A Phase II Study of Decitabine, Venetoclax and Ponatinib for Advanced Phase Chronic Myeloid Leukemia and Philadelphia Chromosome-Positive Acute Myeloid Leukemia

Abstract

Background:

Advanced phase Philadelphia chromosome-positive (Ph+) myeloid disease—consisting of chronic myeloid leukemia (CML) in myeloid blast phase (BP), CML in accelerated phase (AP), and Ph+ acute myeloid leukemia (AML)—is associated with poor outcomes. While previous studies have suggested the benefit of chemotherapy and BCR::ABL1 tyrosine kinase inhibitor (TKI) combinations, the optimal regimen is uncertain and prospective studies for this rare groups of diseases are scant. Preclinical and retrospective clinical data suggest possible synergy between the BCL-2 inhibitor venetoclax and BCR::ABL1 TKIs. We therefore designed a study to evaluate the safety and efficacy of a novel combination of decitabine, venetoclax and the third-generation BCR::ABL1 TKI ponatinib in advanced phase Ph+ myeloid diseases.

Methods:

For this phase II study, patients ≥18 years of age with previously untreated or relapsed/refractory myeloid CML-BP, CML-AP or Ph+ AML and an Eastern Cooperative Oncology Group performance status 0–3 were eligible. Patients were eligible regardless of the number of previous lines of therapy received or prior receipt of ponatinib. Cycle 1 (induction) consisted of a 7-day lead-in of ponatinib 45mg orally daily (days 1–7), followed by combination therapy with decitabine 20 mg/m² intravenously (IV) on days 8–12, venetoclax orally daily with ramp-up to a maximum dose of 400mg on day 8–28, and ponatinib 45mg orally daily on days 8–28. Cycles 2–24 consisted of decitabine 20 mg/m² IV on days 1–5, venetoclax orally 400mg on days 1–21, and ponatinib orally daily on days 1–28. Response-based dosing of ponatinib was implemented in consolidation cycles, with reduction to 30mg daily in patients who achieved complete remission (CR) or complete remission with incomplete hematologic recovery (CRi) and reduction to 15mg daily in patients with undetectable BCR::ABL1 transcripts. The primary endpoint was the composite rate of CR/CRi in the intention-to-treat population. This trial was registered with ClinicalTrials.gov (NCT04188405) and has completed enrollment.

Results:

Between July 12, 2020 and July 8, 2023, 20 patients were treated (14 with CML-BP, 4 with CML-AP and 2 with Ph+ AML). The median age was 42 years (IQR, 32–58 years). Twelve patients (60%) had received ≥2 prior BCR::ABL1 TKIs, and 14 patients (70%) had at least one high-risk additional chromosomal abnormality and/or complex karyotype. The median duration of follow-up was 21·2 months (IQR, 14·1–24·2 months). The CR/CRi rate was 50% (CR in 5%, CRi in 45%); an additional 6 patients (30%) achieved a morphologic leukemia-free state. The most common grade 3–4 non-hematologic adverse events were febrile neutropenia in 8 patients (40%), infection in 7 patients (30%) and alanine or aspartate transaminase elevation in 5 patients (25%). Eight patients (40%) experienced at least one cardiovascular event of any grade. There were 3 on-study deaths, none of which was considered related to the study treatment and all from infections in the setting of refractory leukemia.

Interpretation:

The combination of decitabine, venetoclax and ponatinib is safe and shows promising activity in patients with advanced phase CML, including those with multiple prior therapies and/or high-risk disease features. Further studies evaluating chemotherapy and venetoclax-based combination strategies using newer-generation BCR::ABL1 TKIs are warranted.

Funding:

Takeda Oncology

Introduction

Most patients with chronic myeloid leukemia (CML) are diagnosed in the chronic phase of the disease and have a normal life expectancy in the era of BCR::ABL1 tyrosine kinase inhibitors (TKIs) (1, 2). However, approximately 5% of patients present with advanced phase CML—either blast phase (BP) or accelerated phase (AP)—and another 5% will eventually progress to advanced phase CML (3). Advanced phase CML is associated with a poor prognosis, with patients with CML-BP having an expected median overall survival (OS) less than 1 year (4). Outcomes are particularly poor for patients who have progressed on prior TKI therapy, those with high-risk additional chromosomal abnormalities (ACAs) and those with CML-BP with a myeloid immunophenotype (5).

There is no clear standard of care for patients with advanced phase CML. While some patients with de novo CML-AP can have good long-term outcomes with TKI monotherapy if they achieve optimal molecular response milestones, TKI monotherapy results in suboptimal responses for patients with previously treated CML-AP and those with CML-BP (4, 6). Retrospective studies suggest benefit of combination therapy with either intensive chemotherapy or a hypomethylating agent in combination with a BCR::ABL1 TKI, although prospective studies supporting a specific regimen in advanced phase CML are scant (7). Consensus guidelines recommend that patients with myeloid CML-BP should receive acute myeloid leukemia (AML)-type induction chemotherapy in combination with a BCR::ABL1 TKI, although recommendations for a particular regimen or TKI are lacking. Notably, most of prospective combination studies in advanced phase CML used first- or second-generation TKIs, which are likely not suitable options for patients after failure of earlier-generation TKIs (8–11).

Ponatinib is a potent third-generation BCR::ABL1 TKI that is active against most ABL1 kinase domain mutations, including T315I, which confers resistance to first- and second-generation TKIs (12, 13). Ponatinib-based combination therapies have resulted in promising outcomes in Philadelphia chromosome-positive (Ph+) acute lymphoblastic leukemia (ALL) but have not been widely studied in advanced phase CML (14, 15). The MATCHPOINT study evaluated the combination of FLAG-Ida (fludarabine, cytarabine, granulocyte-stimulating factor, and idarubicin) in combination with ponatinib in 16 evaluable patients with CML-BP (16). Eleven patients (69%) achieved second chronic phase CML, 12 patients (71%) were bridged to allogeneic hematopoietic stem cell transplant (HSCT), and the median OS was 12 months. Notably, this study was enriched with younger patients (median age: 33 years) with de novo CML-BP (59% of the cohort) and enrolled patients with myeloid, lymphoid, or mixed immunophenotype.

Preclinical and retrospective clinical data suggest synergy between BCR::ABL1 TKIs and the BCL-2 inhibitor venetoclax although, to our knowledge, no prospective studies have evaluated this combination in advanced phase CML (17, 18). Given the established potency of ponatinib and to exploit the potential synergy of BCR::ABL1 inhibition with venetoclax, we designed a phase II study to evaluate the triplet combination of decitabine, venetoclax and ponatinib in patients with advanced phase CML or Ph+ AML. By using a backbone regimen of decitabine plus venetoclax, which is widely used for the treatment of older adults unsuitable for intensive chemotherapy (19), we sought to evaluate a novel ponatinib-based regimen that could be administered to patients of all ages, regardless of fitness.

Methods

Study design and participants

This was a single-center, phase 2 study to assess the efficacy and safety of the combination of decitabine, venetoclax and ponatinib in adult patients with CML-BP with myeloid immunophenotype, CML-AP or Ph+ AML. The 2016 World Health Organization (WHO) criteria were used for diagnosis of CML-BP, CML-AP, and Ph+ AML (20). Additional eligible criteria included: Eastern Cooperative Oncology Group (ECOG) performance status 0–3, total bilirubin ≤2 × upper limit of normal (ULN), alanine aminotransferase (ALT) and aspartate aminotransferase (AST) ≤1·5 × ULN, serum amylase and lipase ≤1·5 × ULN, and creatinine clearance ≥30 mL/minutes. Key exclusion criteria included: history of acute pancreatitis within 6 months of study or history of chronic pancreatitis, uncontrolled hypertriglyceridemia, grade III-IV cardiac failure, and clinically significant and uncontrolled cardiovascular disease. There was no restriction on prior therapies. Patients with previously untreated or relapsed/refractory disease were eligible, and notably, patients with prior ponatinib exposure were not excluded. Full inclusion and exclusion criteria are available in the protocol (see Supplemental file). This study was conducted at a single academic center (The University of Texas MD Anderson Cancer Center [MDACC]). This study was approved by the Institutional Review Board of MDACC. All patients provided written informed consent according to institutional guidelines and the study was conducted according to the Declaration of Helsinki. This study was registered at ClinicalTrials.gov (NCT04188405).

Procedures

For patients without ponatinib exposure within the preceding 2 weeks, cycle 1 consisted of a 7-day lead-in of ponatinib 45mg orally daily (days 1–7) to allow for cytoreduction, followed by combination therapy with decitabine 20 mg/m² intravenously (IV) on days 8–12 (5 days total), venetoclax orally daily with ramp-up to a maximum dose of 400mg on day 8–28 (21 days total), and ponatinib 45mg orally daily (given continuously). A bone marrow (BM) examination was performed on cycle 1, day 21 and venetoclax and/or ponatinib could be held if BM blasts were <5% or if aplastic. For patients with recent ponatinib exposure, the 7-day lead-in of ponatinib was omitted. For these patients, cycle 1 was 28 days in length and consisted of decitabine 20 mg/m² IV on days 1–5, venetoclax orally daily with ramp-up to a maximum dose of 400mg on day 1–21, and ponatinib 45mg orally daily on days 1–28. All patients were admitted during the venetoclax ramp-up period (i.e. 3 days) during cycle 1; the rest of the protocol therapy was given in the outpatient setting. Ponatinib was recommended to be given continuously, regardless of the results of interim BM results for most patients. For cycles 2–24, patients received decitabine 20 mg/m² IV on days 1–5, venetoclax orally 400mg on days 1–21, and ponatinib orally daily on days 1–28. Bone marrow assessments were performed at the end of cycle 1 and then every 2–3 cycles as clinically indicated. Response-based dosing of ponatinib was implemented in consolidation cycles, as in previous studies of response-based dose reduction of ponatinib (14, 15, 21). Patients who had not yet achieved CR/CRi received 45mg daily, those who had achieved CR/CRi but still had detectable PCR for BCR:ABL1 received 30mg daily, and those who achieved CR/CRi with undetectable PCR for BCR::ABL1 received 15mg daily. Consolidation cycles were anticipated to be 28 days in length, although cycle delays were allowed due to delayed count recovery or intercurrent illness. The venetoclax dose was reduced by 50% for patients receiving a moderate CYP3A4 inhibitor and by 75% for patients receiving a strong CYP3A4 inhibitor (with the exception of posaconazole for which venetoclax was reduced by 83%, e.g., from 400mg to 70mg daily). The ponatinib dose was not adjusted based on concomitant use of moderate or strong CYP3A4 inhibitors. Patients were recommended to receive low-dose statin and aspirin 81mg daily for platelet count above 50 × 10⁹/L as per our institutional standard of care for patients on ponatinib. To prevent central nervous system relapses, 4 doses of intrathecal prophylactic cytarabine were recommended for all patients.

Outcomes

The primary endpoint was the composite rate of complete remission (CR) and CR with incomplete hematologic recovery (CRi) within 2 cycles of protocol therapy, using AML response criteria from the European LeukemiaNet (ELN) 2017 guidelines (22). These AML response criteria were used rather than CML response criteria—which are largely designed for chronic phase CML—as they represent a clinical endpoint in advanced phase CML that facilitates bridging to potentially curative HSCT. Secondary endpoints included cytogenetic and molecular response rates, rate of subsequent HSCT, relapse-free survival (RFS), overall survival (OS), and safety of the regimen. Cytogenetic and PCR responses were determined using CML response criteria from the ELN 2013 guidelines (23). RFS was calculated from the time of CR/CRi until relapse or death from any cause, censored if alive at last follow-up. OS was calculated from the time of treatment initiation until death from any cause, censored if alive at last follow-up. Safety was assessed with the Common Terminology Criteria for Adverse Events (CTCAE) version 5.

Statistical analysis

An initial 6-patient safety lead-in was performed to confirm the safety of the combination regimen. If ≤1 patient experienced a dose-limiting toxicity (DLT) in the first cycle attributable to the treatment components, then the study would proceed to maximum enrollment. Interim monitoring rules for efficacy and toxicity were used throughout the study period. The study was continuously monitored for efficacy and treatment-related toxicities using a Bayesian design (24, 25). The regimen was considered promising if it achieved a CR/CRi rate of ≥40% and if the drug-related grade ≥3 toxicity rate in the first cycle was <20%. The interim monitoring rules were designed such that the study would be stopped for futility if it were unlikely (i.e., probability <2·5%) that the goal CR/CRi rate would be met, and the study would be stopped for excessive toxicity if it were highly probable (i.e., probability >80%) that the unacceptable toxicity threshold would be met. With 20 patients treated, if 12 patients (40%) achieve CR/CRi and using beta (0·8,1·2), then the 95% credible interval for CR/CRi rate would be (0·21–0·61). Patient characteristics were summarized using the median (IQR) for continuous variables and the frequencies (percentages) for categorical variables. Remission duration, RFS and OS were calculated with Kaplan-Meier estimates, and survival estimates were compared with the log-rank test. All analyses were performed in the intention-to-treat population. The data cutoff for this analysis was February 10, 2024. The data analyses were done using GraphPad Prism 9.

Role of the funding source

The funders provided free study drugs and funding for a research nurse for this study. The funders had no role in study design, data collection, data analysis, data interpretation, or writing of this report.

Results

Between July 12, 2020 and July 8, 2023, 20 patients were treated (Table 1; Figure 1). The median age was 42 years (IQR, 32–58 years), and 5 patients (25%) were ≥60 years of age. Fourteen patients (70%) had myeloid CML-BP, 4 (20%) had CML-AP, and 2 (10%) had Ph+ AML. Among the 4 patients with CML-AP, the AP-defining features were BM blast percentage between 10–19% in all 4 patients, with 2 patients also harboring new ACAs while on TKI therapy. Among patients with CML-AP or CML-BP transformed from prior CML-CP, the median duration of CML-CP was 18·6 months (IQR, 12·3–44·0 months). Fourteen patients (70%) had received prior therapy, and the median number of prior BCR::ABL1 TKIs was 2 (IQR, 0–2·5). Twelve patients (60%) had received 2 or more prior BCR::ABL1 TKIs. Five patients (25%) had received prior ponatinib, all of whom had relapsed on or were refractory to it, including 3 patients who had received ponatinib within 2 weeks of study enrollment. Five patients (25%) had received prior AML-type chemotherapy. Among the 6 patients who had not received prior therapy, 3 had CML-BP, 2 had CML-AP, and 1 had Ph+ AML. Three patients (15%) had a p190 BCR::ABL1 transcript (1 of whom also had concurrent p230); all other patients had the p210 transcript. Sixteen patients (80%) had an ACA, including 14 patients (70%) with at least one high-risk ACA and/or complex karyotype: complex karyotype (n=5), MECOM rearrangement (n=4), trisomy 8 (n=3), isochromosome 17q (n=1) and monosomy 7 (n=1). Eight patients (40%) had an ABL1 kinase domain mutation at study entry, including T315I and E255K in 2 patients each, and Q252H, L384M, G250E, and P465L + A337T in 1 patient each.

Table 1.

Baseline characteristics

Characteristic (n=20)	N (%) or median [IQR]
Age, years	42.5 [32–58]
Age ≥60	5 (25)
Sex
Male	13 (65)
Female	7 (35)
Race or ethnicity
Asian	1 (5)
Black	4 (20)
Hispanic	3 (15)
White	12 (60)
ECOG performance status
0	6 (30)
1	13 (65)
≥2	1 (5)
Disease subtype
Blast phase CML	14 (70)
Accelerated phase CML	4 (20)
Ph+ AML	2 (10)
Previously treated	14 (70)
Blast phase CML	11 (55)
Accelerated phase CML	2 (10)
Ph+ AML	1 (5)
Prior TKI use	14 (70)
Number of prior TKIs	2 [0–2.5]
1 prior TKI	2 (10)
2 prior TKIs	7 (35)
≥3 prior TKIs	5 (25)
Prior ponatinib	5 (25)
Prior AML-type therapy	5 (25)
Extramedullary disease	1 (5)
BCR::ABL1 transcript
p210	17 (85)
p190*	3 (15)
ACAs	16 (80)
Complex	5 (25)
MECOM rearrangement	4 (20)
Trisomy 8	3 (15)
Isochromosome 17q	1 (5)
Monosomy 7	1 (5)
Variant Philadelphia chromosome	1 (5)
Other	1 (5)
High-risk ACA †	14 (70)
ABL1 kinase domain mutation	8 (40)
T315I	2 (10)
E255K	2 (10)
Q252H	1 (5)
L384M	1 (5)
G250E	1 (5)
P465L + A337T	1 (5)
Other mutations
ASXL1	4 (20)
RUNX1	4 (20)
BCOR	2 (10)
TP53	1 (5)

Abbreviations: ECOG, Eastern Cooperative Oncology Group; Ph+, Philadelphia chromosome positive; TKI, tyrosine kinase inhibitor; ACA, additional chromosomal abnormality

^*One patient had concurrent p230 transcript.

^†High-risk ACAs included: complex karyotype, MECOM rearrangement, isochromosome 17q, monosomy 7, and trisomy 8

Figure 1.

Trial profile

There were no DLTs observed in the initial 6-patient run-in part of the trial, and therefore the study proceeded to maximum enrollment. Two patients experienced grade 3 or higher non-hematologic adverse events in cycle 1 (grade 3 rhabdomyolysis and grade 3 febrile neutropenia in 1 patient each) and were considered not related to the study drugs and therefore did not meet DLT criteria. Hematologic recovery for these 6 patients is shown in Supplemental Table 1. Despite 2 patients who achieved MLFS having prolonged myelosuppression, neither met hematologic DLT criteria due to the presence of significant residual disease by flow cytometry and/or fluorescent in situ hybridization or PCR for BCR::ABL1.

Response rates are shown in Table 2. The CR/CRi rate was 50%, including 1 patient (5%) with CR and 9 patients (45%) with CRi. An additional 6 patients (30%) achieved a morphologic leukemia-free state (MLFS), for a total marrow response (i.e. conversion to chronic phase CML) rate of 80%. Four patients did not meet formal response criteria, including 1 patient who had an aplastic marrow without evidence of leukemia but not meeting criteria for MLFS. Best responses after cycle 1 are shown in Supplemental Table 2. After cycle 1, the CR/CRi rate was 35% (all CRi), and the CR/CRi/MLFS rate was 70%. Two patients who had not achieved formal response after cycle 1 achieved a marrow remission by the end of cycle 2. Among 10 patients who achieved CR/CRi/MLFS and had an adequate flow cytometry sample for measurable residual disease (MRD) assessment at the time of best response, 2 (20%) were MRD-negative. Best cytogenetic response was complete cytogenetic response (CCyR) in 8 patients (40%) and partial cytogenetic response (PCyR) in 1 patient (5%). Seven patients were not evaluable for cytogenetic response due to insufficient material for post-treatment karyotype. Best molecular response was undetectable BCR::ABL1 transcripts in 3 patients (15%), major molecular response (MMR) in 1 patient (5%), and MR2 (i.e. 2-log reduction in BCR::ABL1) in 2 patients (10%).

Table 2.

Hematologic, cytogenetic, and molecular responses

Response, N (%)	All patients (n=20)	CML-BP (n=14)	CML-AP (n=4)	Ph+ AML (n=2)
Hematologic response
CR	1 (5)	0	1 (5)	0
CRi	9 (45)	6 (43)	3 (75)	0
MLFS	6 (30)	6 (43)	0	0
No response	4 (20)	2 (14)	0	2 (100)
Composite response
CRc (CR/CRi)	10 (50)	6 (43)	4 (100)	0
Marrow remission (CR/CRi/MLFS)	16 (80)	12 (86)	4 (100)	0
Cytogenetic response
CCyR	8 (40)	4 (29)	4 (100)	0
PCyR	1 (5)	1 (7)	0	0
No response*	4 (20)	4 (29)	0	0
Unevaluable	7 (35)	5 (36)	0	2 (100)
Molecular response
CMR	3 (15)	1 (7)	2 (50)	0
MMR	1 (5)	1 (7)	0	0
MR2	2 (10)	2 (14)
No response†	14 (70)	10 (71)	2 (50)	2 (100)

Abbreviations: CML-BP, CML-blast phase; CML-AP, CML-accelerated phase; CRc, composite complete remission; CR, complete response; CRi, complete response with incomplete count recovery; MLFS, morphologic leukemic-free state; CCyR, complete cytogenetic response; PCyR, partial cytogenetic response; CMR, complete molecular response; MMR, major molecular response; MR2, molecular response 2 (i.e. 2-log reduction in BCR::ABL1 transcripts)

^*Includes all patients not achieving a PCyR or better.

^†Includes all patients not achieving a molecular response of MR2 or better.

Responses rates by disease subgroups are shown in Table 2. CR/CRi was achieved in all 4 patients (100%) with CML-AP and in 6 of 14 (43%) with CML-BP. The overall marrow response rate (CR+CRi+MLFS) in patients with CML-BP was 86% (12 out of 14). CCyR was achieved in all 4 patients (100%) with CML-AP and in 4 of 14 (29%) with CML-BP. Two of 4 patients (50%) with CML-AP achieved complete molecular response (CMR; undetectable BCR::ABL1 transcripts) as best molecular response. One of 14 patients (7%) with CML-BP achieved CMR, 1 (7%) achieved MMR, and 2 (14%) achieved MR2. Neither of the 2 patients with Ph+ AML responded to the regimen. A post hoc analysis of response rates by other baseline characteristics is shown in Supplemental Table 3. The CR/CRi rates were similar for those regardless of prior therapy (50% [3 out of 6] for no prior therapy versus 50% [7 out of 14] for prior therapy), prior exposure to ponatinib (40% [2 out of 5] for prior ponatinib versus 53% [8 out of 15] for those without prior ponatinib, or the presence of an ABL1 kinase domain mutation at the time of study enrollment (42% [5 out of 12] for no mutation versus 63% [5 out of 8] for those with an ABL1 mutation). In contrast, patients with at least 1 high-risk ACA and/or complex cytogenetics had lower rates of CR/CRi than did those without a high-risk karyotype feature (36% [5 out of 14] versus 83% [5 out of 6], respectively; P=0·15). Characteristics of each patient enrolled and their responses are shown in Supplemental Table 1.

The median number of cycles received was 2·5 (IQR, 1–3·5); the median number of cycles was the same for patients who went to HSCT and those who did not. Seven patients (35%) received 1 cycle, 3 patients (15%) received 2 cycles, 5 patients (25%) received 3 cycles and 5 patients (25%) received 4 or more cycles. Two patients who did not undergo HSCT were transitioned to ponatinib monotherapy after 5 and 6 cycles of the triplet regimen, respectively. The median follow-up is 21·2 months (IQR, 14·1–24·2 months). The disposition for the 20 patients is shown in Supplemental Figure 1. Among the 16 patients who achieved a marrow remission (i.e. CR, CRi or MLFS), 8 patients (50%) underwent subsequent allogeneic HSCT, 5 relapsed in the absence of HSCT, and 3 are alive in continuous response without HSCT. Among the 8 transplanted patients, the median time to HSCT was 4·2 months (IQR 3·0–5·1) from the start of protocol therapy; 7 patients received myeloablative conditioning and 1 received reduced-intensity conditioning. Four of the 8 transplanted patients (50%) received post-transplant TKI maintenance (all with ponatinib orally at 15mg daily). Three patients relapsed after HSCT (2 of whom subsequently died), 1 died in remission from cardiac arrest, and 4 remain alive in continuous remission with a median duration of post-transplant remission of 12·5 months (IQR, 6·5–28·0 months). The patient who died in remission from cardiac arrest was 76 years old and did not receive post-transplant ponatinib maintenance. Among the 5 patients who relapsed in the absence of HSCT, 4 died and 1 was salvaged with a regimen of FLAG-Ida, venetoclax and ponatinib, underwent subsequent HSCT, and is still alive and in continuous remission 14·4 months later.

The baseline and relapse characteristics of 8 relapsed patients are shown in Supplemental Table 4. At last follow-up, 9 of the 16 responding patients (57%) have relapsed or died. The median RFS was 7·7 months (95% CI, 4·8-not reached), and the 1-year and 2-year RFS rates were both 37% (Figure 2A). At last follow-up, 4 patients were alive in continuous remission for >12 months, 3 of whom underwent HSCT (2 with CML-BP and 1 with previously treated CML-AP) and 1 of whom did not undergo HSCT and had previously treated CML-BP. At last follow-up 11 patients (55%) had died. The median OS was 11·1 months (95% CI, 8·2-not reached), and the 1-year and 2-year OS rates were 41% and 34%, respectively (Figure 2B). OS rates according to subsequent HSCT and by disease subtype are shown in Supplemental Figures 2 and 3. In the 14 patients with myeloid CML-BP, the median OS was 10·8 months (95% CI, 8·1-not reached). Among the 8 transplanted patients, the median OS was not reached, and the 2-year OS rate was 51%.

Figure 2.

Survival outcomes, (A) relapse-free survival and (B) overall survival

Non-hematologic adverse events are shown in Table 3. Eight patients (40%) experienced at least one cardiovascular event of any grade, of which only 2 events were grade 3 or higher. One patient had a grade 3 myocardial infarction while on ponatinib 45mg during cycle 2. Coronary catheterization shown no obstructive coronary atherosclerosis, and the ponatinib dose was decreased to 15mg daily without further cardiovascular events. Six patients had treatment-emergent hypertension (grade 1–2 in 5 patients and grade 3 in 1 patient). Other cardiovascular adverse events included: grade 2 atrial fibrillation, grade 2 pericardial effusion, grade 1 pulmonary hypertension, and grade 1 and grade 2 thromboembolism, all in 1 patient each. Treatment-emergent infections occurred in 10 patients (grade 1–2 in 1 patient and grade 3–5 in 9 patients [45%]). Non-hematologic adverse events occurring in the first cycle are shown in Supplemental Table 5. There were 3 on-study deaths, none of which was considered related to the study treatment and all from infection in the setting of refractory leukemia. The 60-day mortality rate was 0%.

Table 3.

Non-hematologic adverse events

Adverse event	Any grade	Grade 1–2	Grade 3	Grade 4	Grade 5
Nausea/vomiting	15 (75)	15 (75)	0	0	0
Rash	15 (75)	15 (75)	0	0	0
Constipation	13 (65)	13 (65)	0	0	0
Increased AST/ALT	12 (60)	7 (35)	5 (25)	0	0
Infection	10 (50)	1 (5)	6 (30)	0	3 (15)
Mucositis	10 (50)	9 (45)	1 (5)	0	0
Arthralgias/Myalgias	10 (50)	10 (50)	0	0	0
Headache	9 (45)	9 (45)	0	0	0
Febrile neutropenia	8 (40)	0	8 (40)	0	0
Abdominal pain	8 (40)	7 (35)	1 (5)	0	0
Fluid overload	8 (40)	7 (35)	1 (5)	0	0
Diarrhea	8 (40)	8 (40)	0	0	0
Hypoalbuminemia	8 (40)	8 (40)	0	0	0
Fatigue	7 (35)	7 (35)	0	0	0
Hypocalcemia	7 (35)	7 (35)	0	0	0
Hypokalemia	7 (35)	7 (35)	0	0	0
Hyponatremia	7 (35)	7 (35)	0	0	0
Pain (extremities/back)	7 (35)	7 (35)	0	0	0
Hypertension	6 (30)	5 (25)	1 (5)	0	0
Hyperuricemia	6 (30)	6 (30)	0	0	0
Hypophosphatemia	6 (30)	6 (30)	0	0	0
Acute kidney injury	5 (25)	3 (15)	1 (5)	1 (5)	0
Hyperglycemia	5 (25)	4 (20)	1 (5)	0	0
Increased bilirubin	5 (25)	4 (20)	1 (5)	0	0
Hyperphosphatemia	5 (25)	5 (25)	0	0	0
Insomnia	5 (25)	5 (25)	0	0	0
Dizziness	4 (20)	4 (20)	0	0	0
Dyspnea	4 (20)	4 (20)	0	0	0
Generalized muscle weakness	4 (20)	4 (20)	0	0	0
Increased ALP	4 (20)	4 (20)	0	0	0
Pleural effusion	4 (20)	4 (20)	0	0	0
Paresthesia/neuropathy	3 (15)	3 (15)	0	0	0
Hypotension	2 (10)	1 (5)	1 (5)	0	0
Hypermagnesemia	2 (10)	2 (10)	0	0	0
Thromboembolic event	2 (10)	2 (10)	0	0	0
Tumor lysis syndrome	1 (5)	0	0	1 (5)	0
Cholecystitis	1 (5)	0	1 (5)	0	0
Myocardial infarction	1 (5)	0	1 (5)	0	0
Rhabdomyolysis	1 (5)	0	1 (5)	0	0
Atrial fibrillation	1 (5)	1 (5)	0	0	0
Pericardial effusion	1 (5)	1 (5)	0	0	0
Pulmonary hypertension	1 (5)	1 (5)	0	0	0

Abbreviations: ALP, alkaline phosphatase; AST, aspartate aminotransferase; ALT, alanine aminotransferase

Data are n (%). All cardiovascular adverse events and all grade 3, 4, and 5 adverse events are included.

Any non-cardiovascular grade 1–2 adverse events occurring in ≥10% of patients are included.

Among responding patients, the median time to absolute neutrophil count > 0·5 × 10⁹/L was 41 days (IQR, 38–48) in cycle 1 and 35 days (IQR, 24·5–36·5) in cycle 2. The median time to platelet count > 25 × 10⁹/L was 27·5 days (IQR, 16·5–44·5) in cycle 1 and 25 days (IQR, 12–27) in cycle 2. Ten patients (50%) had at least one dose reduction of one or more study drugs (Supplemental Figure 4). Overall, 3 patients had a dose reduction of decitabine, 8 had a dose reduction of venetoclax, and 3 had a dose reduction of ponatinib (excluding protocol-mandated reductions due achievement of morphologic and/or molecular response). Dose reductions of decitabine or venetoclax were due to concerns of myelosuppression in all cases. The reasons for ponatinib dose reductions were myocardial infarction without angiographic evidence of coronary atherosclerosis (mentioned before), abdominal pain without evidence of pancreatitis or other hepatobiliary disease, and neutropenic fever. No patient discontinued therapy due to treatment-related toxicity.

Discussion

Advanced phase CML is associated with poor outcomes with conventional therapies. While consensus guidelines recommend AML-like chemotherapy for patients with myeloid CML-BP, there is no clear standard of care regimen in this setting. In this study, which was enriched with patients with myeloid CML-BP (70% of the cohort), the triple combination of decitabine, venetoclax and ponatinib resulted in a CR/CRi rate of 50%, a CR/CRi/MLFS rate of 80%, and a median OS of 11·1 months (95% CI, 8·2-not reached). The regimen was delivered with acceptable toxicity using response-based dosing of ponatinib, with no patients discontinuing therapy due to treatment-related toxicity and no treatment-related deaths, although dose reductions were still needed in 50% of patients, primarily due to myelosuppression.

Several prospective studies have evaluated TKI monotherapy in CML-BP. The most relevant to our study is the experience with ponatinib monotherapy in this subgroup of patients in the PACE study (12, 26). In this study, 62 patients with CML-BP (52 of whom had myeloid immunophenotype) received ponatinib 45mg orally daily, 29% of whom achieved a major hematological response (an endpoint equivalent to the composite of CR/CRi/MLFS in the present study) and the median OS was 7 months. In contrast, in the subgroup of patients with myeloid CML-BP in our study, the CR/CRi/MLFS rate was 86% and the median OS was 10·8 months (95% CI, 8·1-not reached). Acknowledging the limitations of cross-study comparisons, the numerically higher response rates and OS in our study suggest benefit of adding decitabine and venetoclax to ponatinib. The activity of the regimen in patients with prior ponatinib exposure, where the CR/CRi rate was 40% and the CR/CRi/MLFS rate was 80%, also support this conclusion. Overall, these results are consistent with retrospective data suggesting superior outcomes when BCR::ABL1 TKI is combined with either intensive chemotherapy or a hypomethylating agent, as compared with a TKI alone, in patients with myeloid CML-BP. In a retrospective analysis from MDACC, the use of combination TKI-based therapy resulted in a 5-year OS rate of 34%, compared with a 5-year OS rate of only 8% with TKI monotherapy (7). Our findings therefore support the consensus recommendations for use of AML-like chemotherapy in combination with a BCR::ABL1 TKI for these patients.

The overall rate of conversion to chronic phase CML (equivalent to a response of CR/CRi/MLFS) in our study was encouraging, as this is an important response endpoint that facilitates use of potentially curative consolidative allogeneic HSCT (4). However, the rate of MMR or deeper was relatively low (20% overall). Treatment strategies that lead to deeper molecular remissions in patients with advanced phase CML will be necessary in order for patients to achieve durable remissions without allogeneic HSCT. Peripheral blood count recovery was also incomplete in most patients in our study (with only 1 patient [5%] achieving CR). While this could reflect the relatively high-risk, heavily treated population of our study, it is possible that use of a hypomethylating agent and venetoclax-containing backbone contributed to the lack of full count recovery in most patients.

To our knowledge, our study is the second to prospectively evaluate ponatinib-based combination therapy for patients with advanced phase AML and the first to evaluate venetoclax and TKI combination regimen. The MATCHPOINT study was a phase I/II study that evaluated FLAG-Ida plus ponatinib in patients with CML-BP (9 myeloid, 4 lymphoid and 4 mixed immunophenotype) (16). Our study used different response criteria than did MATCHPOINT, making a comparison of response rates challenging. However, the survival data from our study appears relatively similar to MATCHPOINT, with a median OS of 11·1 months and 12 months, respectively. While our study also included a small population of patients with CML-AP (n=4), our study generally consisted of poorer risk patients, as our patients were older (median age 42 years versus 33 years in MATCHPOINT), had higher rates of ACAs (80% versus 47%, respectively) and were more heavily pretreated (60% with ≥2 prior TKIs versus 12%, respectively). Acknowledging the challenges of cross-study comparisons, the similar outcomes achieved with our lower intensity regimen question whether an intensive chemotherapy backbone is needed when treating advanced phase CML in younger fit patients. This notion is supported by a retrospective analysis from MDACC that showed similar outcomes with intensive chemotherapy plus a TKI versus a hypomethylating agent plus a TKI in myeloid CML-BP (7, 10). Given preclinical and retrospective clinical data suggesting potential synergy between venetoclax and BCR::ABL1 TKIs (17, 18, 27), it is possible that a combination of a hypomethylating agent, venetoclax and TKI could even be superior to intensive chemotherapy plus a TKI. Additional prospective studies will be needed to answer this important question.

This regimen could be delivered safely with no evidence of synergistic non-hematologic toxicity. We also employed response-based dosing of ponatinib, which has been used in several other ponatinib studies (14, 21, 28), in order to mitigate cardiovascular and other potential toxicities of ponatinib. Forty percent of patients experienced at least 1 cardiovascular event; however, only two grade 3 or higher cardiovascular events were observed (grade 3 myocardial infarction without obstructive coronary atherosclerosis in 1 patient and grade 3 hypertension in 1 patient). The rate of treatment-emergent infections, which occurred in 50% of patients, is also consistent with expectations with hypomethylating agent plus venetoclax combinations, particularly when used in heavily pretreated patients (29).

Our study has several limitations. First, only 20 patients were treated. However, our study is larger than most other studies of combination therapy in CML-BP, apart from a study of “7+3” plus imatinib (n=36), largely reflecting the rarity of this group of diseases (30). Our study also enrolled patients with myeloid CML-BP, CML-AP and Ph+ AML, adding heterogeneity to the patient population. However, the study was enriched with patients with myeloid CML-BP (n=14), providing a sufficient number to evaluate response rates and outcomes in this specific population. It is also worth noting that the most recent WHO guidelines eliminated the term “CML-AP” and replaced it with “high-risk CML.” Our study was designed and conducted prior to this recent change, although the data in the small population of patients with CML-AP (n=4) could reasonably be extrapolated to this newly defined “high-risk CML” group.

In conclusion, the combination of decitabine, venetoclax and ponatinib was safe and shows promising activity in patients with advanced phase CML, including in heavily pretreated patients and those with poor-risk disease-related features. However, despite the high rates of marrow remission with this regimen, long-term survival was suboptimal for most patients. New biological insights into the pathogenesis of CML transformation may allow for better targeting of therapeutic vulnerabilities in this disease. Additional prospective studies of novel regimens combining newer, potent TKIs and venetoclax are needed.

Research in context.

Evidence before this study

A systematic review was not performed before starting this trial. However, during preparation of the study protocol in 2019, we searched PubMed without date restriction for studies published in English on the outcomes of adults with advanced phase Philadelphia chromosome-positive (Ph+) myeloid disease and clinical trials in this population using the keywords “chronic myeloid leukemia” (CML), “Philadelphia chromosome-positive”, “blast phase”, and “treatment.” There is no clear standard of care for patients with advanced phase chronic myeloid leukemia or Ph+ acute myeloid leukemia. Consensus guidelines recommend that patients with myeloid blast phase CML should receive AML-type induction chemotherapy in combination with a BCR::ABL1 tyrosine kinase inhibitor (TKI). Retrospective studies suggest benefit of combination therapy with either intensive chemotherapy or a hypomethylating agent in combination with a BCR::ABL1 TKI, rather than a TKI alone. However, prospective studies supporting a specific regimen in advanced phase chronic myeloid leukemia are scant. Preclinical and clinical data suggest that the combination of venetoclax and a BCR::ABL1 TKI may have synergistic activity.

Added value of this study

To our knowledge, this is the first prospective clinical trial evaluating a low-intensity regimen combining venetoclax and a BCR::ABL1 TKI in patients with advanced phase Ph+ myeloid leukemia. In our study, the combination of decitabine, venetoclax and ponatinib was safe and resulted in high response rates in patients with advanced phase chronic myeloid leukemia, including in heavily pretreated patients and those with poor-risk disease-related features.

Implications of all the available evidence

For patients with advanced phase CML, combination therapy with decitabine, venetoclax and ponatinib induces high rates of marrow response, which may bridge patients to a potentially curative allogeneic stem cell transplant. This study provides further evidence supporting the use of a ponatinib-based combination regimen for this disease. Further prospective studies evaluating chemotherapy and venetoclax-based combination strategies using newer-generation BCR::ABL1 TKIs are warranted.

Data sharing statement

The protocol is available in the appendix (pp 2–28). Data can be shared with qualified researchers upon reasonable request to the corresponding author. No identifying data will be provided.