Evolutionary Action (EA) Score of TP53 Mutations Defines Prognostic Subsets within TP53 Mutated Myelodysplastic Syndromes and Acute Myeloid Leukemia

INTRODUCTION

TP53 mutated myelodysplastic syndromes (MDS) and acute myeloid leukemia (AML) have a poor outcome. TP53 mutations occur throughout the coding region leading to functionally distinct downstream deleterious effects. The impact of the type and functional effects of TP53 mutations on the disease presentation (diagnosis, AML vs. MDS), prognosis and outcome has not been explored. Evolutionary action (EA) score is a novel computational approach to quantify the deleterious impact of missense TP53 mutations. In this study, we evaluated the effect of EAp53 score on these parameters in AML and MDS patients (pts).

METHODS

We selected all newly diagnosed MDS and AML pts (2013-2018) with at least 1 TP53 missense mutation detected by next generation sequencing (NGS) with 1% detection sensitivity. EAp53 score was obtained from the web server. The median overall survival (OS) was calculated by Kaplan-Meier method. Optimal EAp53 cutoffs were determined based on survival independently for AML and MDS using the lowest AIC. Univariate and multivariable Cox proportional hazards were used to identify association between risk factors and survival. Responses were evaluated using the 2006 IWG criteria. TP53 immunohistochemistry (clone DO-7) scoring was based on both intensity and percentage of positive cells.

RESULTS

The study group had 362 newly diagnosed pts with 471 TP53 mutations (221 MDS, 141 AML). This included 216 (59.6%) men and 146 (40.4%) women with a median age of 69 years (range, 18-90.4). 261 pts (72%) had a single TP53 mutation while 101 (28%) had >1 TP53 mutations. The median VAF was 36.4 (range, 1-94.4); median EAp53 score was 78.5 [range, 0.6 - 97.9]. No differences in the TP53 mutation characteristics including type, VAF and EAp53 scores were identified by diagnosis: EAp53 score [MDS: 79.0 (range, 0.6-97.9), AML: 78.5 (1.4-97.7)]; median VAF [MDS: 33.7 (1.0-93.6), AML: 43.9 (1.2-94.4)]. Over a median follow-up of 21.0 months, the median OS for AML was shorter than MDS (4.7 vs. 11.5 months); p < 0.001].

Within both AML and MDS subsets, EAp53 score cut-off of ≥60 predicted worse OS; this was determined as the best cut-off by AIC. Evaluation of outcomes based on both diagnosis and EAp53 scores demonstrated that EAp53 score ≥60 predicted for worse OS when corrected for diagnosis: [p=0.014; HR 1.63 (95% CI: 1.10-2.41)] (Figure 1A). Within MDS, low vs. high EAp53 score had a trend to correlate with OS (median OS 16.8 vs. 11.5 mo, p=0.07) but not in AML (7.1 vs. 4.7 mo; p = 0.17). Since low EA-AML pts and high EA-MDS pts showed similar outcome, EA score identified 3 distinct prognostic subsets within the entire cohort [TIER 1, Low EA-MDS; median OS: 16.8 mo; TIER 2, High EA-MDS plus low EA-AML; median OS: 11.1 mo; and TIER 3, High EA-AML, median OS: 4.7 mo; p < 0.001] (Figure 1B). By multivariate analysis, diagnosis (AML vs. MDS), EAp53 score (cut-off of 60), TP53 VAF, age, gender, platelet count and hemoglobin were independent predictors of OS within the study cohort. EA did not independently predict OS within AML or MDS subsets.

Comparison of TIERS 3 vs. 2 vs. 1 groups showed increasing frequencies of complex karyotype (CK), del(5q), del(17p), lower frequencies of diploid karyotype (p=0.01) and a trend for higher number of TP53 mutations (p=0.08) more likely involving DNA-binding domain (p=0.014). These differences were retained between low and high EA pts in AML and MDS subsets. In addition, low-EA MDS pts belonged to very good/good/intermediate IPSS-R. By immunohistochemistry, median TP53 protein expression level (H-score, intensity and percentage) was significantly lower in low-EA versus high EA pts (p=0.038) and correlated with both TP53 VAF and EA score. In contrast, comparison within TIER 2 (between high-EA MDS and low-EA AML) showed no significant differences including age distribution, prior history of chemoradiation and TP53 VAF [high-EA MDS, 34.1 (1.1-93.6) vs. low-EA AML, 47 (1.2-82.2)]. High-EA MDS pts had a tendency for CK and a higher number of TP53 mutations. High-EA MDS pts were preferentially treated with HMA-based regimens (p=0.0001) while low-EA AML pts received chemotherapy (p=0.0002).

CONCLUSION

EAp53 score identifies prognostic subsets with distinct clinicopathologic characteristics within both AML and MDS pts with missense TP53 mutations, and can independently predict OS along with VAF. EA score should be used as a biomarker for baseline risk assessment in prognostic models and treatment decisions.

Figure 1

View largeDownload slide

Figure 1

View largeDownload slide

Close modal