Venetoclax and hypomethylating agents in TP53‐mutated acute myeloid leukaemia

TP53 mutations occur in nearly 10% of adults with de novo acute myeloid leukaemia (AML); however, these mutations are encountered at higher frequency in therapy-related AML (t-AML) (~35%) and AML with complex cytogenetics (50– 60%) (Rucker et al, 2012; Ok et al, 2015). The presence of the TP53 mutation (TP53m) conveys an extremely poor prognosis (Grossmann et al, 2012; Rucker et al, 2012). Patients with TP53m AML are generally older in age (Rucker et al, 2012), and their physical fitness, especially in those with t-AML, is poor (Granfeldt Ostgard et al, 2015). The clinical response of TP53m AML to standard induction regimens has been disappointing, and if achieved, is usually short-lived (Grossmann et al, 2012; Rucker et al, 2012). Venetoclax (VEN) has shown encouraging activity when combined with hypomethylating agents (HMAs) in both newly diagnosed and relapsed/refractory (r/r) AML (Aldoss et al, 2018; DiNardo et al, 2019). The combination is welltolerated even in frail patients and is associated with low treatment-related mortality (TRM) (DiNardo et al, 2019). Apoptosis mediated by venetoclax appears to be TP53-independent (Anderson et al, 2016), and VEN/HMA activity was observed across various high-risk leukaemia genetics (Aldoss et al, 2018; DiNardo et al, 2019). We conducted a retrospective analysis of AML patients treated with the combination of VEN/HMA at our institution between June 2016 and April 2019. We identified TP53m utilizing an in-house next generation sequencing panel. The study was approved by the City of Hope Institutional Review Board. Response was defined as either complete remission (CR) or CR with incomplete count recovery (CRi). In univariate analysis, descriptive statistics were used to summarize the covariates for both response and non-response groups. The correlations between response to VEN/ HMA and these covariates were assessed by the Pearson Chisquare test for categorical variables. A logistic multivariable regression model was applied to the three covariates with P values not exceeding 0 1; odds ratios and 95% confidence intervals (CI) are listed in Table S1. The median leukaemiafree survival (LFS) and overall survival (OS) were summarized for the response group only. All analyses were performed with R 3.5.1 (http://www.R-project.org) or SAS 9.4 (SAS Institute Inc, Cary, NC). We identified 32 adults with TP53m AML who were treated with VEN/HMA. One patient with r/r AML was unevaluable for response because of death from sepsis 10 days after initiating therapy. The median age was 68 years (22–85). Sixteen (52%) patients had r/r AML while 15 (48%) patients were newly diagnosed. The majority of patients had complex cytogenetics (n = 24, 77%). The median variant allele frequency (VAF) for TP53m was 56% (range, 3–95%), and 9 (29%) patients had more than one TP53 mutation. Nineteen (61%) patients had additional somatic mutations. Decitabine was the HMA used in the majority of cases (90%) and was more frequently given as a 5-day schedule rather than a 10day schedule (52% vs. 39%). Among the 31 evaluable patients, 16 (52%) experienced a response, including 7 CR and 9 CRi. Response was achieved after a median of 2 (range, 1–3) cycles. Minimal residual disease (MRD), defined as >0 01% by multiparameter flow cytometry in a reference laboratory, was evaluated in 10 of the responders, and 7 (70%) achieved MRD negativity. Prior HMA monotherapy was the only factor observed to be associated with a lower response compared to that in patients who were HMA-naive (14% vs. 63%, P = 0 025). There was a trend toward a higher CR/CRi rate in patients with more than one TP53 mutation (78% vs. 41%, P = 0 062) and in those who were treated with VEN/HMA in the frontline setting (67% vs. 38%, P = 0 1). Response was comparable across patient age, sex, AML type, receipt of prior allogeneic haematopoietic cell transplantation (alloHCT), cytogenetics, presence of additional mutations and type and duration of HMA therapy. Table I depicts patient characteristics and response rates. In multivariate analysis of response, prior exposure to HMA was associated with a non-significant trend toward lower response (odds ratio = 0 12; 95%CI: 1 69 to 1 27; P = 0 092) (Table S2). Response to VEN/HMA in TP53m AML according to associated additional mutations is shown in Table S1. The location and pattern of TP53 mutations for responders and non-responders is depicted in Fig 1. The median TP53 VAF was comparable for responders and non-responders (56% vs. 52%). The median LFS and OS for responders was 234 days (95% CI: 101–329) and 329 days (95% CI: 284–not reached), respectively. Five responders (31%) underwent alloHCT in CR. Consistent with previous reports (DiNardo et al, 2019), we illustrate that the response to this regimen in the frontline setting is encouraging for this high-risk group, especially considering the fact that the majority of patients with TP53m AML carried complex cytogenetics. Furthermore, we Correspondence