Minimal Residual Disease Monitoring By Next-Generation Sequencing in Patients with Acute Myeloid Leukemia: MRD Positivity after First Consolidation Chemotherapy Can Better Predict Clinical Outcomes Than That after Induction Chemotherapy
Cheng-Hong Tsai,Jih-Luh Tang,Feng-Ming Tien,Yuan-Yeh Kuo,Chien-Chin Lin,Mei-Hsuan Tseng,Yen-Ling Peng,Ming-Chih Liu,Chia-Wen Liu,Ming Yao,Liang-In Lin,Wen-Chien Chou,Chien-Yu Chen,Hsin-An Hou,Hwei-Fang Tien
DOI: https://doi.org/10.1182/blood-2019-126870
IF: 20.3
2019-11-13
Blood
Abstract:Introduction Presence of minimal residual disease (MRD) detected by multicolor flow cytometry (MCFC) or quantitative polymerase chain reaction has been recognized as an independent important prognosticator for patients with acute myeloid leukemia (AML). Next-generation sequencing (NGS) can simultaneously detect various mutations and be applied to the majority of patients with AML, but the clinical implication of its use in MRD monitoring remains to be clarified. Recently, it was shown that NGS MRD of mutants other than the common mutations occurring in clonal hematopoiesis of indeterminate potential, including the DTA (DNMT3A, TET2, and ASXL1) mutations, carry prognostic impacts on relapse rates and overall survival (OS) in AML patients. However, the proper time point for NGS MRD detection after treatment is still unclear. Our hypothesis is that the NGS MRD detected at different time points might have different clinical implications. In this regard, we aimed to explore the clinical implication of NGS MRD at different time points in AML patients after chemotherapy. Method We enrolled 306 de novo non-M3 and non-M6 AML patients who attained complete remission (CR) after standard induction chemotherapy and received 2-4 courses of post-remission chemotherapy with high-dose cytarabine with or without anthracycline. We analyzed bone marrow samples serially collected at diagnosis, first CR (1st time point for MRD analysis), and after the first consolidation chemotherapy (2nd time point). We used the TruSight myeloid panel (Illumina) to survey the 54 genes related to myeloid malignancies. Because of the sequencing sensitivity issue, we excluded CEBPA mutation and FLT3-ITD in the subsequent analyses. The median follow-up time was 92.0 months. Result At diagnosis, 91% of patients had at least one gene mutation with a median of 2.0 mutations (range 1-6) per patient; 49.4% had molecular gene mutations alone and 41.6% had both cytogenetic changes and molecular mutations. Mutations in NPM1, DNMT3A, NRAS and IDH2 were the most common mutations. According to the 2017 ELN recommendation, 49.3% of patients were in the favorable-risk group; 29.1%, the intermediate-risk group; and 21.6%, the unfavorable-risk group. Among the patients harboring at least one gene mutation at diagnosis, we randomly assigned them into the training (n=167) and validation cohort (n=111); the two cohorts had similar clinical features, and distribution of cytogenetic and molecular abnormalities. Based on the result from the analysis in the training cohort, we set 0.3% as the cut-off for MRD positivity because patients carried gene mutations lower than this limit had a similar outcome as those without detectable mutations. The allele frequencies of the mutants in MRD ranged from 0.3 to 50.5%. Excluding DTA mutations, 47.3% patients in the training cohort had MRD at 1st time point, and 26.9% at 2nd time point. The patients with positive NGS MRD had significantly higher relapse rate (P=0.042 for 1st MRD and P=0.035 for 2nd MRD), shorter disease-free survival (DFS, P=0.037 for 1st MRD and P=0.007 for 2nd MRD) and OS (P=0.015 for 1st MRD and P<0.001 for 2nd MRD, Figure 1). In multivariate Cox proportional hazards regression model incorporating age, white blood cell counts at diagnosis, transplantation status, 2017 ELN risk-stratification, number of chemotherapy cycles to attain CR, and the MRD status into analyses (Table 1), the 2nd MRD was an independent poor prognostic factor (P=0.040 for DFS and P=0.005 for OS) but not 1st MRD (P=0.113 for DFS and P=0.072 for OS). In the validation cohort, 2nd MRD positivity also predicted poorer OS and DFS (P=0.023 and P<0.001) but not 1st MRD (P=0.996 and P=0.461). A comparison of NGS with MCFC for the detection of MRD in 73 patients showed that MRD by NGS had significant additive prognostic value. Conclusion NGS-based MRD monitoring can be applied to more than 90% of AML patients who have detectable mutations at diagnosis. The presence of NGS MRD after treatment can predict outcome of AML patients, especially after the first consolidation chemotherapy (2nd MRD). Positivity of 2nd MRD is an independent unfavorable prognostic factor for DFS and OS. Further prospective trials are warranted to validate these findings and to clarify the role of pre-emptive treatment. Disclosures Tsai: Celgene: Research Funding; Astellas, BMS, Celgene, Chugai, Johnson & Johnson, Kirin, Novartis, Pfizer, Roche, Takeda: Honoraria. Tien:Novartis: Other: Travel Grant. Hou:Celgene: Research Funding; Abbvie, Astellas, BMS, Celgene, Chugai, Daiichi Sankyo, IQVIA, Johnson & Johnson, Kirin, Merck Sharp & Dohme, Novartis, Pfizer, PharmaEssential, Roche, Takeda: Honoraria. Tien:Celgene: Honoraria; Novartis: Honoraria; Alexion: Honoraria; BMS: Honoraria; Roche: Research Funding; Pfizer: Honoraria; Roche: Honoraria; Celgene: Research Funding; Abbvie: Honoraria; Johnson &Johnson: Honoraria; Daiichi Sankyo: Honoraria.
hematology