Genomic context and TP53 allele frequency define clinical outcomes in TP53-mutated myelodysplastic syndromes

Abstract

TP53 mutations are associated with adverse outcomes and shorter response to hypomethylating agents (HMAs) in myelodysplastic syndrome (MDS). Limited data have evaluated the impact of the type, number, and patterns of TP53 mutations in response outcomes and prognosis of MDS. We evaluated the clinicopathologic characteristics, outcomes, and response to therapy of 261 patients with MDS and TP53 mutations. Median age was 68 years (range, 18-80 years). A total of 217 patients (83%) had a complex karyotype. TP53 mutations were detected at a median variant allele frequency (VAF) of 0.39 (range, 0.01-0.94). TP53 deletion was associated with lower overall response rate (ORR) (odds ratio, 0.3; P = .021), and lower TP53 VAF correlated with higher ORR to HMAs. Increase in TP53 VAF at the time of transformation was observed in 13 patients (61%), and previously undetectable mutations were observed in 15 patients (65%). TP53 VAF was associated with worse prognosis (hazard ratio, 1.02 per 1% VAF increase; 95% confidence interval, 1.01-1.03; P < .001). Integration of TP53 VAF and karyotypic complexity identified prognostic subgroups within TP53-mutant MDS. We developed a multivariable model for overall survival that included the revised International Prognostic Scoring System (IPSS-R) categories and TP53 VAF. Total score for each patient was calculated as follows: VAF TP53 + 13 × IPSS-R blast score + 16 × IPSS-R cytogenetic score + 28 × IPSS-R hemoglobin score + 46 × IPSS-R platelet score. Use of this model identified 4 prognostic subgroups with median survival times of not reached, 42.2, 21.9, and 9.2 months. These data suggest that outcomes of patients with TP53-mutated MDS are heterogeneous and that transformation may be driven not only by TP53 but also by other factors.

Graphical abstract

Figure 1.

Landscape of identified TP53 mutations and other co-mutations. (A) Lollipop plot representing all identified mutations in TP53. Height of bar plot represents frequency of a specific mutation. Color of each mutation (circles) represents mutation type. Color of each gene region represents domains. (B) Frequency of identified mutations in other genes by amplicon-based next-generation sequencing. VAFs of identified mutations with median and range. Mutations are ordered by decreasing median VAF. (C) Frequencies of TP53 mutation, TP53 deletion, and complex karyotype among the entire patient cohort (n = 938).

Figure 2.

Influence of TP53 mutation VAF in response and dynamic changes of TP53 VAF throughout therapy. (A) Percentage change in TP53 mutation VAF before and after therapy in patients with and without a clinical response to HMA therapy. (B) TP53 VAF before and at the time of response assessment in patients with and without clinical response to HMA therapy. (C) Time-dependent changes of TP53 VAF among 64 patients with multiple sequencing time points throughout the course of therapy, including response assessment, disease progression, transformation, or relapse. Gray lines represent changes in a specific patient; blue lines represent tendencies of the entire evaluable cohort.

Figure 3.

Mutation and cytogenetic changes at the time of transformation. Frequency of TP53 mutation expansion represents percentage of patients with increase in VAF of TP53 mutation at the time of transformation among evaluable patients (n = 23).

Figure 4.

Cytogenetic evaluation and next-generation sequencing at baseline and time of transformation to AML. For each section of the figure, the upper graph represents VAF of detectable mutations at the time of diagnosis of MDS and at transformation. Lower graphs represent gain (blue) or loss (red) of chromosomal material and translocations (thin black lines connecting chromosomes) present at diagnosis or transformation. A partial semicircle present in the upper right corner of sexual chromosomes in some patients represents marker chromosomes. Inversions or derivatives are shown in green.

Figure 4.

Cytogenetic evaluation and next-generation sequencing at baseline and time of transformation to AML. For each section of the figure, the upper graph represents VAF of detectable mutations at the time of diagnosis of MDS and at transformation. Lower graphs represent gain (blue) or loss (red) of chromosomal material and translocations (thin black lines connecting chromosomes) present at diagnosis or transformation. A partial semicircle present in the upper right corner of sexual chromosomes in some patients represents marker chromosomes. Inversions or derivatives are shown in green.

Figure 4.

Cytogenetic evaluation and next-generation sequencing at baseline and time of transformation to AML. For each section of the figure, the upper graph represents VAF of detectable mutations at the time of diagnosis of MDS and at transformation. Lower graphs represent gain (blue) or loss (red) of chromosomal material and translocations (thin black lines connecting chromosomes) present at diagnosis or transformation. A partial semicircle present in the upper right corner of sexual chromosomes in some patients represents marker chromosomes. Inversions or derivatives are shown in green.

Figure 5.

Prognostic subgroups of patients with TP53-mutant MDS based on TP53 VAF and karyotype. Kaplan-Meier estimate curves for OS of patients based on the presence of TP53 mutation, complex karyotype, and TP53 deletion (A); the number of TP53 abnormalities defined as the sum of mutations and deletions (B); VAF of TP53 mutation (C); and VAF of TP53 mutation and presence of a complex karyotype (D). CCG, complex cytogenetics; LOH, loss of heterozygosity.

Figure 6.

Multivariable Cox proportional hazards model and nomogram for OS. (A) Nomogram for OS. Nomogram used by totaling points identified at top scale for each of the independent variables. This summed point score was then identified on a total point scale to identify the 3-month and 1-year survival probability and median OS. (B) Kaplan-Meier estimate curves for OS, which were based on the 4 groups determined by the median point score of the multivariable model.