Outcomes of <scp><i>TP53</i></scp>‐mutated <scp>AML</scp> with evolving frontline therapies: Impact of allogeneic stem cell transplantation on survival

Abstract

To the Editor: TP53 mutations occur in 10%–20% of patients with acute myeloid leukemia (AML) and are predominantly associated with therapy-related AML and complex cytogenetics (CG).1 TP53-mutated (m) AML is considered a high-risk disease, which is resistant to conventional chemotherapy and confers poor prognosis.1 Recently, several novel therapies have been approved, providing an opportunity to have a risk-adapted approach to AML treatment. Among the newer therapies, CPX-351, a liposomal formulation of cytarabine and daunorubicin, was approved for the therapy of adults with newly diagnosed AML with myelodysplasia-related changes (AML-MRC) and therapy-related AML based on an improvement in OS when compared with 3 + 7 chemotherapy.2 Similarly, in the VIALE-A trial, the combination of azacitidine and venetoclax was found to be superior to azacitdine alone, with improved response rate and overall survival (OS) in older or unfit patients with newly diagnosed AML.3 However, the most effective approach for the treatment of TP53m AML remains unclear. Based on these observations, we analyzed the clinical outcome of patients with TP53m AML in a larger cohort treated over the last decade as novel therapies have been introduced to our therapeutic armamentarium. We conducted a multicenter observational study in collaboration with 8 U.S. academic centers, involving 291 TP53m AML patients diagnosed between November 2012 and June 2021 to analyze clinical characteristics and outcome based on different induction therapies. Patients with an initial diagnosis of AML were divided into 6 groups (GP) based on AML induction: 3 + 7 (GP1, n = 73), hypomethylating agents (HMA) based (GP2, n = 44), high-dose cytarabine based (HDAC) (GP3, n = 24), other low intensity chemotherapy (excluding HMA/venetoclax-based therapy) (GP4, n = 16), CPX-351 (GP5, n = 62), and HMA + venetoclax (GP6, n = 92), to analyze differences in outcome. Response to induction therapy was defined according to European Leukemia Net (ELN) consensus guidelines. Details about mutational and statistical analysis are provided in Data S1. Baseline characteristics across different induction regimen group are summarized in Table S1. The median age for the entire cohort was 65 (range, 18–88). Overall, 46%, 31%, and 22% of patients had de novo, secondary, or therapy-related AML. The proportion of patients with a TP53 VAF ≥40% was 69%, 54%, 50%, 64%, 56%, and 63% in GP 1–6, respectively (p = 0.70). Overall, the most common occurring co-mutations were TET2 (n = 23 [9%]), DNMT3A (n = 25 [9%]), and ASXL1 (n = 16 [6%]); these co-mutation occurrences were not significantly different among different treatment group (p = 0.33, 0.17, and 0.25, respectively). (Table S1, Figure S1). The overall complete remission (CR) rate (CR + CRi) was 29%, and it was not significantly different in different treatment GP (18%, 24%, 33%, 17%, 28%, and 41%; p = 0.10), respectively. Eleven (22%), 2 (4.5%), 6 (25%), 2 (13%), 14 (23%), and 11 (12%) patients were bridged to alloHCT after induction therapy from GP 1–6, respectively (p = 0.002) (Table S2). We performed subgroup analysis, excluding patient with secondary AML (n = 91); among 200 patients, the overall response rate (CR/CRi) was 28%, 15%, 18%, 31%, 29%, 31%, and 40% (from GP1-6, p = 0.11). In univariate analysis using logistic regression, HMA plus venetoclax combination regimens were associated with a superior CR/CRi rate when compared with 3 + 7 as a reference, with an odds ratio of 3.06 (95% CI: 1.34–7.54; p = 0.01). Induction with CPX-351 (p = 0.24) or HDAC based therapy (p = 0.16) were not significantly associated with achievement of complete remission (CR/CRi) (Table S3). The median EFS in months was 7.92 (6.05, 9.99), 7.89 (6.18, 10.12), 5.98 (4.76, 15.47), 1.95 (1.08, 14.366), 6.34 (5.19, 7.62), and 5.85 (3.75, 8.18) from GP1-6, respectively (p = 0.16; Figure 1A). In subgroup analysis, excluding patients with secondary AML, the median EFS based on induction therapy was 7.98 (6.14, 10.81), 6.34 (5.78, 12.48), 9.95 (4.76, 29.70), 1.89 (0.79, NA), 7.26 (5.62, 15.70), and 6.11 (3.84, 8.51) months from GP1-6, respectively (p = 0.14; Figure 1B). Among patients who achieved remission, median EFS was 6.37 months (95% CI: 5.72–9.17) (Figure S2a). Multivariate analysis was conducting including variables, which showed significance in univariate analysis for EFS, patients receiving alloHCT (HR = 0.43, 95% CI; 0.27–0.68; p ≤ 0.001) retained significance for better EFS. Whereas low-intensity chemotherapy (HR = 3.65, 95% CI; 1.88–7.09; p ≤ 0.001), complex CG (HR = 2.36, 95% CI; 1.48–3.78; p ≤ 0.001) and RAS co-mutation (HR = 3.11, 95% CI; 1.62–5.99; p ≤ 0.001) retained significance for inferior EFS (Table S4). The median OS in months was 8.54 (6.14, 10.05), 9.20 (7.10, 12.48), 9.51 (6.54, 18.92), 1.95 (1.08, 25.46), 6.67 (5.75, 10.09), and 6.70 (3.84, 9.13) from GP1-6, respectively (p = 0.41; Figure 1C). In subgroup analysis, excluding patients with secondary AML, the median OS based on induction therapy was 8.67 (6.14, 11.24), 9.56 (6.34, 13.40), 12.98 (7.16, 29.70), 1.89 (0.79, NA), 9.03 (5.91, 23.06), and 7.26 (5.09, 9.95) months from GP1-6, respectively (p = 0.38; Figure 1D). Among patients who achieved remission, the median OS was 11.5 months (95% CI: 9.59–20.20) (Figure S2b). We did an additional analysis to look for survival outcome of TP53m AML patients overtime with approval of newer therapies. Patients with an initial diagnosis of AML were divided into 4 time periods (TPs) based on the progressive use of novel therapies in clinical trials and their approvals as AML induction therapy during different time periods: 2012–2017 (TP1 [prior to CPX-351 approval], n = 68), 2018-June 2019 (TP2 [after approval of CPX-351], n = 112), July 2019–June 2020 (TP3 [utilization of venetoclax plus azacitdine in clinical trials or compassionate use], n = 66) and July 2020–June 2021 (TP4 [after approval of venetoclax plus azacitdine], n = 45), to analyze differences in survival (Figure S3). The median OS in months from TP1-4 was 10.1 (8.3,14.4), 6.6 (4.8–9.6), 7.3 (5.9,9.2), and 5.3 (2.6,9.0), respectively (p = 0.01; Figure S4). Multivariate analysis was conducting including variables, which showed significance in univariate analysis for EFS, patients receiving alloHCT (HR = 0.25, 95% CI; 0.15–0.42; p ≤ 0.001) retained significance for better OS, whereas low-intensity chemotherapy (HR = 3.65, 95% CI; 1.88–7.09; p ≤ 0.001), complex CG (HR = 2.78, 95% CI; 1.67–4.62; p ≤ 0.001), and RAS co-mutation (HR = 3.48, 95% CI; 1.81–6.71; p ≤ 0.001) retained significance for inferior OS (Table S5). In our multicenter observational study of a large cohort of patients with TP53m AML treated over the last decade, we observed a modest response to induction therapy and poor overall survival that has not improved in the era of novel therapies. However, patients able to proceed to alloHCT after responding to induction therapy were found to have improved overall survival. Data from the VIALE-A trial showed an inferior complete remission rate for patients with TP53m AML when compared to those with non-TP53m AML (55.3% vs. 64.7%).4 Similarly, retrospective and post-hoc analyses of patients treated with CPX-351 showed that TP53m AML predicts inferior responses to CPX-351 compared to TP53 wild-type AML and that the benefits of CPX-351 over 3 + 7 for TP53m AML may be abrogated.5 In our analysis, we observed that an increasing proportion of patients received either CPX-351 or HMA plus venetoclax combination therapy; however, this did not translate into improved outcomes, which is consistent with previous reports.4, 5 We acknowledge that response rate observed in our study were lower than published data, one reason could be that approximately 50% of patient had either therapy related or secondary AML (22% + 31% = 53%), respectively, which is relatively higher than published data.6 Secondly, higher proportion of patients in this cohort received venetoclax-based regimen with a relatively higher treatment related mortality resulting in lower response rates. One of the possible reasons for this could be prolong myelosuppression and associated complications with venetoclax-based regimens. Potentially, this can be overcome with careful selection of patients for induction therapy, limiting exposure of venetoclax during induction to 21 days rather than 28 days to allow count recovery. In our analysis, we observed that alloHCT was the only factor that improved survival and retained favorable significance in multivariate analysis. Our observation is consistent with most recent report evaluating outcome of TP53m AML patients.6 In view of recent data, we believe that until we get better and effective therapies to improve long-term outcome of TP53m AML, alloHCT should be considered in first remission in eligible patients. There are conflicting reports regarding significance of TP53m VAF in predicting outcome of MDS/AML patients.6 A recent study reported sub-optimal response to high-dose cytarabine-based induction and inferior survival in patients with TP53m and VAF > 40%.6 In our study, we did not find TP53m VAF < 40% versus ≥ 40% to independently predict response to therapy or outcomes. We believe that there are multiple factors predicting clinical outcome including advance age, co-occurrence of adverse risk mutations, availability of effective therapy, and eligibility for alloHCT rather than TP53m allelic burden alone. We acknowledge the limitations of our retrospective analysis including an imbalance in the number of patients in different GP and some heterogeneity in data from participating centers. Nevertheless, we present the largest experience of patients with TP53m AML analyzed by NGS, demonstrating universally dismal clinical outcome with some benefit of alloHCT in improving long-term survival. Talha Badar received Mayo Clinic Cancer Center support grant (P30 CA015083) and in advisory board for Pfizer; Aaron D. Goldberg received research funding from Celularity, ADC Therapeutics, Aprea, AROG, Pfizer, Prelude, and Trillium; received research funding from and served as a consultant for Aptose and Daiichi Sankyo; served as a consultant and member of advisory committees for Astellas, Celgene, and Genentech; received research funding from, served as a consultant for, and was a member of advisory committees for AbbVie; and received honoraria from Dava Oncology; Anand Patel received research funding from Celgene/BMS and Agios/Servier. None of these relationships were related to the development of this article. Talha Badar: Study design, writing–original draft, and methodology; Ehab Atallah: study design, methodology, contributing patients, and manuscript editing; Rory M.Shallis: contributed patients and manuscript editing; Aaron D. Goldberg: contributed patients and manuscript editing; Anand Patel: contributed patients and manuscript editing; Yasmin Abaza: contributed patients; Jan P. Bewersdorf: contributed patients; Antoine N. Saliba: contributed patients; Guilherme Sacchi De Camargo Correia: contributed patients; Guru Murthy: contributed patients; Adam Duvall: contributed patients; Madelyn Burkart: contributed patients; Maximilian Stahl: contributed patients; Yuanhang Liu and, Shira Dinner: contributed patients; Neil Palmisiano: contributed patients; Mark R. Litzow: contributed patients and manuscript editing; James M. Foran: contributed patients. The data that support the findings of this study are available from the corresponding author upon reasonable request. Figure S1. Heatmap illustration of concurrent mutations with variant allele frequencies. Figure S2. (a) Kaplan–Meier curve for event free survival and (b) overall survival in all TP53-mutated acute myeloid leukemia (AML) patient who achieved remission. Figure S3. Bar chart showing changes in induction therapies over time. Figure S4. Kaplan–Meier curve for overall survival in different time periods. Table S1. Baseline characteristics. Table S2. Responses based on induction therapies. Table S3. Prediction of complete response to induction: logistic regression. Table S4. Cox regression models predicting event free survival. Table S5. Cox regression models for overall survival. Data S1. Supporting Information. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.

Topics & Concepts