https://orcid.org/0000-0003-0504-8233
In this issue of Blood Advances, Shah et al1 provide an evidence-based proposal to harmonize the classification of TP53 mutant myelodysplastic neoplasms (MDS) and acute myeloid leukemia (AML). Patients with MDS/AML with TP53 mutations represent a molecular subset of patients with the poorest outcomes to date, although with some heterogeneity. Specifically, initial studies interrogating differential outcomes based on the variant allele frequency (VAF) of TP53 identified a VAF cutoff of 20% to 23% for the best discrimination of outcomes.2,3
A subsequent seminal publication describing 3324 MDS patients with TP53 mutations identified the allelic status of TP53 to be the most robust predictor of outcome. Specifically, patients who were biallelic/multihit (ie, two-thirds of the cohort) had an inferior overall survival (OS) compared with monoallelic patients who had similar outcomes to patients with the wild type.4 Biallelic was defined as (1) the presence of 2 TP53 mutations or (2) a mutation and deletion or (3) a mutation with copy-neutral loss of heterozygosity. Notably a VAF cutoff of 23% correlated strongly with a multihit status and similarly was shown to represent a probable biallelic status in a recently published cohort.4,5 In parallel, there has been substantial work evaluating the clinical homogeneity of patients with TP53 mutant MDS/AML (n = 2200), in which patients who had ≥5% bone marrow blasts had identical clinical characteristics and survival compared with patients with AML (ie, ≥20 blasts; 2-year OS of 13%).6 Notably, the allelic status did not further stratify patients. Validation of the impact of ≥5% bone marrow blasts was recently shown in a cohort of 747 patients, which additionally highlighted the complex interplay between cytogenetic complexity, VAF, blast count, and allelic status.7
In this regard, the 2022 updates by the World Health Organization (WHO) and the International Consensus Classification (ICC) classified patients with the TP53 mutant with biallelic or multihit status, respectively, albeit with substantial differences between the 2 classifications.8,9 Notably, the 2022 WHO classification only recognizes MDS with biallelic TP53 mutations (ie, bone marrow blasts <20%), has no VAF cutoff, and does not recognize complex karyotype as a biallelic equivalent. In contrast, the ICC recognizes complex karyotype as a multihit equivalent but notably requires a TP53 VAF of >10%. Additionally, patients with bone marrow blasts ≥10% and a single TP53 mutation are classified as TP53 mutant disease (MDS/AML with 10%-19% blasts and AML with ≥20% blasts).
Although the recognition of TP53 mutant MDS/AML is of profound clinical consequence, the heterogeneity in classification and prognostic discrimination has brought forth significant confusion and an overemphasis at times on identifying patients with poor risk. Thus, efforts to harmonize classifications are therefore of critical importance and is the focus of this current article in Blood Advances by Shah et al. Specifically, this is a multicenter study comprising 580 patients from the Mayo Clinic and the South Australia Health Network. Most importantly, 92% of patients with TP53 mutant MDS/AML were found to have poor OS, which contrasted with 37% and 81% by WHO and ICC, respectively (see figure). The main differences between WHO and ICC exist because unlike ICC, WHO does not define TP53 mutant AML and does not account for cytogenetic complexity. However, the data from the current cohort1 support the fact that most patients with TP53 mutant AML have poor outcomes. Further, similar to prospectively treated patients on HOVON-SAKK clinical trials,6 the study by Shah et al also did not identify any prognostic discrimination based on allelic status in patients with ≥5 % blasts (median OS ∼10 months with MDS-excess blasts 1 (EB1) disease with or without biallelic inactivation). Notably, however, 89% of the patients with MDS-EB were biallelic in this study.1
Distribution of poor-risk TP53 mutant MDS/AML vs good-risk TP53 mutant MDS according to Shah et al (92% vs 8%). Poor-risk TP53 mutant MDS/AML is identified by patients with a biallelic/multihit disease, complex karyotype, and ≥5% bone marrow myeloblasts, with the good-risk TP53 mutant patients not meeting any of these criteria. Current cohort also identified a VAF cutoff of 10% to best stratify survival, although multiple past cohorts have identified a VAF cutoff of 20% to 23% as the most optimal.
Distribution of poor-risk TP53 mutant MDS/AML vs good-risk TP53 mutant MDS according to Shah et al (92% vs 8%). Poor-risk TP53 mutant MDS/AML is identified by patients with a biallelic/multihit disease, complex karyotype, and ≥5% bone marrow myeloblasts, with the good-risk TP53 mutant patients not meeting any of these criteria. Current cohort also identified a VAF cutoff of 10% to best stratify survival, although multiple past cohorts have identified a VAF cutoff of 20% to 23% as the most optimal.
Additionally, in the MDS-EB population, a VAF cutoff of 10% was only relevant in patients without complex karyotype. This is in contrast to the ICC whereby patients with TP53 mutant MDS-EB with VAF <10% with a complex karyotype would be classified as monoallelic. In this study,1 patients with MDS-EB with a VAF <10% and without a complex karyotype had a median OS of 26 months vs ∼6 months in patients with VAF <10% and complex karyotype or patients with MDS-EB with VAF ≥10%. In AML, the concordance of complex karyotype and biallelic inactivation was even more striking, occurring in 94% and 88% of patients, respectively, with a median OS of 4 months. Although clearly recognizing patients with TP53 mutant MDS-EB/AML with a monoallelic TP53 mutation and an absence of complex karyotype can identify a subgroup with reasonable outcomes (median OS 26 months), this is strikingly rare because it occurred in only 2.8% of the cohort (n = 16; 8 MDS-EB and 8 AML).1 Clearly, the pivotal role of allelic status evaluation of patients with TP53 mutant MDS lies with patients with low blasts (MDS-LB). Patients with MDS-LB with monoallelic TP53 mutation and without complex karyotype had a median OS of 35 months, representing 8% of the cohort, vs an OS of 13 months of patients with MDS-LB with a complex karyotype or biallelic inactivation. Although one could argue that there is some prognostic difference between biallelic or complex karyotype AML (4 months), MDS-EB (8 months), and MDS-LB (13 months), this could be seen as simply debating how the adjectives “terrible” vs “horrible” are defined. Thus, the authors proposed a hierarchical model with an easy-to-use online calculator.1
Therefore, why is the unification of the TP53 mutant MDS/AML classification of such profound clinical importance? Through international collaboration, we can rapidly enroll TP53 mutant patients with MDS/AML on study because there have been 2 phase 3 pivotal trials to date evaluating eprenetapopt and magrolimab, although notably, the former was using WHO-defined MDS, and the latter used WHO-defined AML, with both studies only requiring a single TP53 mutation.10-13 Although both studies were ultimately negative, a key initiative is to enroll uniform cohorts of patients with poor-risk TP53 on early phase 1 trials, which is not only important to improve the outcomes of TP53 mutant patients but also important to avoid abandoning potentially promising therapies that have may have limited their investigation to patients with TP53 mutations. In summary, and in agreement with Shah et al, I would argue that future classification of TP53 mutant MDS/AML should include all patients who have a biallelic inactivation of TP53, complex karyotype, and/or ≥5 % blasts. Furthermore, this harmonization should be utilized for optimal enrollment on clinical trials, even as first-line therapy given the poor outcomes in this patient population.
Conflict-of-interest disclosure: D.A.S. reports research funding from Aprea Therapeutics and advisory board fees from Gilead.
