), Syndax Pharmaceuticals (scientific advisory board), Kura Oncology (scientific advisory board), and AbbVie Pharmaceuticals (scientific advisory board) outside the submitted work. S. Orwick reports grants from NIH during the conduct of the study. T.L. Lin reports other from Gilead (support to the institution for conduct of the clinical trial) during the conduct of the study, Aptevo (support to the institution for conduct of the clinical trial), Biopath Holding (support to the institution for conduct of the clinical trial), Astellas (support to the institution for conduct of the clinical trial), Novartis (support to the institution for conduct of the clinical trial), Seattle Genetics (support to the institution for conduct of the clinical trial), Celyad (support to the institution for conduct of the clinical trial), Prescient Therapeutics (support to the institution for conduct of the clinical trial), Genentech (support to the institution for conduct of the clinical trial), Incyte (support to the institution for conduct of the clinical trial), Ono Pharmaceuticals (support to the institution for conduct of the clinical trial), AbbVie (support to the institution for conduct of the clinical trial), Celgene (support to the institution for conduct of the clinical trial), Jazz (support to the institution for conduct of the clinical trial), Mateon (support to the institution for conduct of the clinical trial), Pfizer (support to the institution for conduct of the clinical trial), Tolero (support to the institution for conduct of the clinical trial), and Trovagene (support to the institution for conduct of the clinical trial) outside the submitted work. H.E. Crosswell reports other from Bon Secours St Francis (study-related cost and support) during the conduct of the study and reports stock ownership of Gilead, BMS, AbbVie, Nucana, KIYATEC, agios, and Pfizer. D. Zhang reports other from Gilead Sciences, Inc. (employment and stock ownership) outside the submitted work. M.D. Minden reports other from Gilead (funding of clinical trial, no personal fees) during the conduct of the study, Astellas (funding clinical trial, no personal support) and Celgene (funding clinical trial, no personal support) outside the submitted work. V. Munugalavadla reports other from AstraZeneca/Acerta Pharma (current employee and stock ownership) and Gilead Sciences (stock ownership) outside the submitted work, and family member is an employee of Gilead Sciences. J. Liu reports personal fees from Gilead Sciences (employment and stock ownership) outside the submitted work. Y. Pan reports other from Gilead Sciences (employment and stock ownership) outside the submitted work, and a patent for entospletinib (pending). T. Oellerich reports grants from Gilead (preclinical research on SYK inhibitors) during the conduct of the study, grants and personal fees from Merck KGaA (research grants, consultant), personal fees from Kronos Bio (consultant), and Roche (consultant) outside the submitted work, and a patent for International Patent Application No. PCT/EP2017/057671 (Companion Diagnostics for Leukemia Treatment) issued to Goethe University Frankfurt. H. Serve reports grants, personal fees, and non-financial support from Gilead Inc. during the conduct of the study, grants from Gilead Inc. outside the submitted work, as well as a patent for the role of HOX/MEIS as therapeutic response predictor in AML (pending). A.V. Rao reports personal fees from Gilead Sciences Inc (stock options as employee) outside the submitted work. W. Blum reports grants and non-financial support from Gilead during the conduct of the study and personal fees from Syndax outside the submitted work. No potential conflicts of interest were disclosed by the other author.

A.R. Walker: Conceptualization, resources, data curation, formal analysis, investigation, methodology, writing-original draft, writing-review and editing. J.C. Byrd: Conceptualization, writing-original draft, writing-review and editing. J.S. Blachly: Resources, data curation, formal analysis, investigation, writing-original draft, writing-review and editing. B. Bhatnagar: Resources, data curation, formal analysis, writing-review and editing. A.S. Mims: Resources, data curation, methodology, writing-review and editing. S. Orwick: Resources, data curation, formal analysis, writing-review and editing. T.L. Lin: Resources, data curation, formal analysis, methodology, writing-review and editing. H.E. Crosswell: Resources, data curation, formal analysis, methodology, writing-review and editing. D. Zhang: Resources, data curation, formal analysis, writing-original draft. M.D. Minden: Resources, data curation, methodology, writing-review and editing. V. Munugalavadla: Resources, data curation, formal analysis, methodology, writing-review and editing. L. Long: Writing-review and editing. J. Liu: Data curation, formal analysis, writing-review and editing. Y. Pan: Conceptualization, resources, data curation, formal analysis, methodology, writing-review and editing. T. Oellerich: Conceptualization, resources, data curation, formal analysis, writing-review and editing. H. Serve: Conceptualization, resources, methodology, writing-review and editing. A.V. Rao: Conceptualization, data curation, formal analysis, methodology, writing-original draft, writing-review and editing. W. Blum: Conceptualization, resources, data curation, formal analysis, methodology, writing-original draft, writing-review and editing.

We extend our thanks to the patients and their families who participated in this study and to the investigators and coordinators at the clinical sites. We acknowledge Esteban (Steve) Abella, MD, and A. Mario Marcondes, MD, PhD, for their contributions to the study design and conduct. We thank Beth Sesler, PhD, CMPP, of Impact Communication Partners for editorial assistance in preparing the article, with financial support provided by Gilead Sciences, Inc. The study was funded by Gilead Sciences, Inc. J.C. Byrd was supported by R35 CA197734.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.