Prognostic Significance of TP53 Alterations in Patients with Therapy-Related Myeloid Neoplasms

Eleni Vasileiou¹, Marija Juric², Joanna Pawlak³

Introduction

or 7, complex cytogenetics and reduced response to chemotherapy [1,2].

The obvious interpretation was that the treatment caused the mutations. Cytotoxic agents damage DNA, TP53 is the canonical guardian of genomic integrity, and a mutation appearing after exposure looks like a consequence of it [3,4]. That reading shaped how the disease was discussed for years and it turns out to be wrong.

TP53-altered myeloid disease carries the worst outcome of any molecular subgroup, with response rates to induction chemotherapy, hypomethylating agents and venetoclax-based regimens all lower than in unmutated disease and median survival of five to ten months [5,6]. Recent classifications have recognised it as a distinct entity rather than a risk modifier [7,8,9].

This review sets out how often TP53 alterations occur in the therapy-related setting, where the mutant clone actually comes from, what it predicts, and why the allelic-state rules developed for de novo disease do not appear to apply here.

Material and Methods

Search strategy: PubMed and the archives of the major haematology and genetics journals were searched for studies of TP53 in therapy-related myeloid disease. Search terms combined therapy-related myeloid neoplasm, secondary acute myeloid leukaemia and therapy-related myelodysplastic syndrome with TP53, variant allele frequency, allelic state, multi-hit, clonal haematopoiesis and complex karyotype. Reference lists of retrieved sequencing studies and classification documents were screened by hand.

Eligibility: Studies were included if they reported TP53 alteration frequency, clonal origin, variant allele frequency distribution or outcome in a defined myeloid neoplasm population. Studies of de novo disease are cited where they provide the comparator or established the allelic-state framework, and are identified as such throughout.

Data extraction and handling: Design, cohort size, sequencing platform, variant allele frequency threshold and reported survival estimates were extracted as published. The threshold and the method of assessing allelic state are stated alongside every figure, because detection of loss of heterozygosity depends on whether copy-number and single-nucleotide polymorphism array data were available. No pooling or re-analysis was performed and no patient-level data were obtained. Figures 1 to 4 and Tables 1 to 3 reproduce published values only. This is a narrative synthesis and was not registered as a systematic review.

Results

How much commoner the mutation is: TP53 mutations are found in 5 to 10 per cent of de novo myelodysplastic syndrome and acute myeloid leukaemia, where they associate with complex karyotype and shortened survival [10,11,12]. In therapy-related disease the figure is 21 to 38 per cent [1,2]. Where the karyotype is complex, or chromosome 17 or 17p, 5 or 5q, or 7 or 7q is lost, the frequency of TP53 abnormality rises to 70 to 80 per cent (Figure 1) [5,13].

The mutations themselves are unremarkable. In therapy-related disease they cluster in the DNA-binding domains at a median allelic frequency of 37.0 per cent (range 7.1 to 98.8), most are single nucleotide changes with transitions outnumbering transversions at 65.9 against 34.1 per cent, and missense mutations predominate over frameshift and nonsense [2]. That pattern is strikingly similar to what is seen in de novo disease, which is the first clue that something other than direct mutagenesis is at work [2].

The clone predates the treatment: Whole-genome sequencing of 22 patients with therapy-related acute myeloid leukaemia settled the question of causation. The total number of somatic single-nucleotide variants and the proportion of chemotherapy-related transversions were similar to de novo disease, indicating that cytotoxic therapy neither produces genome-wide DNA damage of the expected kind nor directly induces the TP53 mutation [3].

The decisive observation followed. In four patients whose leukaemia carried a TP53 mutation, that exact mutation was detectable at frequencies of 0.003 to 0.7 per cent in mobilised blood leucocytes or marrow collected three to six years before the neoplasm appeared [3]. The clone was present before the chemotherapy that supposedly caused it.

The mechanism is selection rather than mutagenesis. Haematopoietic stem cells carrying a TP53 mutation have a relative fitness advantage under cytotoxic stress, so treatment does not create the clone but clears the competition around it [3,14]. Cellular stressors more generally expand haematopoietic clones of varying leukaemic potential [15], and cancer therapy has been shown to shape the fitness landscape of clonal haematopoiesis, with TP53, CHEK2 and PPM1D clones associated with chemotherapy exposure while DNMT3A and TET2 clones are not [16].

Clinical data support the same picture. Among patients with tubo-ovarian cancer treated with a PARP inhibitor, pre-existing TP53-mutant clonal haematopoiesis at a variant allele frequency of 1 per cent or above was present in 45.0 per cent of those who went on to develop a therapy-related neoplasm against 13.6 per cent of those who did not (p = 0.009) (Figure 3), with clones detectable up to six

Study n Focus Principal reported finding

TP53 frequent and Bowen et al. 2009 Complex — associated with very poor [11] karyotype outcome Correlates with copy Rücker et al. 2012 Complex — number change and dismal [13] karyotype outcome Mutation present 3-6 years Wong et al. 2015 [3] 22 Origin before the neoplasm Papaemmanuil 2016 Genomic classification of — Classification [12] acute myeloid leukaemia Bolton et al. 2020 Clonal Therapy shapes the clonal — [16] fitness haematopoiesis landscape VAF 10% threshold; single Shah et al. 2023 [18] 488 Prognosis and multi-hit equally poor Table 1: Principal studies reviewed, with focus and the finding as reported by the original investigators What the mutation predicts: In therapy-related disease, TP53 mutation was associated with complex karyotype (p < 0.0001), a higher number of chromosomal abnormalities (p < 0.0001) and inferior overall survival, at a median of 6.1 against 14.1 months, significant on both univariate (p < 0.0001) and multivariate analysis (p = 0.0020) (Figure 2) [2]. Those figures sit within the five to ten month median reported for TP53-mutated myeloid disease generally [5].

The poor outcome is not confined to one treatment modality. Response rates are lower with induction chemotherapy, with hypomethylating agents and with venetoclax-based regimens alike [5,6], which is what distinguishes this from a resistance phenotype specific to a class of drug. In transplantation, TP53 mutation is among the strongest predictors of poor outcome after stem-cell transplant for myelodysplastic syndrome [19].

Where the therapy-related setting diverges: Classification of de novo TP53-mutated disease turns on allelic state. A single mutation with the second allele intact behaves differently from biallelic or multi-hit disease, and current criteria from both the International Consensus Classification and the World Health Organization incorporate that distinction [7,8,9]. Mechanistic work supports it: stem cells with monoallelic mutation gain clonal fitness while retaining genomic integrity, whereas biallelic loss produces the genomic instability required for transformation [20].

Analysis of 488 therapy-related patients tested whether the same rules apply and found that they largely do not. Survival was equally poor in single-hit and multi-hit disease, which is the opposite of the de novo pattern (Figure 4) [18]. What did separate outcome was variant allele frequency: mutation at 10 per cent or above, or loss of the TP53 locus, identified the poor-prognosis group, while mutation below 10 per cent behaved more like wild-type disease [18].

An explanation is available within the same data. Among patients with a single mutation at variant allele frequency between 10 and 50 per cent, loss-of-heterozygosity information was available in only a subset, and 80 per cent of those without such information had a complex karyotype and would be considered equivalent to multi-hit under the consensus criteria [18]. Much apparent single-hit disease in this setting may therefore be biallelic disease that the available

Therapy-related

Feature De novo disease

disease

TP53 mutation 5-10% 21-38% [1,2] frequency DNA-binding Mutation spectrum Indistinguishable [2] domain, missense Allelic state separates Yes No; equally poor [18,20] outcome Variant allele Debated 10% discriminates [18] frequency threshold Median overall 6.1 months if mutated 5-10 months survival [2,5] Table 2: How TP53-altered disease differs between the de novo and therapy-related settings A practical question follows from the threshold finding. Variant allele frequency is not a fixed property of the leukaemia: it scales with blast percentage and with the purity of the sample sequenced, so the same clone measured in a hypocellular marrow and in a blast-rich one will return different values [18]. A 10 per cent cutoff is therefore a threshold on a quantity that is partly an artefact of sampling, which is the same difficulty that undermined the FLT3-ITD allelic ratio.

Residual disease in a disease defined by a resistant clone:

Measurable residual disease has become central to risk-adapted management of acute myeloid leukaemia generally [22,23], and its behaviour in TP53-altered disease is instructive. Because the mutant clone often predates the leukaemia and persists in non-leukaemic haematopoiesis, a sequencing assay tracking TP53 may remain positive in a patient who has cleared their leukaemia, detecting clonal haematopoiesis rather than residual disease [16,24].

This is the same interpretive problem that arises with DNMT3A and TET2 mutations, which persist in remission without indicating impending relapse [25]. It argues for pairing molecular measurement with an immunophenotypic or morphological assessment in this subgroup rather than relying on the mutation alone, and it is one reason current recommendations treat certain mutations as unsuitable residual disease markers [22,23].

The revised genetic risk classifications place TP53-altered disease in the adverse category on the strength of outcome data of the kind reviewed here, and validation across independent cohorts supports that placement [7,24]. What classification cannot do is change the outcome, and for this subgroup it remains the worst in the disease.

What the evidence

Question Source shows

Does chemotherapy No; mutational spectrum Genome sequencing cause the matches de novo disease of 22 cases [3] mutation?

When does the Detectable 3-6 years Banked pre-treatment clone appear? before the neoplasm samples [3] Yes; 45% against 13.6% Can it be found Prospective PARP in those who later before treatment? inhibitor cohort [17] develop disease Does allelic state No; single and multi-hit 488 therapy-related matter here? equally poor patients [18] What threshold Variant allele frequency separates Same cohort [18] of 10% or locus loss outcome?

The causal story here was revised by sequencing and the revision matters clinically. If chemotherapy created these mutations, the response would be to give less of it. Because chemotherapy selects clones that already exist, the response is to find them beforehand, and that is a screening question rather than a dosing one [3,16,17].

The evidence for pre-existence is unusually direct. Detecting the identical mutation at 0.003 per cent in a sample banked six years earlier is not an inference from mutational signatures; it is the clone itself, observed before the exposure [3]. The prospective data in PARP inhibitor-treated patients extend it: pre-treatment TP53 clonal haematopoiesis was more than three times commoner in those who later developed a neoplasm [17].

Why the allelic-state rules do not transfer

Classification of de novo TP53-mutated disease depends on distinguishing monoallelic from biallelic involvement, and that distinction is well founded biologically [20]. In the therapy-related setting the distinction loses its discriminating power [18], and there are two candidate explanations that the data do not fully separate.

The first is technical. Establishing biallelic status requires loss-of-heterozygosity and copy-number data that were incomplete in a substantial proportion of cases, and most of those lacking it had complex karyotypes suggesting hidden biallelic loss [18]. On that reading, single-hit therapy-related disease is largely a measurement artefact.

The second is biological. Therapy-related disease arises in a marrow that has already been through selection, so even a monoallelic clone reaching clinical disease has demonstrated fitness that a comparable de novo clone has not. On that reading the equivalence is real. Distinguishing the two would require systematic copy-number analysis in a therapy-related cohort, which has not been reported [18].

Study limitations

The limitations belong to the underlying literature. The sequencing study establishing pre-existence examined 22 patients and demonstrated the mutation in banked samples from four of them, so the proportion of therapy-related disease arising this way is not established [3]. Loss-of-heterozygosity data were incomplete in the largest prognostic cohort, which directly affects the single-hit versus multi-hit comparison [18]. Therapy-related neoplasms are heterogeneous in the primary malignancy treated, the agents used and the interval to presentation, and cohorts differ in their composition [1,21]. Variant allele frequency depends on blast percentage and sample purity, so the 10 per cent threshold is not a fixed biological quantity [18]. And classification criteria changed during the period covered, so the entity being described is not constant across studies [7,8,9].

Two gaps deserve naming. No study has tested whether screening for TP53-mutant clonal haematopoiesis before cytotoxic therapy, and modifying treatment accordingly, reduces the incidence of therapy-related neoplasms [16,17]. And no trial has evaluated whether the poor outcome in therapy-related TP53-mutated disease responds differently to any available regimen than the equivalent de novo disease does [5,6].

Conclusion

TP53 mutations occur in 21 to 38 per cent of therapy-related myeloid neoplasms against 5 to 10 per cent of de novo disease, and reach 70 to 80 per cent where the karyotype is complex. They are not caused by the treatment: mutational spectra are indistinguishable from de novo disease, the burden of chemotherapy-related transversions is no higher, and the identical mutation has been detected at fractions of a per cent in blood banked three to six years before the neoplasm appeared. Cytotoxic therapy selects a pre-existing clone rather than creating it. Once present, the mutation halves survival from 14.1 to 6.1 months, and the allelic-state distinction that governs classification of de novo disease does not separate outcome here, where a variant allele frequency above 10 per cent identifies the poor-prognosis group irrespective of hit number. The actionable implication is upstream of the diagnosis rather than after it.

Two lines of work follow from that. The first is whether pre-treatment screening for TP53-mutant clonal haematopoiesis can identify patients in whom an alternative regimen would avoid the second malignancy, which is testable and has not been tested [16,17]. The second is whether the apparent equivalence of single-hit and multi-hit disease survives systematic copy-number analysis, or whether it dissolves once biallelic loss is properly ascertained [18,20].

Neither question requires a new drug. Both require applying existing assays before the point at which this disease currently comes to attention, which is after it has already formed and when almost nothing works.

Acknowledgements

The authors thank the sequencing consortia and clinical cohorts whose published analyses form the substance of this article.

Funding

No dedicated funding from public, commercial or charitable sources supported this review.

Conflict of Interest

The authors declare no competing interests.

Ethics Statement

This article reviews previously published studies only. No new patient data were collected and no participants were recruited, so ethical approval was not required.

References

1. Godley, LA., Larson, RA. (2008) Therapy-related myeloid leukemia. Semin Oncol, 35(4): 418-429.
2. Ok, CY., Patel, KP., Garcia-Manero, G., Routbort, MJ., Peng, J., Tang, G., et al. (2015) TP53 mutation characteristics in therapy-related myelodysplastic syndromes and acute myeloid leukemia is similar to de novo diseases. J Hematol Oncol, 8: 45.
3. Wong, TN., Ramsingh, G., Young, AL., Miller, CA., Touma, W., Welch, JS., Lamprecht, TL., Shen, D., Hundal, J., Fulton, RS., et al. (2015) Role of TP53 mutations in the origin and evolution of therapy-related acute myeloid leukaemia. Nature, 518(7540): 552-555.
4. Soussi, T. (2000) The p53 tumor suppressor gene: from molecular biology to clinical investigation. Ann N Y Acad Sci, 910: 121-137.
5. Daver, NG., Maiti, A., Kadia, TM., Vyas, P., Majeti, R., Wei, AH., et al. (2022) TP53-mutated myelodysplastic syndrome and acute myeloid leukemia: biology, current therapy, and future directions. Cancer Discov, 12(11): 2516-2529.
6. Kantarjian, H., Kadia, T., DiNardo, C., Daver, N., Borthakur, G., Jabbour, E., et al. (2021) Acute myeloid leukemia: current progress and future directions. Blood Cancer J, 11: 41.
7. Döhner, H., Wei, AH., Appelbaum, FR., Craddock, C., DiNardo, CD., Dombret, H., et al. (2022) Diagnosis and management of AML in adults: 2022 recommendations from an international expert panel on behalf of the ELN. Blood, 140(12): 1345-1377.
8. Hart, SA., Lee, LA., Seegmiller, AC., Mason, EF. (2025) Diagnosis of TP53-mutated myeloid disease by the ICC and WHO fifth edition classifications. Blood Adv, 9(3): 445-454.
9. Kayser, S., Levis, MJ. (2021) Updates on targeted therapies for acute myeloid leukaemia. Br J Haematol, 196(2): 316-328.
10. Bejar, R., Stevenson, K., Abdel-Wahab, O., Galili, N., Nilsson, B., Garcia-Manero, G., et al. (2011) Clinical effect of point mutations in myelodysplastic syndromes. N Engl J Med, 364(26): 2496-2506.
11. Bowen, D., Groves, MJ., Burnett, AK., Patel, Y., Allen, C., Green, C., et al. (2009) TP53 gene mutation is frequent in patients with acute myeloid leukemia and complex karyotype, and is associated with very poor prognosis. Leukemia, 23(1): 203-206.
12. Papaemmanuil, E., Gerstung, M., Bullinger, L., Gaidzik, VI., Paschka, P., Roberts, ND., et al. (2016) Genomic classification and prognosis in acute myeloid leukemia. N Engl J Med, 374(23): 2209-2221.
13. Rücker, FG., Schlenk, RF., Bullinger, L., Kayser, S., Teleanu, V., Kett, H., et al. (2012) TP53 alterations in acute myeloid leukemia with complex karyotype correlate with specific copy number alterations, monosomal karyotype, and dismal outcome. Blood, 119(9): 2114-2121.
14. Klimovich, B., Merle, N., Neumann, M., Elmshäuser, S., Nist, A., Mernberger, M., Kazdal, D., Stenzinger, A., Timofeev, O., Stiewe, T. (2022) p53 partial loss-of-function mutations sensitize to chemotherapy. Oncogene, 41(7): 1011-1023.
15. Wong, TN., Miller, CA., Jotte, MRM., Bagegni, N., Baty, JD., Schmidt, AP., et al. (2018) Cellular stressors contribute to the expansion of hematopoietic clones of varying leukemic potential. Nat Commun, 9(1): 455.
16. Bolton, KL., Ptashkin, RN., Gao, T., Braunstein, L., Devlin, SM., Kelly, D., et al. (2020) Cancer therapy shapes the fitness landscape of clonal hematopoiesis. Nat Genet, 52(11): 1219-1226.
17. Kwan, TT., Oza, AM., Tinker, AV., Ray-Coquard, I., Oaknin, A., Aghajanian, C., et al. (2021) Preexisting TP53-variant clonal hematopoiesis and risk of secondary myeloid neoplasms in patients with high-grade ovarian cancer treated with rucaparib. JAMA Oncol, 7(12): 1772-1781.
18. Shah, MV., Tran, ENH., Shah, S., Chhetri, R., Baranwal, A., Ladon, D., et al. (2023) TP53 mutation variant allele frequency of ≥10% is associated with poor prognosis in therapy-related myeloid neoplasms. Blood Cancer J, 13(1): 51.
19. Lindsley, RC., Saber, W., Mar, BG., Redd, R., Wang, T., Haagenson, MD., et al. (2017) Prognostic mutations in myelodysplastic syndrome after stem-cell transplantation. N Engl J Med, 376(6): 536-547.
20. Bernard, E., Nannya, Y., Hasserjian, RP., Devlin, SM., Tuechler, H., Medina-Martinez, JS., et al. (2020) Implications of TP53 allelic state for genome stability, clinical presentation and outcomes in myelodysplastic syndromes. Nat Med, 26(10): 1549-1556.
21. Grob, T., Al Hinai, ASA., Sanders, MA., Kavelaars, FG., Rijken, M., Gradowska, PL., et al. (2022) Molecular characterization of mutant TP53 acute myeloid leukemia and high-risk myelodysplastic syndrome. Blood, 139(15): 2347-2354.
22. Schuurhuis, GJ., Heuser, M., Freeman, S., Béné, MC., Buccisano, F., Cloos, J., et al. (2018) Minimal/measurable residual disease in AML: a consensus document from the European LeukemiaNet MRD Working Party. Blood, 131(12): 1275-1291.
23. Heuser, M., Freeman, SD., Ossenkoppele, GJ., Buccisano, F., Hourigan, CS., Ngai, LL., et al. (2021) 2021 update on MRD in acute myeloid leukemia: a consensus document from the European LeukemiaNet MRD Working Party. Blood, 138(26): 2753-2767.
24. Mrózek, K., Kohlschmidt, J., Blachly, JS., Nicolet, D., Carroll, AJ., Archer, KJ., et al. (2023) Outcome prediction by the 2022 European LeukemiaNet genetic-risk classification for adults with acute myeloid leukemia: an Alliance study. Leukemia, 37(4): 788-798.
25. Jongen-Lavrencic, M., Grob, T., Hanekamp, D., Kavelaars, FG., al Hinai, A., Zeilemaker, A., et al. (2018) Molecular minimal residual disease in acute myeloid leukemia. N Engl J Med, 378(13): 1189-1199.