The prognosis and durable clearance of <scp>RAS</scp> mutations in patients with acute myeloid leukemia receiving induction chemotherapy

Abstract

The RAS oncogenes, NRAS and KRAS are frequently mutated in AML, occurring in 11% and 5% of patients, respectively.1 These gain-of-function mutations produce altered RAS-GTPase proteins locked in an active GTP-bound state resulting in constitutive activation of the mitogen activated protein kinase (MAPK), and phosphoinositide-3 kinase (PI3K) pathways that impact cell proliferation and survival. Although among patients receiving induction chemotherapy, the presence of RAS mutations do not significantly impact prognosis,1 mounting evidence suggests that RAS-pathway mutated leukemic clones are more likely to be cleared. Targeted next generation sequencing performed at diagnosis and at time of complete remission (CR) after induction chemotherapy revealed that RAS-pathway mutations including NRAS, KRAS, NF1, PTPN11 had higher clearance rates relative to other mutations associated with clonal hematopoiesis of indeterminate potential (CHIP), DNA methylation, and RNA splicing.2, 3 The addition of intensive chemotherapy represents a promising strategy in combination with non-intensive therapies, including targeted therapies of IDH and FLT3 and venetoclax-based regimens, where RAS mutations have emerged as a common mechanism of resistance.4-6 In order for this strategy to be effective additional sequencing at time of relapse is necessary to determine if RAS mutations are persistently cleared after treatment with induction chemotherapy. Here, we performed a single center retrospective study in an unselected adult (age > 18 years) AML population receiving induction chemotherapy. This was to determine the prognosis of NRAS or KRAS mutations and analyze paired NGS at diagnosis and relapse, to evaluate the persistence of RAS-mutant clones after treatment with induction chemotherapy. There were 232 patients with a diagnosis of AML, who were treated with intensive induction chemotherapy from April 1, 2014 to May 15, 2019 and had next-generation sequencing performed prior to treatment were included for analysis [Figure S1]. Over the duration of the study, NGS was performed using one of three sequencing panels: 30-gene panel, 49-gene panel or a 400-gene panel. The details of these assays including the minimum variant allele fraction for mutation calling are reported in the supplemental material [supplemental methods]. A total of 196 patients were WT, 20 patients (9%) were NRAS-mutated, 11 patients (5%) KRAS-mutated, and five patients (2%) were both NRAS and KRAS-mutated. A total of 510 mutations were detected including 27 NRAS mutations (mutation frequency 5%) and 19 KRAS mutations (mutation frequency 4%). The RAS mutations occurred at typical hotspot locations: G12, G13, Q61 [Figure S2]. Baseline characteristics of RAS and WT AML are listed in Table S1. When compared to WT patients, NRAS-mutated patients were associated with a higher proportion with at least four co-mutations (WT 13%, NRAS 30%) and AML with myelodysplastic related changes (WT 28%, NRAS 55%). Characteristics of the KRAS-mutated patients included older median age (WT 60 years, KRAS 64 years), higher median WBC (WT 3 x 109 cells/L, KRAS 13 x 109 cells/L) secondary AML or therapy-related AML (WT 41% vs KRAS 73%), inversion 16 (WT 3%, KRAS 18%), and KMT2A or 11q23 rearrangements (WT 2%, KRAS 36%) at baseline. We also observed that WT1 mutations were more prevalent among RAS-mutated patients, occurring in six out 36 patients (17%). Induction regimens consisted of intravenous cytarabine 100 mg/m2 on days 1–7 and an anthracycline on days 1–3 (idarubicin 12 mg/m2 (6%), daunorubicin 60 mg/m2 (47%), daunorubicin 90 mg/m2 (33%)) or in a liposomal formulation of CPX-351 (14%). Forty patients received combinations of targeted therapy and induction chemotherapy, including midostaurin (n = 12), crenolanib (n = 7), gilteritinib (n = 1), ivosidenib (n = 10), and enasidenib (n = 10) for Flt3, IDH1-, and IDH2-mutated AML, respectively. The type of induction or proportion of patients receiving targeted therapy and induction chemotherapy was similar among RAS and WT groups [Table S1]. Seventy-six patients (33%) received reinduction chemotherapy, most commonly with regimens that contained a cytarabine dose >1 g/m2 [Table S1]. After induction and reinduction chemotherapy, KRAS-mutated patients had lower rates of CR/CRi (WT: 59%, KRAS: 36%) and higher rates of 30-day mortality when compared to WT (WT 3%, KRAS 27%) [Table S2]. The rates of allogeneic stem cell transplantation were also decreased among KRAS-mutated patients (WT 56%, KRAS 27%) [Table S1]. Rates of MRD negativity by multiparameter flow cytometry were similar between WT, KRAS and NRAS-mutated patients and highest in patients with co-occurring KRAS and NRAS mutations [Table S1]. The median follow-up after induction was 21.3 months (range 1.22–58.66 months) among survivors. The median EFS was decreased in KRAS-mutated but not NRAS-mutated patients when compared to WT patients (KRAS 2.9 months vs WT 13.8 months, p < .001) [Figure 1(A)]. The median OS was significantly decreased in KRAS-mutated but not NRAS-mutated patients relative to WT AML patients (KRAS 5.2 vs WT 30.1 months, p < .001) [Figure 1(B)]. We also assessed the impact of tumor mutational burden on relapse risk following response to induction. The 12-month CIR was significantly increased in patients with four or more co-mutations relative to patients with less than four co mutations (≥4 co mutations 50% vs.˂ 4 co mutations 20.4%, p = .001) [Figure 1(C)]. Paired next generation sequencing obtained from bone marrow biopsies at diagnosis and at time of relapse in patients with detectable mutations at diagnosis was available in 35 out of 42 (88%) patients. Durable mutation clearance below a VAF of 1% (MC 1%) occurred in 42 out of the 128 mutations (33%) detected at diagnosis at time of relapse [Table S3]. Out of eight NRAS mutations and one KRAS mutation detected at diagnosis, only two mutations, NRAS Q61R and NRAS Q61K mutation were detected at relapse (Figure 1(D),(E)). Similarly, other RAS-pathway genes including PTPN11 and CBL had a high rate of MC1% with six out of six (100%) and two out of three (66%) mutations cleared respectively. At time of relapse, clearance of at least one mutation below a VAF of 1% was significantly greater for RAS-pathway genes, which included NRAS, KRAS, PTPN11, and CBL genes than non-RAS-pathway genes (MC 1% RAS-pathway vs non-RAS pathway: 81% vs 38%, p = .011) [Table S4]. Among the 35 relapsing patients, 30 new mutations were detected at time of relapse on next generation sequencing. The acquisition of one NRAS Q61H and one KRAS G13D mutation occurred in two patients at time of relapse. For the entire cohort, the most common mutation acquired at time of relapse was TET2 (n = 7) followed by FLT3-ITD (n = 5), TP53 (n = 3), and RUNX1 (n = 3) [Figure 1(D)]. Previous studies have shown high clearance rates of RAS mutations at time of complete remission in patients responding to induction chemotherapy.2, 3 Herein, we demonstrate that RAS mutation clearance persists at time of relapse with only two of the nine RAS mutations found at diagnosis being detected at time of relapse. We also observed high rates of mutation clearance of other RAS-pathway genes including PTPN11 and CBL. Furthermore, the rate of mutation clearance at time of relapse was significantly higher for RAS-pathway genes when compared to mutations of other gene pathways. The high mutation clearance rate after induction chemotherapy at time of relapse suggest that RAS mutations are not a driver of relapse but may be associated with other high-risk molecular abnormalities. We observed that KRAS mutations were more frequent in patients with MLL-rearrangements. Consistent with prior reports, we observed that NRAS mutations were associated with a higher co-mutational burden.6 Furthermore, we found that patients with four or more co-mutations had a higher CIR after response than patients with less than four co-mutations. Due to increased clonal complexity, RAS-mutant AML may require therapeutic strategies capable of eliminating RAS- and non-RAS-mutated leukemic clones. Consistent with previous studies, we have shown that NRAS mutations did not significantly impact survival.1 In contrast, we observed that KRAS mutations were associated with a shorter median OS and EFS. Bowen et al., performed the largest study to date evaluating the impact of RAS mutations and found no difference in survival between KRAS and WT AML. The difference in outcome is likely due to patient selection. Whereas Bowen et al. included younger patients (<60 years) enrolled on clinical trials, the KRAS-mutated cohort in our study was older and had a higher proportion of patients with sAML or t-AML, KMT2A translocations, and prior treatment with hypomethylating agents.1 In our study, the KMT2A rearrangements occurred in 36% of KRAS-mutated patients. These high-risk characteristics translated to a higher rate of early mortality and lower rates of alloSCT among KRAS-mutated patients. We acknowledge several limitations of our study. Our analysis on the impact of NRAS and KRAS mutations after induction chemotherapy was limited by the small number of patients with these mutations. Additionally, the study cohort included patients from a large tertiary cancer center, which may have enriched for a greater proportion of patients with sAML or t-AML. Lastly, we acknowledge that a minimum VAF of 1% may not be sensitive enough to rule out persistence of subclonal RAS mutations. We detected a new NRAS and KRAS mutation at time of relapse that may have been detectable at diagnosis with more sensitive DNA sequencing methods such as error-corrected sequencing or single-cell sequencing. However, the high rates of durable RAS-pathway mutation clearance at time of relapse supports further studies evaluating chemotherapy in combination with targeted therapies in RAS-pathway mutated AML. Serial NGS analyses performed in clinical trials evaluating chemotherapy in combination with therapies targeting IDH, FLT3, or Bcl-2 will ultimately determine if the combination is capable of eliminating the RAS-mutant leukemic clones that mediate resistance to each therapy as monotherapy or in less intensive regimens. This study was supported in part by National Institutes of Health (NIH) award number P01 CA23766 and NIH/National Cancer Institute Cancer Center support grant P30 CA008748. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. The authors would like to thank all the physicians and the medical staff at Memorial Sloan Kettering Cancer Center that took care of the patients and helped to collect data. A.D.G. served on advisory boards or as a consultant for AbbVie, Aptose, Celgene, Daiichi Sanyko, Genentech, received research funding (institutional) from AbbVie, ADC Therapeutics, Aprea, AROG, Daiichi Sanyko, Pfizer, received Honoraria from Dava Oncology; J.L.G. served as a consultant for Gerson Lehman Group; S.F.C. has consulted for Imago Biosciences; A.D.V. is supported by a National Cancer Institute career development grant K08 CA215317, the William Raveis Charitable Fund Fellowship of the Damon Runyon Cancer Research Foundation (DRG 117–15), and an EvansMDS Young Investigator grant from the Edward P. Evans Foundation, received travel support from Mission Bio and is on the Editorial Advisory Board of Hematology News; S.A.G. receives research funding from Miltenyi Biotec, Takeda Pharmaceutical Co., Celgene Corp., Amgen Inc., Sanofi, Johnson and Johnson Inc., Omeros Corp., and Actinium Pharmaceuticals Inc., and is on the advisory boards for Kite Pharmaceuticals Inc., Celgene Corp., Sanofi, Novartis, Johnson and Johnson Inc., Amgen Inc., Janssen, Takeda Pharmaceutical Co., Jazz Pharmaceuticals Inc., SPECTRUM Pharma, and Actinium Pharmaceuticals Inc; B.G. receives research funding from Actinium Pharmaceuticals Inc. for a clinical trial; R.L. is on the supervisory board of Qiagen and is a scientific advisor to Loxo (until 2/2019), Imago, Ajax, Zentalis, Auron, Mana, C4 Therapeutics and Isoplexis, which each include an equity interest. He receives research support from and consulted for Celgene and Roche, he has received research support from Prelude Therapeutics and he has consulted for Lilly, Incyte, Novartis, Astellas, Morphosys and Janssen; M.T. receives research funding from ADC Therapeutics, Biosight, Abbvie, Cellerant, Orsenix, Glycomimetics, Raphael and Amgen, served on the advisory board for Rigel, Nohla, BioLineRx, Oncolyze, Delta Fly Pharma, Daiichi-Sankyo, KAHR, Abbvie, Orsenix, Tetraphase, Jazz, Roche, receives royalties from UpToDate; E.S. served on board of directors or advisory committee for Agios, Astellas Pharma US, Celgene, Daiichi Sankyo, Genentech, Novartis, PTC Therapeutics, Syros, Bioline, and Biotheryx and served as a consultant for Agios. Brian J. Ball, Eytan M. Stein designed the study, collected and interpreted data and wrote the manuscript; Meier Hsu and Sean M. Devlin performed the statistical analysis and analyzed and interpreted data; and all authors Brian J. Ball, Meier Hsu, Sean M. Devlin, Maria E. Arcila, Mikhail Roshal, Yanming Zhang, Chris Famulare, Aaron D. Goldberg, Sheng Cai, Andrew Dunbar, Zachary Epstein-Peterson, Kamal N. Menghrajani, Jacob L. Glass, Justin Taylor, Aaron D. Viny, Sergio S. Giralt, Boglarka Gyurkocza, Brian C. Shaffer, Roni Tamari, Ross L. Levine, Martin S. Tallman, and Eytan M. Stein performed research, collected data, edited and approved the manuscript. The data that support the findings of this study are available from the corresponding author upon reasonable request. Appendix 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