Clinical Outcomes of Adult Patients with Newly Diagnosed Mixed Phenotype Acute Leukemia

Abstract

PURPOSE:

Mixed phenotype acute leukemia (MPAL) is a rare clinical entity with historically poor outcomes.

PATIENTS AND METHODS:

We conducted a retrospective analysis of adults ≥18 years with newly-diagnosed B-cell (B/M) or T-cell/myeloid (T/M) MPAL treated at our institution between 2017–2024.

RESULTS:

We identified 42 patients (median age 70 years); 20 (48%) had B/M and 22 (52%) had T/M MPAL; 57% patients had adverse risk cytogenetics, and 41% had a TP53-mutation. Sixty-two percent of patients were treated with a hybrid regimen and 45% of patients received intensive therapy. A composite complete remission (CRc; CR+ CRi) was achieved in 57% of patients (86% measurable residual disease [MRD]-negative). After a median follow-up of 27.9 months, the median relapse-free survival in patients achieving an overall response (CRc+ morphologic leukemia free state) was 10.1 months, 17.8 months in those who achieved a CRc, and not reached (NR) in patients with MRD-negative CRc. The median overall survival (OS) for all patients was 9.5 months and NR for patients achieving a CRc. Though patients with T/M had a trend towards improved survival compared to B/M (median OS of 9.1 versus 25 months p = 0.28), this difference abrogated when comparison was stratified by treatment intensity. Twelve patients (29%) underwent allogeneic hematopoietic stem cell transplantation (HSCT); on landmark analysis HSCT trended to improve OS (NR versus 22.8, p = 0.12). Multivariate Cox analysis demonstrated TP53-mutation was associated with increased hazards for death (HR 3.5, p = 0.01), while the use of intensive chemotherapy trended to be favorable (HR 0.45, p = 0.11).

CONCLUSION:

Overall, these data demonstrate the need for treatment intensification in MPAL with HSCT in first remission for best outcomes.

Introduction

Mixed phenotype acute leukemia (MPAL) is a rare and aggressive form of leukemia, encompassing less than 5% of all newly diagnosed acute leukemias.¹ Patients with MPAL demonstrate either concomitant, yet separate populations of myeloid and lymphoid phenotypic blasts (bi-lineal), or one population which includes features of both lineage (bi-phenotypic).² There have been variability in the diagnostic criteria for MPAL, however the two more recently developed algorithms include the European Group for Immunological Characterization of Acute Leukemias (EGIL)^3–5 and the World Health Organization (WHO) classification of myeloid neoplasms and acute leukemia.⁶ However, despite these criteria, the diagnosis of acute leukemias with ambiguous lineage (ALAL) and MPAL remain challenging due to the often vague distinctions between cell populations, lack of consistent diagnostic standards, no specific molecular signature in most cases and reliance on expert immunophenotyping.² Furthermore, given the general rarity of MPAL, there is a paucity of data surrounding the clinical characteristics and outcomes for these patients and no consensus guidelines for treatment, leading to significant variability in practice.

In recent iterations of these classification systems by the WHO, there have been greater emphasis on genomic markers, including rearrangements in FGFR1, a specific rearrangement associated with myeloid/lymphoid neoplasia and eosinophilia and amenable to targeted inhibition by pemigatinib⁷, or BCL11B rearrangement leukemias now classified as acute leukemia of ambiguous lineage with specific genetic aberration.⁶ MPAL with BCR::ABL1 rearrangements are relatively more common, thus conferring sensitivity to BCR::ABL1 tyrosine kinase inhibitors (TKIs), while KMT2A-rearranged MPALs are amenable to therapy with menin inhibitors. However, the majority of MPALs lack any defining specific genetic aberration and are sometimes treated with a hybrid regimen containing both myeloid and lymphoid directed regimens, along with the addition of targeted agents if such a genetic lesion is identified. Therapy with lymphoid directed therapies have been generally preferred over myeloid directed therapies, though strong evidence and prospective studies are lacking.^8–10 While allogeneic hematopoietic stem cell transplantation (HSCT) consolidation in first complete remission (CR1) following ALL-directed treatment is the usual treatment algorithm,^2,11 data regarding outcomes of molecularly-annotated cohorts of patients with MPAL remain limited. Additionally, there is a lack of data surrounding the efficacy of B-cell directed immunotherapies such as CD19- or CD22-targeting agents, blinatumomab or inotuzumab ozogamicin (InO), as well as the BCL2-targeting agent, venetoclax on outcomes in patients with MPAL. Venetoclax has potent clinical efficacy in myeloid leukemias and has shown significant preclinical efficacy and early clinical activity in lymphoblastic leukemias, especially T-cell ALL¹².

We therefore conducted a retrospective analysis of a modern cohort of patients with MPAL treated at our institution stratified by treatment regimen to provide insights into the optimal treatment approach for this high-risk group of patients.

Methods

Study Design, Participants and Treatment

We conducted a single-center, retrospective analysis of adult patients ≥18 years of age treated at our institution between the years of 2017 and 2024 to identify patients with newly diagnosed B-cell or T-cell/myeloid MPAL. We defined MPAL as acute leukemia expressing markers of more than one lineage, based on strict immunophenotypic criteria.^6,13 MPAL included mixed phenotype and/or mixed lineage acute leukemia. All patients had our institution’s clinically validated 81-gene panel via next generation sequencing (NGS) panel performed. Patients with chronic myeloid leukemia in blast phase were excluded. Given all patients had a myeloid component, risk stratification was done as per European LeukemiaNet (ELN) 2017 for AML for homogeneity¹⁴. Treatment was stratified as hybrid or AML/ALL directed. Hybrid regimens included combinations in which ALL directed agents (vincristine, asparaginase, blinatumomab, inotuzumab ozogamicin) were added to specific AML directed agents (hypomethylating agents, low-dose cytarabine, cladribine, gemtuzumab ozogamicin). Therapy was considered as intensive when patients received anthracyclines and/or cytarabine at >1gm/m² of body surface area dose as part of their treatment regimen. This study was approved by the Institutional Review Board and was conducted in compliance with the Declaration of Helsinki.

Outcomes and Statistical Analysis

We obtained baseline characteristics of the included patients, of which categorical variables were collected as frequencies and continuous variables were collected as medians. The collected baseline characteristics included age at diagnosis, race, ethnicity, MPAL phenotype, bone marrow blast percentage, cytogenetics, mutations, and baseline blood cell count (white blood cell [WBC] count, hemoglobin and platelet count). Cluster of differentiation (CD) flow markers were collected and were considered to be positive if their expression was at a level of 10% or greater. Differences among variables were determine via the Chi-square test for categorical variables, whereas continuous variables were assessed via the Mann-Whitney U test.

The following outcomes were assessed: composite complete remission (CR) rate ([CRc], composite of CR or CR with incomplete count recovery [CRi]), objective response (OR) rate (CRc + morphologic leukemia free state [MLFS]), overall survival (OS, defined as the time from initiation of treatment to death from any cause with censoring at last follow-up), and relapse-free survival (RFS, defined as the time from achievement of best response to relapse or death). Measurable residual disease (MRD) by multiparameter flow cytometry at a sensitivity of 1 in 10⁻⁴ in remission assessment bone marrows were reported as described before¹⁵. Follow up was estimated by inverse Kaplan Meier and OS and RFS were assessed by the Kaplan-Meier method, and subgroup comparisons were calculated using log-rank testing. Censoring was not performed at the time of HSCT. Multivariate Cox proportional hazard analysis with backward model selection was used to study factors associated with OS. Data analysis was performed with Graphpad Prism, Lumivero XLSTAT, New York, and R-4.4.0.

Results

Patients

Among 2498 patients with newly diagnosed acute leukemia treated between 2017–2024, 42 (1.7%) newly diagnosed MPAL were identified. The median age at diagnosis was 70 years (range 23 – 80 years) and 29 (69%) patients were ≥ 60 years of age (Table 1). Based on immunophenotypic analysis, 22 (52%) patients had T-/myeloid (T/M) leukemia, and 20 (48%) patients had B-/myeloid leukemia (B/M). All patients had bone marrow disease; 3 patients (7%) had central nervous system (CNS) disease at diagnosis (2 B/M and 1 T/M). Twenty-three patients (55%) had a single blast population with a mixed phenotype, whereas the remaining patients (45%) had mixed lineage blasts with 2–3 phenotypically distinct subclones. Of the latter, one patient in the B/M population with a mixed B/M clone in addition to a predominant B-clone, whereas there were 3 patients with T/M leukemia who had a mixed T/M clone in addition to myeloid and T clones. For the B/M cohort, the median involvement by the myeloid clone was 80% compared to 20% median involvement by the B-clone. For the T/M cohort, the median involvement by the myeloid- and T-clones were similar at 40% and 46%, respectively.

Table 1. Baseline Characteristics.

depicts the baseline characteristics of the included patients, including the entire cohort, and stratified by B/M MPAL vs. T/M MPAL.

	N (%), or median [range]			p – value (B/M vs. T/M)
Characteristic	Full Cohort (N = 42)	B/M (N = 20)	T/M (N = 22)
Age (years)	70 [23 – 79.7]	67 [23.9 – 76.1]	67 [23 – 79.7]	0.89
Age ≥60 years	29 (59)	15 (75)	14 (64)	0.51
Sex
Male	24 (57)	12 (60)	12 (55)	0.72
Female	18 (43)	8 (40)	10 (45)
Race
White	28 (67)	15 (75)	13 (59)	0.27
Black	7 (17)	1 (5)	6 (27)	0.05
Asian	4 (10)	2 (10)	2 (9)	0.92
American Indian	1 (2)	1 (5)	0	0.29
Hawaiian or other Pacific Islander	1 (2)	0 (0)	1 (4)	0.33
Unknown	1 (2)	1 (5)	0 (0)	0.29
Ethnicity
Hispanic	5 (12)	4 (20)	1 (4)	0.12
Non-Hispanic	33 (79)	16 (80)	17 (77)	0.83
Declined or Unknown	4 (10)	0 (0)	4 (18)	0.04
Central Nervous System involvement	3 (7)	2 (10)	1 (4)	0.49
Prior chemo-radiotherapy	12 (29)	8 (40)	4 (18)	0.12
Bone marrow blasts (%)	70 [20 – 92]	67.5 [20 – 92]	71.3 [23 – 86]	0.91
Cytogenetics
Adverse	24 (57)	11 (55)	13 (59)	0.79
9;22 translocation	3 (7)	2 (10)	1 (4)	0.49
Baseline Blood Counts
WBC (1 × 109/L )	6.5 [0.4 – 230]	4.4 [0.4 – 53.7]	8.3 [1.1 – 230]	0.06
Platelet (1 × 109/L)	53 [8 – 338]	37 [8 – 154]	65 [19 – 338]	0.22
Immunophenotypic clones
Single	23 (55)	10 (50)	13 (59)
Mixed lineage with or without mixed phenotype (multiple clones)	19 (45)	10 (50)	9 (41)	N/A
Clinically relevant mutations
ASXL1	3 (7)	2 (10)	1 (4)	0.49
FLT3-ITD	6 (14)	4 (20)	2 (9)	0.31
FLT3-D835	3 (7)	2 (10)	1 (4)	0.49
IDH 1/2	5 (12)	2 (10)	3 (14)	0.72
JAK2/3	5 (12)	0 (0)	5 (23)	0.02
RUNX1	8 (19)	7 (35)	1 (4)	0.01
N/K/H RAS	12 (29)	4 (20)	8 (36)	0.24
SRSF2	4 (10)	3 (15)	1 (4)	0.25
TET2	5 (12)	1 (5)	4 (18)	0.19
TP53	17 (41)	8 (40)	9 (41)	0.95
WT1	5 (12)	3 (15)	2 (9)	0.55
Frontline treatment type
Myeloid	14 (33)	6 (30)	8 (36)	0.66
Lymphoid	2 (5)	2 (10)	0 (0)	0.13
Hybrid	26 (62)	12 (60)	14 (64)	0.81

Abbreviations: WBC, white blood cell

^*Indicates mixed clone as one of the sub-clones

There were no patients with an antecedent myeloid disorder and 12 (29%) patients had prior chemo-radiotherapy for non-myeloid cancers (4 patients with T/M and 8 patients with B/M leukemia). Twenty-four (57%) patients had adverse-risk cytogenetics (20 with complex karyotype, 2 with monosomy 7, one each with −17p and MECOM gene rearrangement), 14 (36%) patients had intermediate-risk cytogenetics (10 diploid karyotype and 4 other miscellaneous), and 3 (7%) patients were BCR::ABL1 (Philadelphia [Ph] chromosome) positive. Seventeen (40.5%) patients had a TP53 mutation (14 with complex cytogenetics, one patient with concurrent del 17p). Ten (24%) patients had a KMT2A amplification (5 B/M and 5 T/M), but no patient had a KMT2A rearrangement. Using the ELN 2017 AML risk stratification, 33 (79%) patients had adverse-risk, and 9 (21%) patients had intermediate risk disease. Six (14%) patients were positive for FLT3-ITD (4 B/M and 2 T/M), and 3 patients for FLT3-TKD (including 2 patients who had both). Among 8 patients evaluated through RNA translocation screen, one patient had an FGFR1 rearrangement. When comparing cytogenetic abnormalities in T/M versus B/M, there was a trend towards enrichment of −7/7q in B/M compared to T/M (37% versus 10%, p=0.06). Supplemental (S) Table S1 table compares cytogenetic features between T/M and B/M.

Treatment and response

Twenty-six (62%) patients were treated with a hybrid regimen that included both AML and ALL directed therapy (12 B/M, 14 T/M), 14 (33%) patients were treated with a predominantly myeloid directed regimen (6 B/M, 8 T/M) and 2 (5%) patients were treated with a predominantly lymphoid directed therapy (both B/M) (Figure 1). Treatment regimens were mostly at physician discretion and a detailed description of the regimens received and subsequent best responses for all patients is shown in supplemental Table S2. All 3 patients with Ph-positive disease received frontline BCR:ABL1 tyrosine kinase inhibitor (TKI) (2 ponatinib, 1 dasatinib) with chemotherapy and the one patient with FGFR1 rearrangement received pemigatinib. Venetoclax was administered as part of frontline therapy in 32 (76%) patients (13 B/M, 19 T/M). Seven of twenty (35%) patients with B/M received frontline blinatumomab in combination with chemotherapy. Five patients were treated with a FLT3 inhibitor (4 B/M and 1 T/M). Stratified by treatment intensity, 19 (45%) patients received an intensive chemotherapy backbone, 11 of whom were <60 years of age and 10/19 were hybrid regimens.

Figure 1. Oncoprint of the Study Patients.

Figure 1 depicts an oncoprint of the included patients, depicting key characteristics including the cytomolecular features of each patient, the classification of treatment regimen that they received, the response that each patient achieved, and if they proceeded to transplant.

A best response of CRc was achieved in 24 (57%) patients, which included 10/20 (50%) patients with B/M leukemia and 14/22 (64%) patients with T/M leukemia (Figure 2). Another 7 patients achieved an MLFS leading to an objective response (OR) rate of 74% (31 of 42 patients). MRD-negative CRc (negative for both clones) as best response was achieved in 19/24 (86%) patients (45% of the full cohort). Stratified by treatment intensity, 13/19 (68%) patients treated with intensive therapy achieved a CRc (11 MRD-negative) compared to 11/23 (49%) patents treated LIT (p=0.22); with respect to venetoclax, 17/32 (53%) patients who received venetoclax achieved a CRc (13 MRD-negative) compared to 7/10 (70%) patients who did not receive venetoclax (6 MRD-negative). All 3 patients with Ph+ leukemia achieved MRD-negative CR.

Figure 2. Responses.

Figure 2 depicts the responses of the included patients, stratified by T/M vs. B/M MPAL and non-intensive treatment vs. intensive treatment.

Survival outcomes

At a median follow up estimate of 28 months (95% CI 13.5–45 months), the median RFS of patients with an OR was 10 months (95% CI 7.8-Not reached [NR]), 18 months (95% CI 8.0-NR) in patients who achieved a CRc, while it was NR (95% CI NR-NR) in patients with an MRD-negative CRc (Figure 3A). The median OS of the full cohort was 9.5 months (95% CI 6.7–36.1 months) while it was NR (95% CI NR-NR) in patients with a CRc (2-year OS rate 54%, 95% CI 28.6–78.9%) (Figure 3B). The 2-year OS rate was 63% (95% CI 36.0–90.2%) in patients with an MRD-ve CRc. There was no significant difference in OS in B/M patients with 1 multi-phenotypic clone compared to those with multi-lineal clones (median OS 9.0 months, 95% CI 6.38-NR versus median OS 9.1 months 95% CI 1.71-NR) (Figure S1). Similarly, there was no significant difference in OS in T/M patients with 1 clone compared to those with multiple clones (median OS 25.2 months, 95% CI 6.05-NR versus median OS 36.1 months, 95% CI 9.51-NR) (Figure S2).

Figure 3. Survival Outcomes.

Figure 3 depicts the survival outcomes of the included patients. Figure 3A and 3B demonstrate relapse free survival and overall survival, respectively, stratified by all patients vs. patients achieving a composite complete remission (CRc) vs. patients achieving a measurable residual disease (MRD)-negative CRc.

The median OS in patients with B/M leukemia was 9.0 months compared to 25.0 months in patients with T/M leukemia (2-year OS 23% versus 42%; p=0.28). Among patients treated with intensive regimens, the median OS in patients with B/M leukemia (n=9) was 23.0 months compared to 25.0 months in T/M leukemia (n=10), p=0.91 (2-year OS rate 49.4% vs. 47.6%) (Figure 4A). Among patients treated with LIT, the median OS was 6.4 versus 8.4 months in B/M (n=11) and TM (n=12) leukemia respectively, p=0.16 (Figure 4B).

Figure 4. Survival Outcomes by Treatment Intensity.

Figure 4 depicts the survival outcomes by treatment intensity. Figure 4A demonstrates overall survival in intensively treated patients stratified by B/M vs. T/M MPAL. Figure 4B demonstrates overall survival in patients treated with low-intensity therapy stratified by B/M vs. T/M MPAL.

In the B/M cohort, the median OS of the 12 patients who received a hybrid therapy was 9 months compared to 16 months in the 8 patients who received an AML or ALL type therapy (Figure S3). Of note, 4/12 of the hybrid regimens had an intensive chemotherapy backbone compared to 5/8 AML or ALL directed regimens. In the T/M cohort the median OS of the 14 patients treated with a hybrid regimen was not reached (NR) versus 11.5 months in patients treated with AML or ALL like therapy, p=0.14 (6/14 patients treated with hybrid regimens received an intensive therapy, compared to 4/8 patients treated with AML/ALL like therapy) (Figure S4). Overall, 19/22 patients with T/M received venetoclax (8 with IC and 11 with LIT; 12 with a hybrid regimen and 7 with AML/ALL like regimen); the median OS in the IC+ venetoclax treated patients was 25 months while it was 10.7 months in LIT+ Ven treated patients.

Allogeneic hematopoietic stem cell transplantation

Twelve (29%) patients had undergone an allogeneic HSCT (Table S3, Figure S5); 4/20 (20%) patients with B/M leukemia and 8/22 (36%) patients treated with T/M leukemia. The median age of the transplanted patients was 60 years (range 23–75 years), and 6 (50%) patients were ≥60 years of age at diagnosis. The median time to HSCT from treatment initiation was 3.7 months (range 2.5–6.6 months); 11/12 patients had an MRD-negative CRc before HSCT and one patient had an MLFS (MRD-negative on an inadequate sample). Genomically, one patient had Ph+ leukemia, 7 patients had intermediate risk cytogenetics, and 4 patients had adverse cytogenetics (complex karyotype). Three patients had a TP53 mutation. Nine transplanted patients were treated with IC, and 8 patients had received a hybrid therapy. The median RFS and OS of the transplanted patients were 18 months (95% CI 7.7-NR) and NR (95% CI 9.8-NR) and 2-year rates were 62.3% and 72.7% respectively (Figure 5A). The 2-year OS rate among transplanted patients ≥60 years of age was 60% while it was 83.3% in patients <60 years of age. We performed a landmark comparison of all transplanted to non-transplanted patients including patients alive for >3.7 months (landmark), with an OR and age <75 years (n=16); the median OS was NR versus 22.8 months in the transplanted and comparator cohort respectively (2-year OS 72.7% versus 38.2%), p=0.12 (Figure 5B). The baseline characteristics and outcomes for TP53-mutated patients in our cohort is mentioned in supplemental Table 4 (Table S4).

Figure 5. Outcomes of Transplanted Patients.

Figure 5 depicts the survival outcomes of transplanted patients. Figure 5A demonstrates the overall survival and relapse free survival of all patients who proceeded to transplant. Figure 5B depicts a landmark analysis of all transplanted compared to non-transplanted patients where the landmark time included patients alive for >3.7 months. Patients in this analysis included those under the age of 75 years old with an overall response.

Factors affecting survival

We used a stepwise Cox regression (alpha <0.15 for inclusion) with age (continuous), type of MPAL (B/M versus T/M), cytogenetics (adverse or not; Ph+ was included in the non-adverse category), TP53 status (mutated or wild-type), treatment intensity (IC or LIT), use of venetoclax and HSCT (as a time dependent variable). Relevant factors selected by the model included TP53 status, with TP53 mutation associated with increased hazards of death, HR 3.5 (95% CI 1.35–9.2; p 0.01) while use of an IC backbone trended towards lower hazards of death, HR 0.45 (95% CI 0.17–1.2, p=0.11) (Table S5). In the backward elimination model (Table S5B), only TP53 was relevant with an independent HR of 5.1 (95% CI 2.1–12.1, p<0.01). On a treatment intensity stratified Cox regression with backward elimination only TP53 status was associated with hazards of death (HR 3.74, 95% CI 1.32–10.60, p=0.01). For the full cohort, HSCT univariately trended to be favorable, HR 0.33 (95% CI 0.10–1.05, p=0.06). ¹⁶

Discussion

Given the overall lack of consensus guidelines for the treatment of MPAL, we aimed to evaluate our own cohort of patients with newly diagnosed immunophenotypically characterized MPAL treated with modern ALL-, AML-, and hybrid regimens to ascertain pertinent clinical characteristics and outcomes. In our cohort of 42 patients with newly diagnosed MPAL, the median OS was only 9.5 months, which is consistent with previously reported poor outcomes in this population.^17,18 However, we show promising survival outcomes in the subgroup of patients treated with IC, with a median OS of between 23 and 25 months. The use of an intensive chemotherapy backbone was associated with an improved survival on univariate KM estimation and trended to be favorable on MV Cox stepwise regression. Younger patients, more frequently treated with IC, demonstrated durable survival outcomes with a clear plateauing of the survival curve.

The majority of included patients in our study (62%) received treatment with a hybrid-based approach, which was associated with a high CRc rate of between 70 and 80%. When assessed by MPAL subtype, patients with B/M MPAL treated with ALL or AML-directed therapy fared reasonably well with a median OS of 16 months, however those treated with a hybrid approach did poorly with a median OS of only 9 months. Conversely, those with T/M MPAL responded more favorably to hybrid regimen compared to ALL or AML directed therapy, and thus these patients should be likely offered such an approach. We observed inferior outcomes in patients with a TP53 mutation, and while HSCT was associated with a trend for improved survival univariately, this was not significant on multivariate analysis, possibly due to the small number of patients. The benefit of HSCT in MPAL has been demonstrated in prior studies, with a 3-year OS of 56% in a cohort of 519 patients with MPAL who underwent HSCT in CR1,¹⁸ which is consistent with our data, as demonstrated by our landmark analysis of a median OS NR for patients proceeding to HSCT. Thus, our data further supports the use of consolidative HSCT in CR1 for patients with MPAL.

Previous studies have attempted to identify genetic signatures and interrogate ontogeny of MPAL types (B/M and T/M) compared to AML and ALL. Integrated genomic analysis was conducted on 31 samples with MPAL and showed that patients with MPAL carried both AML- and ALL-type mutations¹⁹. In this study, B/M MPAL and T/M MPAL varied significantly in terms of their DNA methylation signatures, which were ultimately associated with a differential expression of lineage-commitment genes and could identify patients who benefitted from AML or ALL type therapy.¹⁹ Additionally, recently published data has suggested that MPAL and AML with a mixed phenotype (AML-MP) are two clinically and biologically distinct entities with unique molecular characteristics and response to treatment, thus suggesting a new classification model delineating these two groups.²⁰ Given the smaller number of patients included in our study, we were unable to validate these findings. Of note, spliceosome mutations were rare in our cohort (only 4 cases with SRSF2 mutations), in contrast to their higher prevalence in de novo or secondary AML with antecedent myelodysplasia²¹. Given that spliceosome mutations are associated with inferior outcomes to intensive chemotherapy in the absence of venetoclax^22,23 their near-absence may reflect distinct disease biology compared with typical myeloid disorders. Additionally, we observed a high incidence of TP53 mutations in our cohort of patients. This group in particular warrants further investigation given their overall poor outcomes, which is consistent with previous reports.²⁴ Our study is strengthened by the fact that patients had a long duration of follow-up and well curated data as majority of data was prospectively collected on clinical trials. Furthermore, patients received targeted therapy whenever eligible, thus ensuring our data is relevant with respect to novel treatment approaches. Limitations include single-center, retrospective design of our study as well as the relatively small number of included patients.

Based on our findings, regimens incorporating newer myeloid directed agents including venetoclax and lymphoid directed agents (blinatumomab and/or inotuzumab ozogamicin for B/M MPAL; nelarabine, pegylated asparaginase for T/M MPAL) warrants evaluation in prospective studies. While intensive therapy often constitutes drugs that have reasonable activity against myeloid and lymphoid blasts (intermediate/high dose cytarabine, fludarabine, anthracyclines) the same is not true with LIT and incorporating agents in LIT regimens to cover both phenotypes of MPAL is likely important in these patients. This is supported by our data as well given the poor outcomes in patients treated with LIT. In summary, in our cohort of newly diagnosed MPAL patients, intensive therapy followed by HSCT leads to promising long-term survival. Ongoing treatment optimization is warranted in patients with MPAL who are ineligible for intensive therapy.

Context Summary.

Key objective:

Treatment recommendations for mixed phenotypic acute leukemia (MPAL) based on prospective studies are lacking.

Knowledge generated:

In this retrospective series of 42 newly-diagnosed adult patients with MPAL, with 69% patients ≥60 years of age, 52% patients had T-myeloid and 48% had B-myeloid MPAL. Overall, 45% of patients received intensive therapy, 76% received venetoclax, and 35% of B-myeloid MPAL patients received blinatumomab. Composite complete response was achieved in 50% of patients with B-myeloid and 64% with T-myeloid MPAL and 29% of the cohort underwent allogeneic hematopoietic stem cell transplantation (HSCT) in first remission. On univariate survival analysis and multivariate Cox analysis, the best survival outcomes were achieved in patients who received intensive therapy followed by allogeneic HSCT.

Relevance:

Optimal frontline therapy followed by HSCT leads to best outcomes in MPAL. Prospective studies to evaluate what is an optimal regimen for patients with MPAL are strongly warranted.

Data availability statement:

Data available from corresponding author on reasonable request