Trial eligibility, treatment patterns, and outcome for venetoclax-based therapy in AML: a prospective cohort study

Venetoclax (Ven)-containing therapies have recently become a standard of care for older patients with newly diagnosed (ND) acute myeloid leukemia (AML) and for those deemed ineligible for intensive induction therapy.¹ In the phase 3 VIALE-A clinical trial, the addition of Ven to azacytidine (Aza) was associated with significantly prolonged survival, as well as significant increase in remission and transfusion independence (TI) rates compared with treatment with Aza alone.^2,3 In the long-term follow-up of VIALE-A, the median overall survival (mOS) with Ven in combination with Aza was 14.7 months (95% confidence interval [CI], 12.1-18.7) as compared with 9.6 months (95% CI, 7.4-12.7) in the Aza-monotherapy arm (hazard ratio [HR], 0.58 [95% CI, 0.47-0.72]; P < .001).⁴ 

Previous studies have demonstrated that use of novel therapeutic approaches in routine clinical practice significantly differ from those tested in the registrational clinical trials leading to their approval.^5,6 Retrospective studies from different countries and health systems are therefore complementary to the data obtained in clinical trials and often vary in terms of patterns of patient selection, treatment exposure, and postremission management.^7-11 Better understanding of the factors that influence outcome in the real world may help in choosing patients for Ven-containing therapy and for optimal management of patients receiving this regimen in the clinic.

In Israel, Ven + hypomethylating agent (Ven-HMA) and Ven + low-dose cytarabine therapy were approved for use within the National Health Insurance service in parallel to the approval of the US Food and Drug Administration in January 2019 and are fully reimbursed for intensive chemotherapy–ineligible patients with ND AML. With the understanding that patterns of patient selection, treatment dosing and management, as well as postremission treatment management may differ from those applied in the context of a registration clinical trial, we initiated a real-world, prospective observational multicenter trial with the aim to analyze Ven-containing therapy in clinical practice.

This is a prospective observational study conducted at 12 sites across Israel, between August 2019 and December 2022, that enrolled patients with ND-AML (REVIVE study). The study was conducted in accordance with the principles outlined in the Declaration of Helsinki and approved by the institutional review board in all 12 centers. All participating patients provided written informed consent before study participation. The study included adults (age of ≥18 years) with ND-AML according to the World Health Organization 2016 classification.¹² An inclusive enrollment approach was applied, and all ND patients treated with Ven-containing therapy were eligible for inclusion, excluding only patients that participated in an interventional clinical trial within 30 days before Ven-containing treatment initiation. The decision to treat a patient with Ven-containing therapy was at the discretion of the treating physician and therapy was administered according to routine clinical practice in Israel and according to the local label. Patients receiving any AML therapy other than Ven-containing lower-intensity therapy were excluded from this study. Antifungal and antibacterial prophylaxis or treatments were given at physician discretion and per local protocol in each medical center. Granulocyte colony-stimulating factor was also given at physician discretion and per local protocol in each medical center.

The primary end point was OS. Secondary end points included composite complete remission (CRc; complete remission (CR) or complete remission with incomplete hematologic recovery [CRi]), red blood cell [RBC] and platelet TI, treatment pattern, exposure, and safety profile. Exploratory end points included event-free survival (EFS) and subgroup analyses of CRc and OS.

OS was defined as time from treatment initiation to death from any cause; EFS was defined as time elapsed from diagnosis to disease progression, treatment failure (failure to achieve CR or <5% bone marrow blasts after at least 6 cycles of treatment), or death.² Response assessment was done as per local standard. CRc, CR, CRi, morphological leukemia-free state (MLFS), and partial response (PR) were defined according to the previously published criteria outlined in the European LeukemiaNET (ELN) criteria.¹³ ELN 2017 cytogenetic risk group¹³ was determined at physician discretion, based on available cytogenetic and molecular data. TI was defined as the absence of a red cell (RBC and whole blood) or platelet transfusion for at least 56 days (8 weeks) between the first and last day of treatment.² Patients who discontinued Ven-containing treatment were followed up for survival.

The data cutoff date was 30 June 2023. For the primary end point analysis, OS, a conservative estimation based on previous publications of 50% at 12 months was assumed.² A sample size of 100 patients enabled us to estimate the OS rate at 12 months with a 95% CI with a precision of ±10%. The intention-to-treat population included all 207 patients who were treated with at least 1 dose of Ven. Efficacy and safety analyses were performed in the intention-to-treat population. The distribution of OS was estimated for each treatment group with the use of the Kaplan-Meier method and compared with the use of the log-rank test. Survival analysis was censored at the time of allogeneic stem cell transplant. The HR between the treatment groups was estimated with the Cox proportional hazards model.

A total of 209 ND patients were enrolled in the REVIVE clinical trial; 207 patients initiated treatment with Ven-containing therapy and were included in the intention-to-treat efficacy and safety analysis; 2 patients were not included in the analysis because they did not receive therapy (1 patient died within a week of enrollment, before therapy initiation, and 1 patient was ineligible for study; Figure 1). The study patient population largely mirrored the patient population expected to be treated in the clinic (Table 1). The median age at diagnosis was 75 years (range, 43-94; interquartile range, 71-81), with 54.6% of patients aged ≥75 years at diagnosis. One-hundred and eleven patients (53.7%) were diagnosed with secondary AML including AML with prior myelodysplastic syndrome (MDS) or MDS-related cytogenetic changes (AML arising from MDS, 46.8%), therapy-related AML (t-AML, 28.8%), and AML arising from a previous myeloproliferative neoplasia (blast-phase MPN, 24.3%). Baseline characteristics of the different subgroups are described in supplemental Table 1. Of the 52 AML arising from MDS, 13 (25%) received prior Aza therapy (Table 1).

Favorable, intermediate, and poor ELN 2017 disease risk¹³ were documented in 9.7%, 38.2%, and 35.3% of patients, respectively. Normal karyotype and complex karyotype were reported in 31.4% and 15.5% of patients, respectively. Mutations in FLT3 and NPM1 were found in 10.4% and 16.0% of patients tested, respectively (Table 1).

Grade 3/4 neutropenia, anemia, and thrombocytopenia at diagnosis were reported in 61.4%, 19.8%, and 41.1%, respectively. Requirements of at least 1 RBC or platelet transfusion within 8 weeks preceding diagnosis were noted in 38.2% and 15.5% of cases, respectively (Table 1).

The median time from diagnosis to Ven-containing treatment initiation was 7.5 days (range, 1-83), with 49.8% initiating therapy after >7 days and 22.7% after >14 days from diagnosis.

Per study protocol, the treating physician documented after diagnosis the main reason for choosing to treat the patient with a Ven-containing treatment approach rather than with any other therapeutic intervention. Age of ≥75 years was the main reason for choosing Ven-containing treatment, accounting for 42.5% of cases, Eastern Cooperative Oncology Group (ECOG) performance status (PS) of ≥2 accounted for 15.9% of cases, and patient comorbidities were reported as the main reason for treatment choice in 17.8% of cases. In 22.2% of the patients in this study, the main reason for choosing Ven-containing therapy was reported as “disease-risk,” defined as factors related to disease biology (such as genetics, cytogenetics, or predicted resistance) (Table 1).

Overall, 202 patients (98.1%) initiated Ven in combination with a hypomethylating agent (Ven-HMA), of them, 196 patients (94.7%) received Ven in combination with Aza (Ven-Aza) and 6 (2.9%) patients initiated Ven with decitabine. Only 2 (1.0%) patients received Ven in combination with low-dose cytarabine, and 3 patients (1.4%) received Ven monotherapy. Treatment was administered in a tertiary cancer center for 71.5% of patients and 68 patients (32.8%) initiated therapy in an outpatient setting (defined as patients who were not admitted for ramp-up dosing). Ramp-up to the target dosing of 400 mg was accomplished in 91.6% of patients, after a median of 3 days.

Bone marrow assessment at diagnosis was reported in 96.6% of patients. Response marrow assessment was done at the end of cycle 1 in 64.7% of patients, at end of cycle 3 in 53% of patients, and at end of cycle 6 in 38.6% of patients (representing multiple assessments for some patients). Full description of bone marrow assessment method and results are described in supplemental Table 2.

Almost one-half of the patient population had at least 1 documented tumor lysis syndrome (TLS) risk factor at baseline (supplemental Figure 1). TLS prophylaxis with hypouricemic agents and hydration was administered in 77.8% of patients whereas a minority of patients received only hydration (9.2%) or hypouricemic agent (5.3%). Of patients with ≥25 × 10⁹ cells per μL at presentation (n = 26 patients), cytoreduction with hydroxyurea was used in 73.1% of cases.

Antibacterial prophylaxis was administered to 27% of patients and comprised ciprofloxacin (n = 51), trimethoprim-sulfamethoxazole (n = 4), or piperacillin (n = 1). Antifungal prophylaxis was administered in 19.7% of patients and comprised systemic fluconazole (n = 37) or voriconazole (n = 6).

Administration of granulocyte colony-stimulating factor support during remission cycles was administered in 26.6% of patients.

Treatment exposure was analyzed only in patients who received Ven-Aza combination and had complete information for cycle 1 and 2 of both Ven and Aza exposure days in these cycles (Figure 2). A total of 146 (70.5%) patients had sufficient data to be included in this analysis. The median number of cycles administered was 3 (range, 1-30). Cycle was defined as the start of a 28-day cycle until the start of the next cycle (including dose holds) and was calculated based on the interval between Aza treatments. The median cycle length, defined as the start of a cycle until the start of the next cycle (including dose holds), was 35 days, and the median of Aza treatment was 7 days. The median duration of Ven therapy per cycle was 25 days and was highest during the first 3 cycles (median of 28 days), decreasing thereafter over time to a median of 14 days on cycle 15 (Figure 2).

With a median follow-up of 22.5 months (range, 0.1-43), the mOS for the entire cohort was 11.7 months (95% CI, 9.9-15.4) (Figure 3A). Overall response rate was documented in 75.4% of patients and CRc was documented in 65.2% of patients (CR and CRi in 44.4% and 20.8%, respectively; Figure 3B). The median time to first response (CR/CRi/MLFS/PR) was 35 days (95% CI, 32-39) (Figure 3C). In responding patients (attaining an overall response), the median duration of remission was 12.8 months (95% CI, 10.2-19.3) (Figure 3D). The median time for EFS was 3.86 months (95% CI, 3.05-5.34) (Figure 3E).

Survival analysis showed a significant correlation with ELN 2017 risk (P = .01). CRc rates were higher than 50% across the 3 ELN risk groups, with the highest rates observed in favorable ELN risk patients (supplemental Figure 2).

AML ontogeny significantly impacted OS of patients (P = .024). The mOS of de novo AML was highest, with a mOS of 16.6 months, (95% CI, 9.9-22.3), followed by (t-AML; mOS, 14.9 months; 95% CI, 7.3 to not reached [NR]), AML arising from MDS (mOS, 9.9 months; 95% CI, 7.6-15.5) and AML arising from myeloproliferative neoplasm (MPN; mOS, 8.8 months; 95% CI, 5.7-10.7; Figure 4A). CR/CRi rates were lower for patients with AML arising from MDS (57.7%) and patients with AML arising from MPN (59.2%) than de novo AML (69.9%) and t-AML (68.8%; Figure 4B). AML ontogeny significantly affected duration of remission (P = .025). Median duration of remission for de novo AML was 16.23 months (95% CI, 10.6-27.2), 5.9 months for AML arising from MPN (95% CI, 3.8-10.8), 7.6 months for patients with AML arising from MDS (95% CI, 4.6-19.7), and NR for patients with t-AML (95% CI, 4.5 to NR; Figure 4C).

Body mass index (BMI) emerged as a factor associated with OS. mOS for patients with a BMI of <18.5, 18.5-25, 25-30, and >30 was significantly different (P = .0015; supplemental Figure 3); with every increase in 1 BMI unit, the risk for death decreased in a ratio of 0.956 (P = .0298).

Modified Charlson Comorbidity Index, in which leukemia is excluded from the model,¹⁴ was significantly correlated with OS (P = .0017; supplemental Figure 4).

No statistically significant differences in OS were observed when patients were stratified by prior exposure to HMAs, age (≥75 or < 75 years), sex, main reason patient not eligible for induction chemotherapy, ECOG PS, karyotype abnormalities, days elapsed from diagnosis to treatment initiation, and molecular subgroups (supplemental Figures 5-12).

Multivariant Cox regression analysis confirmed an independent association of OS and AML arising from prior MPN or MDS (MPN vs de novo AML: HR, 1.72; 95% CI, 1.0-2.96; MDS vs de novo AML: HR, 1.56; 95% CI, 1.02-2.4; P = .01), modified Charlson score (≥5 vs 0-2: HR, 2.0; 95% CI, 1.19-3.38; P = .03) and ELN disease risk at baseline (poor vs favorable: HR, 2.16; 95% CI, 1.1-4.23; P = .02; supplemental Figure 13). A trend toward significance was also observed for age (HR, 1.13; 95% CI, 1.0-1.27; P = .05) and BMI (HR, 0.81; 95% CI, 0.64-1.02; P = .08; supplemental Figure 13).

Analysis of EFS demonstrated that it is significantly correlated with ELN disease risk (P < .0001; supplemental Figure 2), monosomal and complex karyotype (P = .016; supplemental Figure 10), and BMI (P = .036; supplemental Figure 3).

The changing patterns of patient selection and treatment management over time were also assessed. Patients were divided into 4 quartiles (Q1-Q4) reflecting different time frames of treatment initiation over the study accrual span. A nonsignificant trend toward better OS for later-enrolled patients was demonstrated when comparing all 4 quartiles (P = .12; supplemental Figure 14A). When stratifying patients into 2 groups of early enrollment (Q1 + Q2) vs later enrollment (Q3 + Q4), survival was significantly better for patients enrolled later in the course as the trial progressed (Q1 + Q2: mOS, 9.9 months; Q2 + Q3: mOS, 15.5 months; P = .026).

Patient baseline characteristics as well as disease characteristics and remission rates did not trend differently between the 4 quartiles. The reasons for the increment in survival over time is not yet clear and may suggest a potential role for treatment management in responding patients on outcome (supplemental Figure 14C; supplemental Table 3). For example, although not statistically powered for evaluation, days of Ven exposure per cycle were numerically lower in Q3 + Q4 than Q1 + Q2 (supplemental Figure 14D).

The median number of RBC, platelet, and both RBC and platelet transfusions during the first 6 months of Ven-containing therapy were 4, 5, and 8 transfusions, respectively. Eight-week TI was achieved by 43 (48.9%) patients within a median time of 82 days (95% CI, 70-99). Of those, 38 (46.9%) had RBC TI within median time of 75 days (95% CI, 67-92), and 17 (53.1%) patients had platelet TI within median time of 64 days (95% CI, 59-75). TI of both platelet and RBC was achieved in 44% of patients with a median time of 103 days (95% CI, 62-152). The median duration of 8-week TI was 112 days (95% CI, 68-169) for any transfusion type, 124 days for RBC transfusions (95% CI, 56-169), 142 days (95% CI, 43-460) for platelet transfusions, and 80 days (95% CI, 12-190) for both RBC and platelet transfusions. Median time to last transfusion, defined by the time elapse between treatment start date and last transfusion date, was 14.5, 7.0, and 43 days for RBC, platelets, or both, respectively.

Thirty-three patients proceeded to hematopoietic stem cell transplant (HSCT; 15.9% of patients) that represented 21.1% of responding patients. Of these patients, 13 patients (39.4%) were in CR at time of HSCT, 16 (48.5%) in CRi, 1 patient (3.0%) had MLFS, and 1 patient each with PR, refractory disease, or missing information.

The median age of patients who received HSCT was 69.0 years (range, 43-77). Most patients had adverse ELN risk (51.5%) and an ECOG PS of ≤1 (supplemental Table 4). The median time from treatment initiation to transplant was 108 (range, 58-407) days. Reduced-intensity conditioning was applied in 15 (45.5%) patients, nonmyeloablative conditioning in 10 (30.3%), and a myeloablative conditioning was administered in 6 (18.2%) patients. mOS from transplant was 35.2 months (95% CI, 8.26 to NR; supplemental Figure 15).

Assessment of minimal residual disease (MRD) was not mandatory and was reported for a minority of patients in this study. MRD at the end of cycle 3 were available for 24 patients (18 by flow, 5 by polymerase chain reaction, and 1 by next-generation sequencing) and were negative in 10 patients (41.7%). At the end of cycle 6, MRD was available for 15 patients (11 by flow and 4 by polymerase chain reaction) and were reported to be negative in 6 patients (40%). Although these results are encouragingly similar to those reported by Pratz et al for the VIALE-A study,³ patient numbers were not sufficient for statistical analysis.

In the attempt to better understand patterns of patient selection and their relation to outcome in the real-world setting, we analyzed separately patients that would not have been eligible for the VIALE-A registration trial. In total, 90 patients (43.5%) were designated ineligible for the registration clinical trial based on the original exclusion criteria used in the VIALE-A study protocol² (referred to hereafter as “trial-ineligible”). These included patients who were trial ineligible because of disease ontogeny (prior MPN, n = 27), prior HMA exposure (n = 13), concomitant active malignancy (n = 23), ECOG PS of 4 (n = 3), isolated extramedullary disease (n = 2), and severe kidney or hepatic disturbances (n = 13 and n = 15, respectively). One patient had favorable risk core-binding factor genetics and was also considered trial-ineligible.

The CR/CRi rate in the trial-ineligible group (n = 90) was 60%, which translated into a mOS of 10.7 months (95% CI, 7.9-11.8) whereas trial-eligible patients (n = 117) demonstrated higher remission rates (69.2%) and longer mOS of 17.8 months (95% CI, 9.9-23.4; P = .027; Figure 5A-B). Although CR/CRi rates were numerically higher in the trial-eligible group than for trial-ineligible patients, this did not achieve statistical significance (P = .17; χ² test).

When looking at the mOS, stratified by the ineligibility criteria, kidney function, and prior MPN, showed a significant association with survival (Figure 5C). Isolated extramedullary disease and ECOG PS of 4 also showed a significant association, although patient numbers were low (n = 2 and n = 3, respectively). When assessing the effect of the cumulative number of ineligibility criteria on OS, a nonsignificant trend toward shorter survival was observed in patients that had >1 cumulative ineligibility criteria as compared with patients with only 1 exclusion criteria (mOS of 11.34 months vs 8.23 months; P value = .163).

Safety analysis included 207 patients. Mortality at 30 and 60 days was 4.8% and 14%, respectively. Overall, 85.5% of patients had at least 1 adverse event (AE); 60.4% of patients had a serious AE. Common AEs are summarized in Table 2. The most frequently reported hematologic AE of grade ≥3 included neutropenia (25.1%), febrile neutropenia (28.5%), and thrombocytopenia (7.3%). Gastrointestinal AEs of any grade were common and included diarrhea (14.5%) and constipation (14%). TLS (as defined by the Cairo criteria¹⁵) occurred in 4.8% of patients, with 2.4% of cases recorded as grade ≥3. Notable serious AEs were neutropenia and febrile neutropenia (15.9% and 20.8%, respectively), pneumonia (7.2%), sepsis (5.8%), and pyrexia (9.7%).

Treatment discontinuation because of AEs occurred in 20.8% of patients (43 patients) and were primarily because of febrile neutropenia and neutropenia (3.4% each).

Dose interruption occurred in 65.3% of patients at least once, 81.1% of patients had dose discontinuation and 32.1% of patients had at least 1 dose modification. The median number of dose interruption/discontinuation per patient was 2 (range, 1-30) and the median number of dose modifications (as defined by the local label) per patient was 1 (range, 1-10).

The REVIVE study was designed as a prospective observational trial, recruiting patients from several sites in the context of an inclusive and global reimbursement system. This enabled us to better dissect the patterns of patient selection for therapy and outcome in varying populations. Indeed, patients in this cohort were different from those enrolled to the VIALE-A trial and may mirror the patient population encountered in a real-world clinical practice. Of patients in the REVIVE cohort, >50% had secondary AML including patients that would be excluded from the registration clinical trial such as those previously exposed to HMAs (25% of patients with history of MDS in the current cohort) and patients that presented with MPN in blast-crisis (almost one-quarter of patients with secondary AML).

Several previous retrospective analyses reported on the efficacy of Ven-containing combinations in a real-world setting.^7-11 In our prospective data set, OS for the whole cohort was 11.7 months. Because OS rate may be related to patterns of patient selection to therapy, we focused on patients that would not be eligible for the VIALE-A. In the current analysis, 43.5% of patients would have not been eligible for the VIALE-A trial based on the original exclusion criteria.² This trial-ineligible group was characterized by lower remission rates and shorter survival (median survival, 10.72 months). Conversely, VIALE-A trial-eligible patients in our study had longer OS (median survival, 17.8 months). Our observations (Figure 5) are in line with previous retrospective analyses, suggesting worse outcome in Ven-Aza–treated patients with MPN blast crisis.^16,17 With respect to recent reports,¹⁸ we were able to show a statistically significant association between ELN 2017 criteria and OS in Ven-based treated patients, although we also noted it does not provide clinically meaningful stratification for favorable and intermediate-risk groups.¹⁸ Co-occurrence of somatic alterations was recently shown to further predict patient outcome. In that regard, mutations in TP53 were shown to drive the poor response and prognosis of patients with complex karyotype treated with Ven-Aza.¹⁹ Further analysis to investigate and validate the association of somatic genetic alterations with outcomes as previously described is planned.

BMI was previously studied in younger patients with acute leukemia. Higher BMI has traditionally been linked to increased mortality in patients with AML and acute lymphoblastic leukemia.^20,21 Interestingly our analysis demonstrated an inverse correlation with better survival for patients with a higher BMI, an advantage that retained borderline significance in the multivariate analysis (P = .073; supplemental Figure 16).

These findings suggest that the effect of BMI on outcome in AML may be context dependent and age related. The survival advantage of older patients with AML and higher BMI treated with Ven-containing therapy has, to our knowledge, not been described previously and is in line with recent reports correlating higher BMI with better survival in older adults (“the obesity paradox”).²² 

The importance of balancing treatment exposure and toxicity, as previously reported,²³ is further highlighted in this prospective trial. A time-dependent analysis demonstrated a consistent improvement in survival over time, although the response rates were unchanged. This significant difference in survival cannot be explained by baseline differences in patient characteristics, disease biology, or the use of HSCT over time, and may potentially reflect increasing experience in patient management on a national level. A recent retrospective analysis from a large US registry demonstrated that patients that were treated with Ven-containing lower-intensity regimens (after VIALE-A publication) were more likely to have an earlier marrow response assessment and more liberal dose modifications because of cytopenia.²³ Analysis of the patterns of cycle administration and dose reductions in our study are generally aligned with the recent publication on the long-term follow-up data from the VIALE-A⁴ showing a median cycle length of 34 days, defined as the start of a cycle until the start of the next cycle (including dose holds), with a reduction in the duration of Ven treatment days per cycle over time.

Retrospective studies demonstrate the utility of allogeneic HSCT after Ven-containing induction and suggest comparable outcome with that achieved after intensive remission induction.^24-30 In this analysis just over 20% of patients received HSCT in response. These data reflect the growing use of HSCT as a postremission therapeutic approach with curative intent after Ven-Aza therapy, even in patients not originally eligible for intensive chemotherapy. These patients were younger and fitter and demonstrated an excellent mOS from transplantation of 35.2 months. Although this observational study was not designed to assess the utility of HSCT as compared with ongoing Ven-Aza therapy, it does highlight current patterns of postremission therapy and the need for better controlled data to assess the role of HSCT consolidation after Ven-Aza therapy.^24,28 

A limitation of this analysis is the lack of more extensive annotated genetic data that may affect risk ascription, as well as lack of analyzable data regarding MRD that precluded the incorporation of these factors to the analysis. Although these data gaps are significant, they do reflect the routine practice in many centers in evaluation and treatment of patients with Ven-containing therapy.

In conclusion, this large prospective observational study demonstrates the effectiveness of Ven-containing therapy in the real world and identifies unique patterns of patient selection, therapy, and supportive care that may inform the clinical approach to treatment with Ven-containing combinations in the clinic.

The authors thank the patients, their families, and the physicians who were involved in this study. The authors acknowledge Michal Klein and Lea Price from BioForum LTD for assistance with statistical analysis. The authors also acknowledge the contribution of Stav Vilkovski in coordinating data management in the different participating centers, and all the clinical research coordinators from the centers: Lior Gadot, Tali Fialkov, Dafna Laufer, Anna Turkot, Anastasya Vernik, Ayala Yaron, Andrea Shoukair, Hadas Bar, Maya Shaked Rabi, Oli Berkov, and Tehila Azualos.

This study was sponsored by AbbVie Israel.

Contribution: O.W., Y.M., N.F., and M.G. designed the study; O.W., I.L., B.N., S.T., I.A., J.C., Y.O., C.G., T.Z., D.O., I.H., T.T., N.D., G.S., and Y.M. collected and analyzed data; O.W., A.J.B., and Y.M. wrote the manuscript; and all authors critically revised the manuscript.

Conflict-of-interest disclosure: O.W. reports research support from AbbVie; speaker honoraria from AbbVie, Astellas, and Novartis; and an advisory role with AbbVie, Astellas, Novartis, Pfizer, Medison, and Teva. Y.M. reports an advisory role with AbbVie, Astellas, Bristol Myers Squibb, Gilead, Medison, Neopharm, Novartis, and Pfizer; and honoraria from and consultancy with AbbVie, Astellas, Medison, Novartis, and Takeda. B.N. reports speaker honoraria from and consultancy with AbbVie. Y.O., S.T., and G.S. report consultancy with AbbVie. T.T. reports research support from AbbVie. J.C. reports speaker honoraria from and consultancy with AbbVie. M.G., A.J.B., N.F., J.B., A.B., M.B., and R.C. report employment with AbbVie, and may hold stock or other options. The remaining authors declare no competing financial interests.

A complete list of the members of the Israeli Acute Leukemia Working Group appears in “Appendix.”

Correspondence: Ofir Wolach, Institute of Hematology, Davidoff Cancer Center, Beilinson Hospital, Rabin Medical Center, Ze'ev Jabotinsky Rd 39, Petah Tikva 4941492, Israel; email: owolach@gmail.com.

The members of the Israeli Acute Leukemia Working Group are: Ofir Wolach, Itai Levi, Boaz Nachmias, Sigal Tavor, Irina Amitai, Yishai Ofran, Chezi Ganzel, Tsila Zuckerman, Doaa Okasha, Ilana Hellmann, Tamar Tadmor, Najib Dally, Jonathan Canaani, Galia Stemer, Yakir Moshe, Jacob Rowe, Pia Raanani, Arie Apel, Shlomzion Aumann, May Basood, Shlomo Bulvik, Avraham Frisch, Sharon Gino-Moor, Alexander Gural, Maya Koren-Michowitz, Baher Krayem, Alon Rozental, Vladimir Vainstein, Itay Zilbershatz, Tzvika Porgas, Adrian Duek, Alex Kolomansky, Miriam Neaman, Moshe Mittelman, Drorit Merkel, Maya Koren-Michowitz, Ariel Aviv, Eitan Kugler, Shai Shimony, and Meira Yisraeli Salman.