The Current Understanding of and Treatment Paradigm for Newly-Diagnosed TP53-Mutated Acute Myeloid Leukemia

About 10% of newly diagnosed and 20–30% of therapy-related acute myeloid leukemia (AML) harbors a TP53 mutation (mTP53-AML). Unfortunately, this biological subset predicts one of the worst prognoses among patients with AML, specifically a median overall survival of about 7 months with fewer than 10% of patients eventually cured of disease. Although remission rates appear to be increased with venetoclax-based, less-intensive regimens when compared with contemporary, intensive chemotherapy (55–65% vs. 40%), survival appears to be no different between the two approaches. Attempts to discern whether or not the prognosis of mTP53-AML is universally poor have centered around the study of concurrent cytogenetic risk and predicted TP53 allelic state, measurable residual disease status and the impact of conditioning intensity for patients proceeding to allogeneic hematopoietic stem cell transplantation. We discuss these considerations in this review and offer the current treatment approach to TP53-mutated AML.


Introduction
The TP53 gene, located on chromosome 17p13.1, encodes the transcription factor and tumor suppressor p53 whose cellular levels increase in response to deoxyribonucleic acid (DNA) damage among other cellular stressors; the consequent transcription of numerous p53 target genes, including those responsible for DNA damage repair, cell cycling and differentiation, induces cell cycle arrest and apoptosis [1]. Consequently, loss-of-function of this critical regulator via mutation and/or deletion threatens both cellular integrity and susceptibility to therapies targeted against cells harboring the abnormality. The plurality of mutations observed across all human tumor types are those within TP53 (approximately 40%) [2]. Approximately 10% of all new cases of AML are characterized by the presence of a TP53 mutation (mTP53-AML) [2][3][4]. Therapy-related AML is particularly enriched for this molecular subgroup of disease with TP53 mutations found in up to 25-35% of cases [5][6][7].
Given the critical importance and pleotropism of TP53 mutations and relatively unfavorable prognosis of most patients with AML, it is predictable that mTP53-AML is an informative example of why poor-risk disease is aptly named so. The presence of a TP53 mutation within AML has consistently predicted, irrespective of other relevant covariates like cytogenetic risk, a poor prognosis [8,9]. The median overall survival (OS) of patients with mTP53-AML is approximately 7 months when treated with standard-of-care therapies many of which will be discussed in this review [10][11][12]. Additionally, fewer than 10% of patients with mTP53-AML remain alive three years following allogeneic hematopoietic stem cell transplantation (alloHCT) performed while in complete remission (CR) [11,[13][14][15].
In this review we review the evolution of mTP53-AML management to its current form and also briefly discuss the novel treatment options that hold promise in light of these currently dismal outcomes.

Associated Factors and Important Considerations
Several factors associate with the presence of TP53 mutations in AML and bear relevance in appraising clinical studies or formulating a treatment plan for mTP53-AML. The only consistent patient-specific factor predicting a higher rate of TP53 mutated disease among patients with mTP53-AML appears to be age with patients with mTP53-AML being older than their TP53 wild type counterparts [8,16,17].
Disease-specific factors are more readily associated with the likelihood of detecting a TP53 mutation within AML. Interestingly, mTP53-AML appears to be associated with a lower burden of marrow blasts at diagnosis when compared with TP53 wild-type disease [13,16,17]. One-third of therapy-related AML (t-AML) harbors a TP53 mutation, which is detected in only 5-10% of patients with de novo disease [18][19][20]. In apparent contrast, the rate of TP53 mutations in AML with myelodysplasia-related changes (AML-MRC) may not be higher than that observed in de novo AML [6,16]. These observations are likely better explained by disease biology rather than histologic or historical factors, which are likely indirect surrogates. For instance, t-AML is more likely to be characterized by complex karyotype [21], defined as ≥3 acquired chromosomal aberrations, which might explain why the latter is more commonly seen in mTP53-AML (80-90%) [17,22]. Conversely, about 50-60% of AML with a complex karyotype will have a detectable TP53 mutation [8,17,22,23]. A higher rate of complex karyotype has been associated with increasing TP53 variant allele frequency (VAF) [16].
In contrast to the frequent cytogenetic abnormalities that are associated with mutated TP53, consistent co-mutations are not identified. In fact, mTP53-AML is noted to have statistically-significant decreased frequency of NPM1 (2-3%) and FLT3 mutations (2-7%) when compared with patients with TP53 wild type AML in which 30% of cases have these mutations, arguably the most commonly observed in AML [8,13,16,17,24]. The reasons for this lack of association are incompletely understood.
Several multivariable analyses after adjusting for age, performance status and other relevant disease-and patient-specific factors have demonstrated that even among patients with AML harboring poor-risk cytogenetics, the presence of a TP53 mutation predicts worse OS [8,9]. Similarly, among patients with mTP53-AML, the presence of high-risk cytogenetics like complex or monosomal karyotype have worse OS than patients with non-high risk cytogenetics [16,[25][26][27]. These observations have turned recent attention to TP53 allelic state in AML as patients with mono-allelic loss of TP53 in the pathobiologically similar disease myelodysplastic syndrome may have a prognosis similar to that with wild type TP53; bi-allelic loss of TP53 (predicted by a concurrent TP53 point mutation, deletion or copy neutral loss of heterozygosity) appears to drive the poor prognosis seen with the disease as a whole [28]. However, current techniques and resources do not allow consistent or expedient detection of these aberrations, which are often cryptic.
Retrospective analyses have shown that the burden of detectable TP53 appears to impact outcomes, specifically that a higher mTP53 variant allelic fraction (VAF) is predictive of a worse median OS [25,29]. This may be related to the observation that a higher mTP53 VAF is associated a higher rate of complex karyotype [16], which is more likely (approximately 70% of cases) to be associated with bi-allelic TP53 alteration predicting complete loss of TP53 protein function [17]. In an attempt to dichotomize a continuous variable, some groups identified specific VAF "cutoffs" to refine prognostic estimates. A mTP53 VAF > 40% predicts worse median OS when compared with a VAF ≤ 40% in patients treated for mTP53-AML (hazard ratio [HR] = 1.61, 95% confidence interval [CI]:1.17-2.21; p = 0.003) [25], suggesting that this VAF cutoff should be accounted for in efficacy analyses [25]. However, the impact of mTP53 VAF may depend upon the intensity of remission induction therapy. One of the largest retrospective analyses of 202 patients with mTP53 found that a mTP53 VAF cutoff of 40% was only predictive in intensivelytreated patients and not those treated less-intensively [25]. Conversely, other studies have shown no association at all between mTP53 VAF and median OS [16,24].

Intensive Therapy and Its Limitations
Intensive induction therapy has represented the standard-of-care for patients appropriate to receive it, is typically constituted by cytarabine and an anthracycline and has largely been agnostic of disease subgroups defined by cytogenetic or molecular abnormalities. Subgroup analyses of both retrospective and prospective studies have demonstrated the rates of CR of mTP53-AML to cytarabine + anthracycline intensive range from 28% to 34% [6,8,17,30]. Although some analyses have reported CR rates as high as 48% [16], a consistent statistically-significant difference in the rate of CR is observed between patients with mTP53-AML and AML with wild-type TP53 for whom the rate of CR is estimated to be 50-85% [8,16,17,30]. Furthermore, patients with mTP53-AML are observed to have a statistically-significant higher rate of primary induction failure/refractory disease when compared with their counterparts with TP53 wild-type disease (approximately 50% vs. 15-35%) ( Table 1) [17,30]. A randomized, multi-center, phase 3 trial compared CPX-351, the liposomal combination of cytarabine and daunorubicin, to a standard combination of cytarabine and daunorubicin (7 + 3) in older patients with newly-diagnosed AML-MRC or t-AML [31]. CPX-351 demonstrated superior rates of CR (38% vs. 26%, p = 0.036) [31]. Furthermore, CPX-351 was associated with superior median OS (9.6 vs. 5.9 months, p = 0.005) with more CPX-351-treated patients proceeding to alloHCT in first CR (20% vs. 12%) and an overall comparable safety profile [31]. Based on these data CPX-351 was approved by the Food and Drug Administration (FDA) for the treatment of patients with newly-diagnosed t-AML or AML-MRC. However, similar to the inferior rates of CR/CRi and survival for mTP53-AML patients treated with standard 7 + 3 induction, subsequent analyses demonstrated only a 29% CR/CRi rate for patients with mTP53-AML treated with CPX-351, in comparison to 66% for patients with TP53 wild-type disease (p = 0.0353) [6,11]. Real-world, multivariable analyses have also reaffirmed the lower response rates of mTP53-AML to CPX-351 (CR/CRi 41% vs. 66%, p = 0.04) [33]. Furthermore, post hoc analyses of the original phase 3 trial have shown that CPX-351 appears to be less effective in its ability to induce CR/CRi for patients with mTP53-AML when compared with those treated with classical 7 + 3 induction, although this was not statistically significant (29% vs. 40%; odds ratio [OR] = 0.62, 95% CI: 0.20-1.87) [6]. Real-world analyses have demonstrated similar results (Table 1) [32].
Improved rates of remission do not necessarily correlate adequately with outcomes, however, when specifically evaluating the mTP53-AML sub-population. The median OS of patients with mTP53-AML treated with CPX-351 on the original phase 3 trial was similar to those treated with 7 + 3 (4.5 vs. 5.1 months; HR = 1.19, 95% CI: 0.70-2.05) [6]. This is likely due to the fact that the remission durations and event-free survival (EFS) were similar between both arms and limited salvage options for this particular disease subset; however, a lesser proportion of patients treated with CPX-351 eventually proceeded to alloHCT (13% vs. 31%) [6].
The limited expectations with intensive therapy for mTP53-AML must be rectified with its expected toxicity. Patients who receive intensive therapy accept an appreciable risk of toxicity as well as treatment-related early death with the latter typically defined as that occurring within the first 30 days of therapy. No large, high-quality estimates of intensive therapy-attributable early mortality of patients with mTP53-AML exist, but disease biology is unlikely to influence this risk. The early mortality for all patients treated with CPX-351 on the phase 3 trial was 5.9%, increasing to 13.7% up to 60 days out and this compared favorably with 7 + 3-treated patients (10.6% and 21.2%, respectively), but did not reach statistical significance (p = 0.149 and p = 0.97, respectively) [31]. Real-world analyses of patients with newly-diagnosed AML treated with intensive induction have estimated a higher early mortality rate of approximately 15% and this increases to about 25% when evaluating older patients (age ≥65 years), who as previously discussed are more likely to have mTP53-AML [34]. The early mortality rate of intensively-treated older patients remains up to 25-30% when analyzing those treated in the largest clinical trials, which typically enroll less frail and more "fit" patients [35][36][37][38][39][40][41]. Age is an imperfect surrogate for intensive therapy appropriateness but often associates with decreased end-organ reserve and decreased performance status [42,43]. The latter correlates with higher post-intensive therapy early mortality with an ECOG PS of 3-4 predicting rates of approximately 50-60% [36,[44][45][46][47]. This must be taken into account when formulating a treatment plan for the patient with mTP53-AML.

The Evolution of Less-Intensive Therapy for mTP53-AML
All patients with AML, irrespective of age, should be considered for disease-directed therapy. However, the previously discussed risks that associate with increased chronological age make less-intensive induction therapy a more attractive option for the older and comorbidity-burdened patient. This may be even more so when considering the relatively unattractive risk:benefit ratio associated with intensive therapy for mTP53-AML.

Hypomethylating Agent Monotherapy
The hypomethylating agents (HMAs) azacitidine (AZA) and decitabine (DEC), since their availability in the mid-2000s, were the standard of care for the treatment of patients with AML who were inappropriate for intensive therapy [48]. A pre-specified sensitivity analysis of the data from the randomized phase 3 AZA-AML-001 trial found that AZA 75 mg/m 2 daily for 7 days improved the median OS of patients diagnosed with AML at age ≥ 65 years when compared with a "conventional care regimen (CCR)" comprised of either low-dose cytarabine (LDAC), intensive induction therapy or best supportive care (12.1 vs. 6.9 months; HR = 0.76, 95% CI: 0.60-0.96; p = 0.019) [49,50]. DEC was shown in a randomized, phase 3 trial to non-significantly increase median OS in a similar population of patients when compared with supportive care or LDAC (7.7 vs. 5.0 months, p = 0.108) [51]. As AZA-AML-001 was a positive study and included a more externally valid comparator group, AZA had generally been considered the optimal HMA for the treatment of the patient with AML and inappropriate for intensive therapy. However, this study was not designed to evaluate differences in outcome between patients treated with AZA and a specific "CCR." Furthermore, the majority of patients in the "CCR" arm received low-dose cytarabine monotherapy, an arguably ineffective therapy. Large, high-quality, population-based analyses and randomized trials, however, have found no difference in outcomes between AZA-and DEC-treated patients without accounting for TP53 status [52,53]. The CR/CRi rate of mTP53-AML to AZA is reported to be as low as 0% in randomized prospective trials [15], but as high as 40% in retrospective, single-center studies [13]. Subsequent analyses restricted to patients with mTP53-AML treated on AZA-AML-001 revealed that, as expected given the low rates and duration of remission, these patients have outcomes inferior to patients with TP53 wild-type disease, specifically a median OS of 7.2 months (vs. 12.0 months) for AZA-treated patients ( Table 2) [54]. DEC is traditionally administered at 20 mg/m 2 on a 5-day schedule (DEC5), but may have improved activity when extended to a 10-day schedule (DEC10). A retrospective study by Welch, et al. reported a 100% CR/CRi rate with DEC10 for mTP53-AML in addition to robust, albeit incomplete, mutation clearance and survival rates that seemed to remit to those observed in intermediate-risk AML populations [59]. However, other studies of DEC10 have not duplicated that striking response rate, affirmed the minority of patients with mutation clearance, and comparative studies of DEC10 or DEC5 for mTP53-AML have reported no differences in response rates between the two groups [51,[55][56][57][58]. The only randomized, prospective study comparing DEC10 to DEC5 reported a numerically higher rate of CR/CRi (47% vs. 29%, p = 0.40) and better median OS for patients with mTP53-AML treated with DEC10 (8.5 vs. 5.5 months, p = 0.55), however these differences were not statistically significant [62]. The risk of neutropenic fever and infection appeared to be higher in the DEC10 group [62]. It should be noted that the pre-treatment median mTP53 VAF was higher in the DEC10 arm (50.3% vs. 23.1%) and specifically above the 40% "cut-off" that has been identified as being of prognostic significance in the previously mentioned retrospective analyses; similarly the median mTP53 VAF of patients in the DEC5 arm was below the ascribed VAF with negative prognostic implications [25,55]. Given this shortcomings, the question whether or not DEC10 is more effective than DEC5 for mTP53-AML remains unanswered.

The Addition of Venetoclax
AML BCL-2 overexpression likely mediates apoptotic evasion and has been clinically associated with chemoresistance and ultimately poor survival [63,64]. Venetoclax, an oral, small molecule inhibitor of BCL-2 initially demonstrated robust leukemic cell kill in AML cell lines [65]. An initial non-comparative study of the combination of venetoclax with either AZA or DEC5 reported a CR rate of 37% (and CR/CRi rate of 67%) in the intention-to-treat population sufficient in 2018 to garner an FDA approval for the treatment of older or frail patients inappropriate for intensive therapy [10]. In sub-group analysis of patients with mTP53, the combination was associated with a CR/CRi rate of 47%, median duration of response of 5.6 months (95% CI: 1.2-9.4 months) and a median OS of 7.2 months (95% CI: 3.7 months-not reached) [10]. The subsequent confirmatory VIALE-A trial, which compared AZA monotherapy to AZA + venetoclax combination therapy in patients with newly diagnosed AML, demonstrated that AZA + venetoclax significantly prolonged the median OS in older patients with AML (median age 76 years, range 49-91 years) when compared with azacitidine monotherapy (14.7 vs. 9.6 months, p < 0.001)( Table 2) [15].
The presence of TP53 mutation predicts resistance, both primary and adaptive, to venetoclax-based therapy [60,66,67]. Although VIALE-A demonstrated a statisticallysignificant improved rate of CR/CRi with AZA + venetoclax combination therapy for mTP53-AML when compared with AZA monotherapy (55% vs. 0%, p < 0.001), an unplanned subgroup analysis of these 52 patients found no improvement in OS (HR = 0.76, 95% CI: 0.40-1.45) [15]. Retrospective analyses have also tempered enthusiasm for the addition of venetoclax to AZA for this specific patient population. Among patients with mTP53-AML and poor-risk cytogenetics, although associated with a higher rate of CR (20% vs. 11%), the addition of venetoclax to AZA does not appear to impact median OS (5.1 vs. 4.8 months). Recent real-world analyses have also demonstrated no survival advantage with venetoclax combination therapy over AZA monotherapy for patients with mTP53-AML and poor-risk cytogenetics, which are found in the majority of such patients, as previously discussed [68].
Given the equivocal "benefit" imparted by AZA + Ven for the treatment of mTP53-AML and the debate regarding the most favorable DEC monotherapy dosing schedule, the combination of DEC10 + venetoclax has been studied. A single-center, single-arm phase 2 trial of DEC10 + venetoclax in 118 patients, including 35 with mTP53-AML, reported 57% rate of CR/CRi among patients with mTP53-AML (vs. 77% for TP53 wild-type disease, p = 0.29) with 29% being measurable residual disease (MRD)-negative by flow cytometric analysis (MFC) [24]. However, median remission duration and OS was only 3.5 months and 5.2 months, respectively [24]. Additionally, when compared with the historical rates reported with DEC10 monotherapy, and despite a higher CR/CRi rate, DEC10 + venetoclax offered no apparent improvement in rate of MRD-negativity, time to remission, RFS or OS (Table 2) [24,55]. However, similar to AZA or DEC5, it is unclear if the addition of venetoclax to DEC10 imparts any true benefit relating to the depth, duration of remission and consequently overall outcome.

Allogeneic Hematopoietic Stem Cell Transplantation
Patients with poor/adverse risk AML are generally recommended to be considered for alloHCT as part of their consolidation strategy [69]. The exceedingly poor outcomes overall observed for patients with mTP53-AML when paired with the non-trivial risks associated with alloHCT call into question the unequivocal recommendation for consolidative alloHCT for this specific disease subset. However, a high-quality appraisal of the benefit associated with alloHCT have been hampered by the relatively uncommon rate of mTP53-AML among all AML and concerns regarding equipoise in conducting a randomized trial.
Earlier retrospective studies have been limited by a small number of patients with mTP53-AML proceeding to alloHCT, lack of dedicated subset analyses or the use of imperfect surrogates for loss of TP53 such as patients with chromosome 17p abnormalities [12,14,60,70,71]. Relatively large retrospective analyses of patients with disease harboring chromosome 17p abnormalities predicted to negatively impact TP53 function and proceeding to alloHCT have informatively shown two-or three-year OS rates that are as low as 10-15% with nearly all relapses occurring within six months post-alloHCT [70,71]. Some studies specifically evaluating patients with mTP53-AML have shown favorable outcomes at one-year post-alloHCT [12,60]. Other studies with slightly longer follow-up have demonstrated two-year rates of survival that essentially approximate 0% [8,54]. Comparative analyses attempting to describe any potential benefit to alloHCT are likewise limited, but a multivariable analysis of a retrospective cohort of 174 patients with mTP53-AML found that patients able to proceed to alloHCT in first remission had more favorable OS than their non-transplanted counterparts (HR = 0.28, 95% CI: 0.15-0.53; p = 0.001) [27].
The largest retrospective study of mTP53-AML to date reported on 98 patients, 18 of whom (55%) underwent alloHCT [16]. A significant improvement in EFS was noted for patients undergoing alloHCT when compared with those undergoing only chemo-consolidation or autologous stem cell transplantation (HR = 0.25, 95% CI: 0.11-0.58; p = 0.001) [16]. However, this is very likely confounded by selection bias. Although the poor median OS observed for patients with mTP53-AML may be attributed to the minority of patients proceeding to a seemingly definitive consolidative modality like al-loHCT, post-alloHCT outcomes still remain poor. Retrospective studies offering more granular descriptions of mTP53-AML populations proceeding to alloHCT have still demonstrated a median EFS or progression-free (PFS) survival of 5.0-7.5 months, even in studies with most patients receiving myeloablative conditioning [14,72]. Post-alloHCT relapse of mTP53-AML appears to herald forthcoming death as EFS approximates OS in this circumstance at approximately 8-10 months [14,72]. Only 5-10% of patients are alive at two years post-alloHCT (Table 3) [14,60,73].  [14] Abbreviations: AlloHCT, allogeneic hematopoietic stem cell transplantation; CIR, cumulative incidence of relapse; EFS, event-free survival; KT, karyotype; NRM, non-relapse mortality; MAC, myeloablative conditioning; OS, overall survival; PFS, progression-free survival; R, range; VAF, variant allele frequency.
The study of particular subgroups of patients within those diagnosed with mTP53-AML are likely to be important in identifying those truly benefitting from alloHCT. Patientspecific factors like age appear to impact the post-alloHCT outcomes of patients with mTP53-AML with some data suggesting a benefit with alloHCT for patient younger than 60 years at time of diagnosis [13,74]; one study reported no difference in outcomes for patients age ≥ 60 years and a striking difference in one-year OS based on this age cut-off (75% vs. 29%, p = 0.012), although the analysis was limited to 6 patients and may be confounded by selection bias [13]. Age, however, as previously-mentioned is likely just a reflection of fitness and comorbidity burden. A hematopoietic cell transplant-comorbidity index (HCT-CI) ≥ 4 and a Karnofsky Performance Status ≤ 80% at time of alloHCT predicts worse OS for patients with mTP53-AML [72]. These factors are also likely to influence the likelihood of a provider offering myeloablative conditioning, which is generally believed to increase the likelihood of cure [74,75]. Although studies with statistical comparisons are limited, patients with mTP53-AML clearly appear to have a higher rate of post-alloHCT relapse when compared with those receiving reduced-intensity/non-myeloablative conditioning [74]. This apparent difference needs to be rectified with the observed higher rate of alloHCT-related mortality [74].
Disease-specific features like cytogenetic status have been evaluated in this specific setting and appear to have prognostic impact. Post-alloHCT relapse rates among patients with mTP53-AML appear to be higher for those with poor-risk cytogenetics compared with those with intermediate cytogenetic risk [14,74]. mTP53-AML with complex karyotype predicts an inferior prognosis when compared with the absence of complex karyotype (p = 0.04), suggesting that patients with mTP53-AML and complex karyotype might not derive benefit from alloHCT; the presence of concurrent chromosome 17p abnormality, specifically, and thus likely bi-allelic loss, predicts worse post-alloHCT survival (8.0 vs. 13.3 months) [14,74]. Similarly indicative of bi-allelic TP53 loss, patients with a pre-alloHCT TP53 VAF > 60% are found to have a lesser chance at longer-term survival than their counterparts with a TP53 VAF ≤ 60% (2-year OS 0% vs. 25%; p = 0.20) (Table 3) [73]. Overall, the benefit of alloHCT for patients with mTP53-AML appears to be modest at best, but some sub-populations may benefit.

Formulating a Treatment Plan
Patients with newly diagnosed AML with poor-risk features including t-AML are shown to have a lower likelihood of receiving leukemia-directed therapy and this appears to be irrespective of age [76][77][78]. As no evidence suggests that mTP53-AML is more likely to present with acute disease-related complications that preclude appropriate consideration for therapy, a perceived unfavorable risk:benefit of therapy for a nearly incurable, aggressive disease may be responsible.
Given the almost universal poor prognosis and limited data to support an unequivocal frontline therapy, we ultimately recommend that patients with newly-diagnosed mTP53-AML be treated in the context of a clinical trial.

Clinical Trials
Fortunately, an increasing understanding of the mechanisms of TP53-associated chemoresistance as well as the cellular and marrow microenvironment features with which mTP53-AML is associated has offered promise for more effective therapies.
APR-246 or eprenetapopt is a molecule theorized to restore p53 protein to its wildtype conformation and in ex vivo experiments was shown to abrogate the chemoresistance of mTP53-AML to conventional chemotherapies as well as upregulate p53 protein levels [79][80][81]. An intention-to-treat analysis of the patients with oligoblastic mTP53-AML studied in a phase 2 study of APR-246 + AZA combination therapy demonstrated a median OS of 10.8 months [82]. An updated analysis with median follow-up exceeding two years also found that patients achieving CR/partial remission (PR) with TP53 mutational clearance prior to alloHCT had a median OS that was not reached [83]. A trial studying a triplet combination of APR-246 with AZA + venetoclax is ongoing (NCT04214860).
The targeting of CD47, or the transmembrane protein that interacts with macrophageexpressed signal-regulatory protein alpha, to increase macrophage-mediated phagocytosis of malignant cells has also shown promise for the treatment of mTP53-AML [84,85]. This is an enthusiastic area of clinical research as the surface of leukemic hematopoietic stem/progenitor cells (HSPCs) from primary AML patient blood and marrow samples overexpress CD47, thus promoting evasion from immunosurveillance, which in murine models is shown to be restored with anti-CD47 therapy [86][87][88]. The anti-CD47 monoclonal antibody magrolimab was associated with a favorable toxicity profile and a median OS of 12.9 months, the longest reported, in patients with mTP53-AML [89]. A phase 3 trial comparing magrolimab + AZA to AZA monotherapy (NCT04313881) as well as a phase 1b/2 trial of AZA + venetoclax + magrolimab triplet therapy (NCT04435691) are ongoing. Early results of the latter study demonstrate a 100% rate of CR/CRi in seven evaluable patients with newly-diagnosed mTP53-AML with a 57% rate of MFC-MRD-negativity; data the proportion of patients with mTP53-AML proceeding to alloHCT and long-term outcomes are eagerly awaited [90].
Other emerging therapies have provided optimism beyond mutant p53 protein refolding/reactivating or anti-CD47 therapies. The NEDD8-activating enzyme inhibitor pevonedistat led to four of five patients with newly-diagnosed mTP53-AML achieving CR/PR and a majority responding for >10 cycles at last follow-up [91]. Triplet therapy with pevonedistat and AZA + venetoclax led to CR/CRi/morphologic leukemia-free state in six out of eight patients (75%) with newly-diagnosed oligoblastic mTP53-AML, although median OS at last data cutoff was only minimally better than that expected with current therapies at 8.9 months [27]. Furthermore, the mTP53-AML marrow immune microenvironment appears to be one in an immunosuppressive state. HSPCs and leukemic blasts isolated from patients with mTP53-AML are found to overexpress the negative immune checkpoint regulators PD-L1 and TIM-3, thus hampering the T-cell-mediated anti-AML response [92][93][94][95][96][97]. Indeed, the mTP53-AML marrow is characterized by lesser infiltration by CD8+ cytotoxic T-cells, natural killer cells and helper T-cells as well as an expansion of regulatory T-cells, which promotes immune evasion [97]. Anti-PD-L1 or anti-TIM-3 therapies, either as monotherapy or in combination with standard-of-care backbones, to modulate this immunosuppressive state have been studied in unselected patients with activity in mTP53-AML to be determined by unplanned subset analyses (NCT02775903, NCT04150029, NCT04266301, NCT03946670, NCT03066648 and NCT03940352). Newer preclinical data also demonstrate that the combined inhibition of BCL-2 as well as the antiapoptotic BH3 protein MCL-1 may abrogate the resistance imparted by TP53 mutations harbored by AML and thus support the development and study of dual BCL-2/MCL-1 inhibitor therapy for mTP53-AML [98]. A more granular discussion on these ongoing research efforts is beyond the scope of this review.
It is our hope that ongoing and emerging therapies eventually establish a more effective standard of care for these patients. However, if a clinical trial is not available, providers are left to devise a treatment plan using current standard therapies and interpret the nuances of the data supporting them.

Intensive vs. Less-Intensive Induction
The poor outcomes observed among patients with mTP53-AML may be irrespective of therapy intensity. There have been no prospective, randomized trials comparing intensive with non-intensive therapies for patients with mTP53-AML and thus it is only feasible to use indirect comparisons using historical data at present. As previously-discussed, older patients with mTP53-AML treated with classical induction therapy including CPX-351 are predicted to only have a CR/CRi rate of 30-40% and median OS of only about 6-7 months [6,11,33,99]. Less-intensive options such as HMA + venetoclax are associated with CR/CRi rates of about 55-65% in mTP53-AML and this favorably, although indirectly, compares with intensive therapy [10,15,61]. The largest retrospective analysis of patients with mTP53-AML (n = 174) recently demonstrated no difference in the survival between patients treated with AZA + Ven when compared with other therapies like CPX-351 or 7+3 as well as no difference between patients treated with CPX-351 vs. other therapies [27]. However, one recent multivariable analysis of 95 patients with mTP53-AML demonstrated that patients treated with CPX-351 fared better than those treated with AZA + venetoclax when evaluating relapse-free survival (RFS)(HR = 0.37, 95% CI: 0.14-0.96; p = 0.04) and OS (OR = 0.41, 95% CI: 0.22-0.74; p = 0.003) [100]. The authors duly note that the survival advantage attributed to CPX-351 may be confounded by bias including the higher rate of alloHCT in CPX-351-treated patients likely indicating these patients were more fit with less comorbidity.
The uncertainty in whether efficacy differs based on therapy intensity implores providers to consider whether differences in toxicity or quality of life (QoL) exist between these two approaches. Like the efficacy evaluation, the absence of prospective, comparative trials limits the strength of any conclusion regarding toxicity/QoL. Indirectly, the 3-7% 30-day mortality rate associated with HMA + venetoclax approximately in older patients compares favorably to the 6-11% associated with modern intensive induction in similar populations [10,15,31]. The toxicity profiles between the two modalities are slightly different, as venetoclax combination therapies are typically continuous and may be associated with more protracted myelosuppression when indirectly compared with cyclic, but finite intensive therapy. Although early mortality may be no different for patients treated with AZA + venetoclax and CPX-351, the rate of febrile neutropenia and microbiologically-proven infection as well as length of stay appear to be higher for patients treated with CPX-351 [101]. QoL is another important consideration when evaluating the most appropriate remission induction therapy for a disease with a poor prognosis and therapies that are almost universally associated with at least some toxicity. Differences in QoL measures between patients treated intensively and with less-intensive therapies such as venetoclax combinations are lacking and cannot guide decision-making at the moment.
It is unclear if AZA + venetoclax represents a meaningful advance as patients with mTP53-AML garner only a 5-7 month median OS when treated with this therapy, illustrating the discordance between initial response and event-based outcomes for this subset of patients. Given the similarly poor OS observed for patients treated with AZA monotherapy and AZA + venetoclax therapy and the higher rate of protracted and deep myelosuppression associated with the latter, the older patient with mTP53-AML who is inappropriate to receive intensive therapy and alloHCT can be reasonably considered for HMA monotherapy. Deciding upon an induction strategy for the younger, intensive therapy appropriate patient may be more nuanced and requires accounting for the presence of concurrent high-risk features like concurrent monosomal/complex karyotype, TP53 VAF or mono-vs. bi-allelic TP53 loss as best predicted by current techniques. The more favorable survival observed for patients with mTP53-AML with TP53 VAF ≤ 40% and without additional high-risk cytogenetic features [16,25,26] may warrant more consideration for intensive induction, perhaps favoring 7 + 3 over CPX-351, for the younger patient with such disease and who is appropriate to receive intensive therapy including alloHCT with myeloablative conditioning.

Allogeneic Hematopoietic Stem Cell Transplantation
Beyond remission induction, the benefits of alloHCT in the treatment of mTP53-AML, as previously discussed, remain unanswered. The previously discussed available data, albeit limited, might suggest that patients in first remission who are able to proceed to alloHCT should do so. This consolidative strategy should be more strongly considered for the younger patient without disease harboring complex karyotype or predicted biallelic TP53 inactivation who can achieve MRD-negative remission and ultimately receive myeloablative conditioning. Patients with MFC-MRD positivity entering alloHCT may be accepting an exceedingly high rate of post-alloHCT relapse for the expected alloHCTrelated mortality risk with reduced-intensity/non-myeloablative condition as opposed to myeloablative conditioning. However, the decision must always be individualized with appropriate patient input regarding goals of treatment/care after review of what is known of the current nature and treatment of this disease. Post-remission, less-intensive therapy to deepen the response for patients in CR or convert MRD-positive to MRD-negative disease prior to alloHCT may be considered, but is not a practice guided by strong evidence. The role of post-alloHCT maintenance therapy is similarly unknown. Prospective, randomized studies are needed and are fortunately being planned to answer these questions.

Conclusions
The management of mTP53-AML represents one of the most critical areas of need within myeloid oncology with a slim minority of patients likely cured of disease. Although intensive therapies have traditionally been the backbone of remission induction for AML including mTP53-AML, the advent of novel and targeted therapies have offered some promise in the ability to induce similar if not better rates of remission with lesser toxicity and treatment-related mortality. Despite this apparent progress, improvements in survival have not yet been realized. Additional novel therapies and perhaps combinations of novel agents with other novel agents are needed to help rectify the disconnect observed between improved rates of remission and prolonged survival for patients with mTP53-AML. For these reasons, we recommend frontline clinical trial for patients with newly-diagnosed mTP53-AML. The benefits of alloHCT in the management of mTP53-AML are unclear and require a more nuanced accounting for relevant covariates like allelic state, MRD status, conditioning intensity among others. For these reasons, dedicated clinical trials for patients with mTP53-AML are needed and should be the preferred option for the newly diagnosed