Prognostic heterogeneity and clonal dynamics within distinct subgroups of myelodysplastic syndrome and acute myeloid leukemia with TP53 disruptions.

INTRODUCTION

A decade has elapsed since comprehensive sequencing efforts from The Cancer Genome Atlas revealed that TP53 mutations are recurrent in acute myeloid leukemia (AML), with a mutational frequency of 8% [ 1 ]. AML and myelodysplastic neoplasms (MDS) harboring TP53 mutations are associated with exceptionally poor survival, typically on the order of weeks to months [ 2 , 3 , 4 , 5 , 6 , 7 ]. Specifically, very recent efforts that have clarified this genetically defined subgroup include the IPSS‐M model for MDS, the European LeukemiaNet (ELN) 2022 classification for AML, and the 2022 International Consensus Classification (ICC) [ 8 , 9 , 10 ]. The IPSS‐M model has shown an adjusted hazard ratio (HR) of 3.27 (the highest of all) for multihit TP53 aberration; the ELN 2022 classification now includes a new category termed “AML with mutated TP53 ” for patients with TP53 variant allele frequency (VAF) ≥ 10%; and the 2022 ICC has designated a unique category for MDS and AML with TP53 mutation at the height of the myeloid neoplasm classification hierarchy.

Our group has recently explored the basis for these distinctly adverse outcomes in TP53 ‐aberrant MDS and AML using clonal dynamic modeling, which suggested that TP53 ‐aberrant clones drive malignant hematopoiesis and constitute measurable residual disease (MRD) for most patients, which in turn contributes to relapse [ 3 ]. Genomic aberration affecting the TP53 locus is often associated with disruption of other cellular machinery including regulators of the epigenome, the spliceosome complex, and the cohesin complex [ 3 ]. However, our study size was relatively small, and our analysis was conducted under the assumption that all TP53 ‐mutant MDS and AML are equal (a homogeneous population). The population of patients with TP53 disruptions is more heterogeneous than current classification and prognostication systems suggest, and data on prognostic discrimination within this heterogeneous population are limited. The distinction between TP53 ‐mutant MDS versus AML using the time‐honored World Health Organization (WHO) blast cutoff of 20% is arbitrary, and very recent studies have shown negligible differences in both genetic profiles and clinical outcomes between MDS and AML [ 11 ].

MATERIALS AND METHODS

Patient subjects

UMass Institutional Review Board approval was obtained via protocols H00019393 (“AML Database”) and H00013131 (“Blood & Marrow Banking”; NCT01174615 ), as previously described [ 3 ]. The UMass Leukemia Registry identified 76 patients harboring one or more aberrations in TP53 plus with an ICD‐10 code of either D46.9 (MDS, including its subentities) or C92.00 (AML, including its subentities) between 2011 and 2023. TP53 aberrations were defined as mutations within the TP53 locus (missense, frameshift deletion, splice site deletion, in‐frame deletion, frameshift insertion) or large‐scale disruptions of 17p (deletion, addition, translocation, isochromosome, duplication, inversion, dicentric chromosome, ringed chromosome), confirmed by fluorescence in situ hybridization (FISH) to have TP53 deletion.

Cytogenetic analyses

Bone marrow cells from patients with TP53 aberrations were cultured, and metaphase cells were harvested at 24 and 48 h. For these 76 patients, conventional karyotyping by G‐banding using trypsin and Giemsa was performed on metaphases. For each specimen, at least 20 (or all available) metaphase cells were analyzed for numerical and structural chromosomal anomalies. Criteria defined by the International System for Human Cytogenetic Nomenclature were used to describe abnormal clones. FISH was performed for TP53 deletion assessment using the Vysis LSI TP53/CEP17 probes (Abbott Molecular).

Profiling of TP53 somatic variants and allelic states

Genomic DNA was purified from bone marrow aspirates, as previously described [ 3 ]. A library was created using multiplex PCR targeting oncogene “hot spot” mutations and full exon sequencing (CTMPv3 panel). PCR amplicons were isolated, followed by the ligation of a sequencing linker, and a barcode for patient identification. Real‐time PCR was used to quantify the library. A sequencing template was prepared by diluting the library, and performing emulsion PCR to amplify and link the amplicon to a sequencing bead. The sequencing template was loaded onto an IonTorrent PGM next‐generation sequencer. For TP53 exome sequencing, the entire coding region from codons 1–393 of isoform A was covered, with coverage depth ranging from 500× to 6500× (Supporting Information Table S1 ). Patients were stratified by the presence or absence of multihit status and by TP53 VAF. Multihit status was defined as per the 2022 ICC: two or more distinct TP53 mutations with VAF ≥ 10%, or a single TP53 mutation plus one of the following abnormalities: (1) TP53 deletion, (2) TP53 VAF > 50%, (3) copy‐neutral loss‐of‐heterozygosity, or (4) any complex karyotype [ 10 ]. Monoallelic TP53 status was defined as non‐multihit.

Clonal dynamic modeling and copy number variation (CNV) analysis

Clonal and subclonal structures for each patient with available serial bone marrow aspirates were modeled via R (R Core Team, 2013) and R packages “circlize,” “timescape,” and “BioCircos” as previously described [ 3 , 12 ]. VAF for each mutant gene was adjusted for gene copy number variation (CNV) in the event of copy number gain or loss. VAF was defined as the ratio of mutant read number to total read number (total reads = mutant reads plus wild‐type reads). CNV was performed on sequencing data for samples containing ≥20% tumor load and analyzed on IonXpress. CNV annotation against VAF was under the assumption that, for a given bone marrow sample, mutations with VAFs in a similar range likely occur in the same clone.

Hierarchical clustering for VAFs for TP53 and cooperating mutations

Unsupervised hierarchical clustering of VAFs for TP53 and co‐occurring mutations was performed by the R package “pheatmap.” Clades were generated based on type and VAF for co‐occurring mutations. VAF data were analyzed using IonTorrent Variant Caller and Softgenetics Next Gene software packages using GRCh37_3 (Ver HG19) as the reference sequence.

Multivariate analysis for survival

Cox proportional hazard models were used to analyze the binary dependent variable of interest (alive vs. deceased). HR and 95% confidence intervals were reported for each model. Variables were removed one by one from the adjusted model based on the statistical significance in the model. Age was kept in the model despite not being statistically significant in the adjusted model because of its clinical association with mortality. Analyses were conducted using SAS software (version 9.4; SAS Institute, Cary, NC). α was set as 0.05 for all tests.

Other statistical analyses

Pearson's chi‐squared test for significance was used to calculate observed vs. expected frequencies for sex and ethnicity distributions. A Paired student's t ‐test assuming heteroscedasticity was performed to assess differences in fold changes in TP53 VAF with treatment. Log‐rank tests (nonparametic) were performed to compare survival distributions of subsets of TP53 ‐aberrant MDS and AML.

RESULTS

Demographics and baseline features of TP53 —Aberrant MDS versus AML

We queried the laboratory information system and the UMass Leukemia Registry for patients with a diagnosis of MDS or AML between 2011 and 2023. A total of 598 patients were identified. Of the 598 patients, 76 (12.7%) harbored TP53 aberrations. Ten patients had disruption of the TP53 locus without an intragenic mutation. We assessed baseline features based on morphologic classification into MDS and AML, as defined by the WHO Revised 4th Edition [ 13 ]. Of these 76 patients, 35 (46.1%) had MDS and 41 (53.9%) had AML. Of the 35 patients with MDS, 9 (25.7%) had therapy‐related MDS (the most prevalent subtype of MDS). Of the 41 patients with AML, 17 (41.5%) had AML, not otherwise specified (NOS) and 14 (34.1%) had AML with myelodysplasia‐related changes (AML‐MRC), which were the two most prevalent subtypes of AML (Table 1 ). The majority of patients with therapy‐related myeloid neoplasms or multihit states also had a complex karyotype.

The median age at diagnosis for MDS was significantly higher than that of AML: 76.2 years (range: 41.6–93.4) versus 67.3 years (range: 46.3–89.5), p = 0.0079 (Table 1 ). There was a nonuniform distribution between sexes; most were males ( p = 0.0476). Ethnicity representation among patients with TP53 disruptions was significantly different compared to that of the reference population from the 2020 Worcester County Census ( p = 0.0289) ( www.census.gov ). Most patients were non‐Hispanic White. Patients with MDS and AML had no significant differences in median WBC and hemoglobin values at diagnosis, but patients with AML had lower median platelet count compared to those with MDS (Table 1 ).

Subtypes of TP53 aberrations in MDS versus AML

We assessed subtypes of TP53 aberrations in WHO‐defined MDS vs. WHO‐defined AML for 76 patients with TP53 ‐aberrant MDS or AML. The most prevalent subtype of aberration was the missense mutation (Figure 1A ). There were no significant differences in the frequencies of these aberrations between MDS and AML. We assessed all gross structural abnormalities found in patients with TP53 aberrations, then segregated the abnormalities by MDS versus AML (Figure 1B ). The structural genomic landscape from aggregate analysis did not appear different between MDS and AML: eight major subtypes of structural disruptions were observed in both MDS and AML, dispersed among nearly all chromosomes. We then assessed the topographic distribution of intragenic TP53 disruptions in the open reading frame for TP53 ‐encoding p53 isoform A for MDS (Figure 1C ) and AML (Figure 1D ). Among 30 unique TP53 intragenic mutations in patients with MDS, 28 (93.3%) localized to the DNA‐binding domain, with only 1 mutation (3.3%) each in the transactivation domain and the proline‐rich domain. Among 35 unique TP53 intragenic mutations in patients with AML, 33 (94.3%) localized to the DNA‐binding domain, with only 1 mutation (2.9%) each in the transactivation domain and the basic domain.

Clonal combinations among varying TP53 allelic states of MDS versus AML

We next assessed the co‐occurring mutational profiles for MDS versus AML with TP53 aberrations based on VAFs for mutations from diagnostic bone marrow specimens (Figure 2A, B ). Full exon sequencing was performed for known recurrent somatic mutations. There were 16 distinct co‐occurring mutations among patients with TP53 ‐aberrant MDS (Figure 2A ) and 20 distinct co‐occurring mutations among patients with TP53 ‐aberrant AML (Figure 2B ).

Unsupervised hierarchical clustering showed a largely dichotomous distribution with respect to co‐occurring mutational profiles for both MDS and AML (Figure 2A, B , dendrograms ). The clades with high versus low TP53 VAF showed no obvious differences in co‐occurring mutations, and there was no significant difference in the dendrograms of co‐occurring mutations for MDS versus AML. Therefore, we mapped co‐occurring mutational profiles as a function of TP53 allelic state (monoallelic vs. multihit) and absolute VAF (≤50% or >50%) for MDS (Figure 2C ) and AML (Figure 2D ). The frequency of multihit status was higher for AML compared to MDS (73.2 vs. 60.0%), suggesting that multihit status correlates with more aggressive myeloid neoplasm. The frequency of patients with multihit status plus VAF >50% was also higher for AML compared to MDS (36.6 vs. 17.1%).

There was no significant difference in the mean TP53 VAF for patients with AML (46.8 ± 5.3%) versus MDS (35.0 ± 4.8%) ( p = 0.0571) (Figure 2E ). There was no significant difference in the subclonal diversity for MDS versus AML (Figure 2F ).

Prognostic discrimination within a heterogeneous population of TP53 —aberrant MDS and AML

We next assessed overall survival (OS) for varying TP53 allelic states and mutational burdens (VAFs) for MDS and AML. We identified 68 patients who met the 2022 ICC definition for myeloid neoplasm with mutated TP53 (after exclusion of patients with VAF < 10%) and stratified patients based on hit status and VAF. Median follow‐up was 214 days (very short due to the deaths of patients). Median OS was not significantly different among patients with monoallelic TP53 ‐mutant MDS, multihit TP53 ‐mutant MDS, and monoallelic TP53 ‐mutant AML (median OS: 439 days vs. 376 days vs. 439 days, respectively) (Figure 3A ). In contrast, patients with multihit TP53 ‐mutant AML had distinctly shorter survival (median: 64 days) (Figure 3A ).

We then assessed OS for patients in the multihit subgroup only ( n = 51) (Figure 3B ). We substratified patients based on TP53 VAF, given that VAF > 50% confers an obligatory biallelic disruption within the multihit subgroup. OS for patients with MDS with VAF ≤ 50% and VAF > 50% was not significantly different (median: 376 days vs. 435 days, respectively) (Figure 3B ). Patients with AML with TP53 VAF > 50% had significantly inferior OS (median: 46 days) (Figure 3B ). OS was also analyzed for the entire cohort (Supporting Information Figure S1A ) and also as a function of disease category (Supporting Information Figure S1B ), karyotype (Supporting Information Figure S1C ), and antecedent chemotherapy exposure (Supporting Information Figure S1D ).

Unadjusted survival models showed significant associations with mortality in multivariate analysis (Figure 3C ). Patients with AML had 2.5 times the risk of death as those with MDS. Death was predominantly due to disease progression. Patients with a complex karyotype were at 3 times the risk of death than those without it. Patients with multihit status were 2.3 times at risk of death as compared to those without. Those who did not receive hematopoietic stem cell transplant (HCT) were at 3.9 times the risk of death as compared to HCT recipients. The hazard was analyzed for the following continuous variables: age, TP53 VAF, blast percentage at diagnosis, and CD34 + cell percentage (Figure 3D ). For every absolute increase by 10% in TP53 VAF, blast percentage at diagnosis, and CD34‐ cell percentage, the risk of death increased by 20, 10, and 20%, respectively. We also assessed OS as a function of TP53 VAF as a continuous variable, regardless of the MDS versus AML designation, and noted an inverse correlation between OS and TP53 VAF for patients with complex karyotypes (Supporting Information Figure S2A ). There was also an inverse correlation between OS and age at diagnosis (Supporting Information Figure S2B ).

The results from the adjusted model also showed an association between mortality and AML, HCT status, and change in TP53 VAF (Supporting Information Table S2 ). Patients with AML had 3.3 times the risk of death as those with MDS while holding constant HCT status, TP53 VAF, and age. Those without HCT were at 4.3 times the risk of death than HCT recipients, while holding constant AML and MDS status, TP53 VAF, and age. For every absolute increase in TP53 VAF by 10%, the risk of death increased by 20% while holding constant AML and MDS status, HCT status, and age (Figure 3D ).

There was no difference in survival for t ‐MN versus non‐ t ‐MN (Figure 3C and Supporting Information Figure S1D ), suggesting that TP53 aberration is a confounding variable with regard to the association between t ‐MN and increased risk for death (for patients with TP53 ‐aberrant MDS or AML).

Clonal and subclonal dynamics as a function of therapeutic intervention

Our prior studies suggested that inferior survival for patients with TP53 disruptions was associated with the persistence of TP53 ‐mutant clones [ 3 ]. However, we previously did not discern clonal dynamics between disease designations (MDS vs. AML) or TP53 allelic states. We therefore modeled the clonal dynamics of heterogeneous subsets of TP53 ‐aberrant MDS versus AML (with varying TP53 allelic states) as a function of therapeutic intervention (Figure 4 ). VAFs were adjusted for CNV, as shown in the representative pictogram (Supporting Information Figure S3 ).

For patients with TP53 ‐aberrant MDS, 17 patients underwent serial bone marrow biopsies during treatment. Of these, 11 (64.7%) had multihit status (Figure 4 , yellow backdrop ). Eight patients (47.1%) underwent HCT, and 7 of these 8 patients had received HMA‐based front‐line therapy (Figure 4 , green backdrop ). For patients with MDS, we assessed the effect of HMA‐based treatment on TP53 VAF. The mean TP53 VAF (sum of VAFs) was significantly lower after HMA‐based treatment (35 vs. 12.6%) ( p = 0.0075) (Figure 5A ).

For patients with TP53 ‐aberrant AML, 19 patients underwent serial bone marrow biopsies during treatment. Of these, 14 (73.7%) had multihit status (Figure 4 , yellow backdrop ). Only two patients (10.5%) underwent HCT (Figure 4 , green backdrop ). Five patients (26.3%) received araC‐based induction, while 11 patients (57.9%) initially received HMA‐based induction. Notably, 100% of araC‐based therapy recipients eventually went on to receive HMA later in their treatment course. Only one patient with multihit TP53 underwent HCT, and this patient had previously received HMA plus venetoclax. For patients with AML with multihit TP53 , we assessed the effect of HMA‐ versus araC‐based front‐line therapy on TP53 VAF. HMA therapy led to a mean fold reduction in TP53 VAF (pre‐ to post‐treatment) of 11.7 ± 5.8 ( p = 0.0189), while araC‐based therapy actually led to a mean fold increase (but not statistically significant) in TP53 VAF (pre‐ to post‐treatment) of 2.6 ± 1.05 ( p = 0.3186) (Figure 5B ). HMA‐based treatment resulted in a significantly greater reduction in the VAF in aberrant clones and subclones compared to araC‐based treatment. Some patients were started on first‐line therapy after diagnosis but were unable to proceed with serial bone marrow biopsies, mostly due to death prior to response assessment (Supporting Information Figure S4 ).

Clinical fate and post‐transplant outcomes

We assessed the clinical trajectory of patients with TP53 ‐aberrant MDS versus AML based on key management decision points including decisions about curative versus palliative intent therapy and candidacy for HCT (Figure 6A ). Compared to AML, most patients with MDS (62.9%) proceeded with palliative intent therapy at the time of diagnosis, largely due to their older age and suboptimal performance status at the time of diagnosis. Most patients with AML (53.7%) proceeded initially with curative intent, in contrast to MDS. However, only 14.6% of patients with AML eventually proceeded with HCT (despite curative intention at diagnosis), largely due to clinical decompensation and partly due to disease progression during front‐line therapy. Among HCT recipients for MDS ( n = 10), 7 (70%) remained in remission (Figure 6B ). The best long‐term outcomes were seen in patients with low TP53 VAF. Among HCT recipients for AML (n = 6), 4 (66.7%) remained in a morphologic leukemia‐free state (remission) (Figure 6C ). Median OS was 734 days for transplant recipients.

DISCUSSION

To date, multiple studies have shown the impact of TP53 mutations on survival, though most studies evaluated this genetically defined subset as a homogeneous group, in comparison to other mutation‐specific groups [ 2 , 14 , 15 ]. Bernard et al. showed survival implications for multihit TP53 ‐mutant MDS, which we were unable to reproduce perhaps due to our smaller sample size [ 2 ]. In our current study, we noted significant heterogeneity with respect to genomic states and outcomes for patients with TP53 disruptions. Although patients with MDS and AML shared common gross structural aberrations and intragenic TP53 aberrations (Figure 1 ), there was substantial diversity in each patient's allelic state, response to front‐line therapy, HCT candidacy, and ultimately clinical outcomes. Our data enhances upon prognostic discrimination within this heterogeneous group of patients; such heterogeneity within this population was mostly a function of the genomic state of TP53 for each patient, with less impact of morphologic state or clinical disease label. Our data suggests that the boundary between classification as MDS or AML remains indistinct, but a functional and clinically relevant distinction is observed with varying TP53 genomic status. Importantly, among all patients with AML (single‐hit or multihit), only one patient remained alive at the longest time to follow‐up (∼5.5 years).

Our data are consistent with findings by Grob et al. showing no significant differences in molecular characteristics between MDS with excess blasts and AML [ 11 ]. Regarding survival, Grob et al. suggested that AML and MDS with excess blasts do not differ with respect to survival [ 11 ]. In contrast, our findings suggest significant prognostic heterogeneity between MDS and AML when patients are stratified based on allelic state and VAF. Within the multihit category, we found that the most inferior outcomes are seen in patients with AML with VAF >50%, which confers obligatory biallelic disruption in at least some percentage of cells. Early determination of allelic state and VAF may also guide optimal front‐line therapeutics, especially when treatment is given with curative intent.

There is limited literature on clonal and subclonal diversity in a heterogeneous group of patients with TP53 aberrations. Such clonal considerations are important from the standpoint of therapeutics, as certain interventions may only eliminate a fraction of cells within a given patient. We identified a distinct clonal architecture within each patient, and no two patients had the same disease features. In our cohort, HMA‐ and araC‐based regimens had differential effects on various subclones (Figure 4 ). HMA was more effective in reducing the mutant TP53 burden compared to cytotoxic induction therapy, and more patients were able to proceed with HCT if they received first‐line HMA‐based therapy. In fact, araC‐based therapy increased the TP53 VAF, which was likely due to selection pressure in favor of TP53 ‐mutant cells. We did not observe improved OS with the addition of venetoclax ( data not shown ): this may be related to a higher risk for neutropenia and subsequent infections with venetoclax, which can ultimately impair HCT candidacy (Figure 6 ). This is consistent with prior reports on the outcomes of TP53 ‐mutant myeloid neoplasms after treatment with venetoclax [ 16 ]. Prospective trials may be needed to identify optimal front‐line regimens that lead to the best long‐term outcomes.

We did not observe progression from monoallelic to multihit state for any patients: the contributing reasons were that patients died (thus no subsequent bone marrow aspirate showed biallelic state), and patients received treatment (HMA) which blocked the progression and reduced the TP53 VAF. For example, if a patient had a single mutation (monoallelic) and received HMA, the VAF decreased. If hypothetically there was a group of monoallelic patients who were untreated but lived long enough to follow with serial bone marrow aspirates, then we might capture them converting into biallelic status. For the patients with multihit status, all were biallelic at a time of diagnosis. We can presume that all multihit patients started out as monoallelic but were not captured and diagnosed as monoallelic. The ideal in vitro study to address clonal dynamics is to establish a fresh set of monoallelic hit cells and then follow those cells longitudinally to assess for conversion to biallelic hit. This can also be tested in patient‐derived xenografts.

One limitation to our study is the lack of single‐cell resolution [ 12 ]. Our data infers clonal composition from VAFs, CNV, and cytogenetics [ 17 ]. The presumption in our study is that if VAF is above 50%, then at least a fraction of the cells has obligatory biallelic disruption. For example, if VAF was 51%, then we can guarantee that at least one cell in that bone marrow has a biallelic disruption (there might be more than one cell with a biallelic disruption, but there is at least one). If VAF is 50% or less, then there are two possible scenarios: (A) every cell has monoallelic disruption at most (none are biallelic), or (B) some cells have biallelic disruption and some have monoallelic disruption.

Another limitation of our study is the varying time to clinical presentation for each patient: the disease likely developed months to years prior to clinical detection, and this can skew the analysis from these diagnostic bone marrow specimens. Lead‐time bias may be present for patients with more indolent disease courses (such as patients with monoallelic states and low TP53 VAF).

A crucial question continues to revolve around curative intent therapy for this heterogeneous subset, as long‐term survival is seen more often in HCT recipients compared to non‐HCT recipients. Long‐term survival may depend on HCT candidacy from the time of diagnosis, though one's performance status is not static and may depend on the type of front‐line therapy received. Since HCT is a time‐dependent covariate and subject to selection bias, we recognize that our data may not objectively reflect a favorable effect of HCT as an independent variable, as patients who were selected for HCT were likely healthier at baseline than those who were not selected for HCT. Our data supports the use of HMA‐based therapy as front‐line therapy to help reduce the mutant TP53 allele burden, followed by HCT for patients who remain candidates for HCT. HCT affords durable remission lasting beyond 1 year in most cases (Figure 6B, C ). HCT also appears to be the most effective intervention to lead to MRD negativity. Furthermore, we acknowledge another limitation to our study: it is difficult to mix survival studies between patients treated or untreated, as well as those who are more or less intensively treated, or even those who have received an allograft.

Translational efforts have focused on targeting TP53 ‐mutant cells; such agents include the p53 reactivator eprenetapopt (APR‐246), the nutlin analogs, magrolimab, and sabatolimab [ 18 , 19 , 20 , 21 , 22 , 23 , 24 ]. However, there are no Food & Drug Administration (FDA)‐approved medications to date, and there are no consensus guidelines for definitive management of TP53 ‐mutant MDS and AML [ 25 ]. Recent clinical trials of TP53 mutation‐directed therapy have not met primary endpoints in Phase III settings [ 18 ]. Single‐center studies have shown that HMA with or without venetoclax has some efficacy and may serve as a bridge to HCT, though the median OS is still very short [ 3 , 6 , 16 , 26 ]. We note the enrichment of TP53 mutations in non‐Hispanic White patients, and racial disparities in therapeutic approaches are also worthy of consideration [ 27 ]. Further risk stratification using details of the TP53 genomic state in conjunction with morphologic assessment may guide tailored therapeutics and personalized treatment plans for this heterogeneous group [ 28 , 29 , 30 , 31 , 32 , 33 , 34 ].

AUTHOR CONTRIBUTIONS

SAP and JMG conceived the study. SAP, WKG, and BT wrote the manuscript. JC, AGS, MR, SS, AH, LH, SK, XM, PM, WS, BW, and JMG contributed to data acquisition. SAP, JC, AE, JF, BB, KS, and JMG contributed to critical analysis and revisions. SAP and JMG gave final approval.

CONFLICT OF INTEREST STATEMENT

SAP served on the COMMANDS advisory board and the Acute Myeloid Leukemia advisory board for Bristol Myers Squibb and served on the Multiple Myeloma advisory board for Pfizer. JC served on advisory boards for Jazz Pharmaceuticals, Bristol Myters Squibb, MERIT CRO, Pfizer, and Amgen; served as a Data Safety Monitoring Board Member with ICON‐AlloVir and ICON‐Prolacta; owns stock in Actinium Pharmaceuticals, Bluebird Bio/2Seventy, Cellectar Sciences, Dynavax Technologies, aTyr Pharma, Gamida Cell Ltd, Novavax Inc, Ovid Therapeutics, Sorrento Therapeutics, TG Therapeutics, Vaxart, and Veru. JMG serves on the Advisory Board for Novartis and holds US Patent No. 9 012 215, US Patent No. 10 222 376, and US Patent No. 11 209 435.

FUNDING INFORMATION

SAP receives research funding from the UMass Center for Clinical and Translational Science (CCTS) Pilot Project Program grant (NIH/NCATS Grant UL1TR001453). AE was supported by the National Cancer Institute (NCI) award F32CA250304, the Advanced Residency Training Program at Stanford, and the American Society of Hematology Scholar Award.

ETHICS STATEMENT

Approval was obtained through the UMass Institutional Review Board (IRB) via protocols H00019393 (“AML Database”) and H00013131 (“Blood & Marrow Banking”; NCT01174615 ).

PATIENT CONSENT STATEMENT

Approval was obtained through the UMass Institutional Review Board (IRB) via protocols H00019393 (“AML Database”) and H00013131 (“Blood & Marrow Banking”; NCT01174615 ).

CLINICAL TRIAL REGISTRATION

The authors have confirmed clinical trial registration is not needed for this submission.

Supporting information

FIGURE S1 . OS for all 76 patients with TP53 aberrations (A) . OS for subgroups after stratification based on disease label (B) , karyotype (C) , and history of chemotherapy exposure (D) . Log‐rank p values are shown.

Click here for additional data file.

FIGURE S2. (A) OS as a function of TP53 VAF, stratified by complex karyotype ( blue ) versus noncomplex karyotype ( pink ). (B) OS as a function of age at diagnosis, stratified by complex karyotype ( blue ) versus noncomplex karyotype ( pink ). Coefficients of correlation are shown.

Click here for additional data file.

FIGURE S3 . Representative pictogram of CNV analysis for a patient. Green probe dots represent significant gains. Red probe dots represent significant deletions. Total coverage for sample and control is shown.

Click here for additional data file.

FIGURE S4 . Clonal landscape for patients with only one bone marrow biopsy (diagnostic sample) available. These patients were not able to proceed with subsequent bone marrow biopsies due to death or lack of follow‐up.

Click here for additional data file.

Supporting Information

Click here for additional data file.

Supporting Information

Click here for additional data file.

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available in Supporting Information of this article.

REFERENCES

1 Ley TJ , Miller C , Ding L , Raphael BJ , Mungall AJ , Robertson A , et al. Genomic and epigenomic landscapes of adult de novo acute myeloid leukemia . N Engl J Med . 2013 ; 368 : 2059 – 2074 . 23634996 10.1056/NEJMoa1301689 PMC3767041

2 Bernard E , Nannya Y , Hasserjian RP , Devlin SM , Tuechler H , Medina‐Martinez JS , et al. Implications of TP53 allelic state for genome stability, clinical presentation and outcomes in myelodysplastic syndromes . Nat Med . 2020 ; 26 : 1549 – 1556 . 32747829 10.1038/s41591-020-1008-z PMC8381722

3 Patel SA , Lloyd MR , Cerny J , Shi Q , Simin K , Ediriwickrema A , et al. Clinico‐genomic profiling and clonal dynamic modeling of TP53‐aberrant myelodysplastic syndrome and acute myeloid leukemia . Leuk Lymphoma . 2021 ; 62 : 3348 – 3360 . 34496723 10.1080/10428194.2021.1957869

4 Bewersdorf JP , Shallis RM , Gowda L , Wei W , Hager K , Isufi I , et al. Clinical outcomes and characteristics of patients with TP53‐mutated acute myeloid leukemia or myelodysplastic syndromes: a single center experience . Leuk Lymphoma . 2020 ; 61 : 2180 – 2190 . 32362171 10.1080/10428194.2020.1759051 PMC7603787

5 Hunter AM , Komrokji RS , Yun S , Al Ali N , Chan O , Song J , et al. Baseline and serial molecular profiling predicts outcomes with hypomethylating agents in myelodysplastic syndromes . Blood Adv . 2021 ; 5 : 1017 – 1028 . 33591325 10.1182/bloodadvances.2020003508 PMC7903224

6 Zeidan AM , Bewersdorf JP , Hasle V , Shallis RM , Thompson E , de Menezes DL , et al. Prognostic implications of mono‐hit and multi‐hit TP53 alterations in patients with acute myeloid leukemia and higher risk myelodysplastic syndromes treated with azacitidine‐based therapy . Leukemia . 2022 ; 37 : 240 – 243 . 36437356 10.1038/s41375-022-01766-z

7 Short NJ , Montalban‐Bravo G , Hwang H , Ning J , Franquiz MJ , Kanagal‐Shamanna R , et al. Prognostic and therapeutic impacts of mutant TP53 variant allelic frequency in newly diagnosed acute myeloid leukemia . Blood Adv . 2020 ; 4 : 5681 – 5689 . 33211826 10.1182/bloodadvances.2020003120 PMC7686900

8 Bernard E , Tuechler H , Greenberg PL , Hasserjian RP , Arango Ossa JE , Nannya Y . Molecular international prognostic scoring system for myelodysplastic syndromes . NEJM Evidence . 2022 ; 1 : 1 – 14 . 10.1056/EVIDoa2200008 38319256

9 Döhner H , Wei AH , Appelbaum FR , Craddock C , DiNardo CD , Dombret H , et al. Diagnosis and management of AML in adults: 2022 recommendations from an international expert panel on behalf of the ELN . Blood . 2022 ; 140 : 1345 – 1377 . 35797463 10.1182/blood.2022016867

10 Arber DA , Orazi A , Hasserjian RP , Borowitz MJ , Calvo KR , Kvasnicka HM , et al. International consensus classification of myeloid neoplasms and acute leukemias: integrating morphologic, clinical, and genomic data . Blood . 2022 ; 140 : 1200 – 1228 . 35767897 10.1182/blood.2022015850 PMC9479031

11 Grob T , Al Hinai ASA , Sanders MA , Kavelaars FG , Rijken M , Gradowska PL , et al. Molecular characterization of mutant TP53 acute myeloid leukemia and high‐risk myelodysplastic syndrome . Blood . 2022 ; 139 : 2347 – 2354 . 35108372 10.1182/blood.2021014472

12 Ediriwickrema A , Aleshin A , Reiter JG , Corces MR , Köhnke T , Stafford M , et al. Single‐cell mutational profiling enhances the clinical evaluation of AML MRD . Blood Adv . 2020 ; 4 : 943 – 952 . 32150611 10.1182/bloodadvances.2019001181 PMC7065471

13 Arber DA , Orazi A , Hasserjian R , Thiele J , Borowitz MJ , Le Beau MM , et al. The 2016 revision to the World Health Organization classification of myeloid neoplasms and acute leukemia . Blood . 2016 ; 127 : 2391 – 2405 . 27069254 10.1182/blood-2016-03-643544

14 Papaemmanuil E , Gerstung M , Bullinger L , Gaidzik VI , Paschka P , Roberts ND , et al. Genomic classification and prognosis in acute myeloid leukemia . N Engl J Med . 2016 ; 374 : 2209 – 2221 . 27276561 10.1056/NEJMoa1516192 PMC4979995

15 Tazi Y , Arango‐Ossa JE , Zhou Y , Bernard E , Thomas I , Gilkes A , et al. Unified classification and risk‐stratification in acute myeloid leukemia . Nat Commun . 2022 ; 13 : 4622 . 35941135 10.1038/s41467-022-32103-8 PMC9360033

16 Kim K , Maiti A , Loghavi S , Pourebrahim R , Kadia TM , Rausch CR , et al. Outcomes of TP53‐mutant acute myeloid leukemia with decitabine and venetoclax . Cancer . 2021 ; 127 : 3772 – 3781 . 34255353 10.1002/cncr.33689 PMC10462434

17 Ediriwickrema A , Gentles AJ , Majeti R . Single‐cell genomics in AML: extending the frontiers of AML research . Blood . 2023 ; 141 : 345 – 355 . 35926108 10.1182/blood.2021014670 PMC10082362

18 Sallman DA , DeZern AE , Garcia‐Manero G , Steensma DP , Roboz GJ , Sekeres MA , et al. Eprenetapopt (APR‐246) and azacitidine in TP53‐mutant myelodysplastic syndromes . J Clin Oncol . 2021 ; 39 : 1584 – 1594 . 33449813 10.1200/JCO.20.02341 PMC8099410

19 Konopleva MY , Röllig C , Cavenagh J , Deeren D , Girshova L , Krauter J , et al. Idasanutlin plus cytarabine in relapsed or refractory acute myeloid leukemia: results of the MIRROS trial . Blood Adv . 2022 ; 6 : 4147 – 4156 . 35413116 10.1182/bloodadvances.2021006303 PMC9327534

20 Daver N , Jonas BA , Medeiros BC , Patil U , Yan M . Phase 1b, open‐label study evaluating the safety and pharmacokinetics of atezolizumab (anti‐PD‐L1 antibody) administered in combination with Hu5F9‐G4 to patients with relapsed and/or refractory acute myeloid leukemia . Leuk Lymphoma . 2022 ; 63 : 2711 – 2714 . 35834732 10.1080/10428194.2022.2092853

21 Schwartz S , Patel N , Longmire T , Jayaraman P , Jiang X , Lu H , et al. Characterization of sabatolimab, a novel immunotherapy with immuno‐myeloid activity directed against TIM‐3 receptor . Immunother Adv . 2022 ; 2 : ltac019 . 36196369 10.1093/immadv/ltac019 PMC9525012

22 Cluzeau T , Sebert M , Rahmé R , Cuzzubbo S , Lehmann‐Che J , Madelaine I , et al. Eprenetapopt plus azacitidine in TP53‐mutated myelodysplastic syndromes and acute myeloid leukemia: a phase II study by the Groupe Francophone des Myélodysplasies (GFM) . J Clin Oncol . 2021 ; 39 : 1575 – 1583 . 33600210 10.1200/JCO.20.02342 PMC8099409

23 Mishra A , Tamari R , DeZern AE , Byrne MT , Gooptu M , Chen YB , et al. Eprenetapopt plus azacitidine after allogeneic hematopoietic stem‐cell transplantation for TP53‐mutant acute myeloid leukemia and myelodysplastic syndromes . J Clin Oncol . 2022 ; 40 : 3985 – 3993 . 35816664 10.1200/JCO.22.00181

24 Sallman DA , Al Malki MM , Asch AS , Wang ES , Jurcic JG , Bradley TJ , et al. Magrolimab in combination with azacitidine in patients with higher‐risk myelodysplastic syndromes: final results of a phase Ib study . J Clin Oncol . 2023 ; 41 : 2815 – 2826 . 36888930 10.1200/JCO.22.01794 PMC10414740

25 Patel SA , Cerny J . TP53‐mutant myelodysplastic syndrome and acute myeloid leukemia: the black hole of hematology . Blood Adv . 2022 ; 6 : 1917 – 1918 . 35090167 10.1182/bloodadvances.2021006580 PMC8941466

26 Welch JS , Petti AA , Miller CA , Fronick CC , O'Laughlin M , Fulton RS , et al. TP53 and decitabine in acute myeloid leukemia and myelodysplastic syndromes . N Engl J Med . 2016 ; 375 : 2023 – 2036 . 27959731 10.1056/NEJMoa1605949 PMC5217532

27 Badar T , Litzow MR , Shallis RM , Patel A , Saliba AN , Burkart M , et al. Disparities in receiving disease‐directed therapy, allogeneic stem cell transplantation in non‐Hispanic Black patients with TP53‐mutated acute myeloid leukemia . Cancer . 2023 ; 129 : 934 – 945 36545710 10.1002/cncr.34604

28 DiNardo KW , LeBlanc TW , Chen H . Novel agents and regimens in acute myeloid leukemia: latest updates from 2022 ASH Annual Meeting . J Hematol Oncol . 2023 ; 16 : 17 . 36869366 10.1186/s13045-023-01411-x PMC9983204

29 Marks JA , Wang X , Fenu EM , Bagg A , Lai C . TP53 in AML and MDS: the new (old) kid on the block . Blood Rev . 2023 ; 60 : 101055 . 36841672 10.1016/j.blre.2023.101055

30 Badar T , Atallah E , Shallis R , Saliba AN , Patel A , Bewersdorf JP , et al. Survival of TP53‐mutated acute myeloid leukemia patients receiving allogeneic stem cell transplantation after first induction or salvage therapy: results from the Consortium on Myeloid Malignancies and Neoplastic Diseases (COMMAND) . Leukemia . 2023 ; 37 : 799 – 806 . [In Press]. 36807649 10.1038/s41375-023-01847-7

31 Contadini C , Monteonofrio L , Virdia I , Prodosmo A , Valente D , Chessa L , et al. p53 mitotic centrosome localization preserves centrosome integrity and works as sensor for the mitotic surveillance pathway . Cell Death Dis . 2019 ; 10 : 850 . 31699974 10.1038/s41419-019-2076-1 PMC6838180

32 Mijit M , Caracciolo V , Melillo A , Amicarelli F , Giordano A . Role of p53 in the regulation of cellular senescence . Biomolecules . 2020 ; 10 : 420 . 32182711 10.3390/biom10030420 PMC7175209

33 Hu WL , Jin L , Xu A , Wang YF , Thorne RF , Zhang XD , et al. GUARDIN is a p53‐responsive long non‐coding RNA that is essential for genomic stability . Nat Cell Biol . 2018 ; 20 : 492 – 502 . 29593331 10.1038/s41556-018-0066-7

34 Weinberg OK , Siddon A , Madanat YF , Gagan J , Arber DA , Dal Cin P , et al. TP53 mutation defines a unique subgroup within complex karyotype de novo and therapy‐related MDS/AML . Blood Adv . 2022 ; 6 : 2847 – 2853 . 35073573 10.1182/bloodadvances.2021006239 PMC9092405