TP53 mutation characteristics in therapy-related myelodysplastic syndromes and acute myeloid leukemia is similar to de novo diseases

BACKGROUND: TP53 mutation is more prevalent in therapy-related myeloid neoplasms (t-MN) than their de novo counterparts; however, the pattern of mutations involving TP53 gene in t-MN versus de novo diseases is largely unknown. METHODS: We collected 108 consecutive patients with therapy-related myelodysplastic syndrome (t-MDS)/acute myeloid leukemia (t-AML). Clinical, hematological, and cytogenetic data were collected by searching the electronic medical record. TP53 sequencing was performed in all patients using a clinically validated next-generation sequencing-based gene panel assay. A previously published patient cohort consisting of 428 patients with de novo MDS/AML was included for comparison. RESULTS: We assessed 108 patients with t-MN, in which 40 patients (37%) had TP53 mutations. The mutation frequency was similar between t-MDS and t-AML; but significantly higher than de novo MDS/AML (62/428 patients, 14.5%) (p < 0.0001). TP53 mutations in t-MN were mainly clustered in DNA-binding domains, with an allelic frequency of 37.0% (range, 7.1 to 98.8). Most mutations involved single nucleotide changes, of which, transitions (65.9%) were more common than transversions (34.1%). Missense mutations were the most frequent, followed by frameshift and nonsense mutations. This TP53 mutation pattern was strikingly similar to that observed in de novo MDS/AML. TP53 mutations in t-MN were associated with a complex karyotype (p < 0.0001), a higher number of chromosomal abnormalities (p < 0.0001), and an inferior overall survival in affected patients (6.1 vs 14.1 months) by univariate (p < 0.0001) and multivariate analyses (p = 0.0020). CONCLUSIONS: Our findings support the recent notion that heterozygous TP53 mutation may be a function of normal aging and that mutated cells are subject to selection upon exposure to cytotoxic therapy. t-MN carrying TP53 mutation have an aggressive clinical course independent of other confounding factors.