Delayed Hematologic Recovery in AML Patients after Induction Chemotherapy is Associated with Inferior Relapse-Free Survival and Persistence of Preleukemic Mutations
Tracy Murphy,Jinfeng Zou,Georgina S. Daher-Reyes,Vikas Gupta,Caroline McNamara,Mark D. Minden,Aaron D. Schimmer,Hassan Sibai,Karen Yee,Tracy Stockley,Suzanne Kamel‐Reid,Dawn Maze,Scott V. Bratman,Andre C. Schuh,Steven M. Chan
DOI: https://doi.org/10.1182/blood-2018-99-115518
IF: 20.3
2018-01-01
Blood
Abstract:Introduction:Induction chemotherapy debulks the leukemic burden in AML patients. Blood count recovery usually occurs during the fourth week of starting chemotherapy in patients who achieve a morphologic remission in bone marrow. However, a subset of patients experience significantly delayed recovery. The relevance of delayed recovery on long-term clinical outcomes and its contributing factors have not been well studied. Specifically, the association between recurrent mutations in AML and hematologic recovery is unknown. Methods:We studied a total of 262 newly diagnosed adult AML patients treated between September 2014 and December 2017 at Princess Margaret Cancer Centre who achieved a complete remission (CR) or CR with incomplete count recovery (CRi) after one cycle of induction chemotherapy. The regimens consisted of 3+7 (N=194) and FLAG-IDA (N=68). We collected information on disease characteristics and blood count results at baseline and during chemotherapy. Mutation profiling of diagnostic samples was performed using a 54-gene next generation sequencing panel (TruSight Myeloid Sequencing Panel, Illumina). Detection of persistent mutations in remission samples was performed using a custom 37-gene duplex sequencing platform with a lower detection limit of ~0.05% variant allele frequency (VAF). Results:Of the cohort of 262 patients, 256 patients (97.7%) achieved neutrophil recovery (defined as > 1x109/L), with time to recovery ranging from 17 to 84 days. Two hundred forty-four (93.1%) patients achieved platelet recovery (defined as > 100x109/L); time to recovery ranged from 17 to 117 days. The percentage of patients who achieved neutrophil and platelet count recovery before day 35 was 82.4% and 84.0% respectively (Fig. 1). To evaluate the prognostic significance of delayed recovery, we categorized patients who achieved CR into two groups, "normal" or "delayed" recovery, according to whether they achieved recovery before or after day 35, respectively. Relapse-free survival (RFS) of patients with delayed recovery was significantly worse than those with normal recovery and only marginally better than those with CRi (P=0.02; Fig. 2). Analysis restricted to 3+7 treated patients showed the same trend (P=0.02), excluding the possibility that the inferior outcome was due to treatment of higher risk patients with more intensive regimens. To study the factors associated with delayed recovery, we performed multivariable Cox regression analysis that included clinical factors and mutations identified at the time of diagnosis as covariates. Four factors were found to be independently correlated with delayed recovery: treatment with FLAG-IDA, truncating ASXL1mutations, SRSF2mutations, and DNMT3AR882 mutations (Table 1). Because FLAG-IDA is the preferred frontline regimen for higher risk patients at our institution, we performed a secondary analysis restricted to patients treated with 3+7 to exclude chemotherapy regimen as a potential confounding variable. This analysis identified six independent factors: AML with myelodysplasia-related changes, lower hemoglobin levels at presentation, truncating ASXL1mutations, TET2mutations, CEBPAmutations, and DNMT3AR882 mutations (Table 1). Somatic mutations in DNMT3A, TET2, ASXL1, and SRSF2(DTAS) mutations are associated with preleukemic conditions, such as myelodysplastic syndrome and age-related clonal hematopoiesis, and frequently persist in remission. These mutations are acquired in hematopoietic stem cells resulting in their propagation to progenitors and terminally differentiated blood cells. We hypothesized that the persistence of DTAS mutations in progenitors might compromise their capacity for reconstitution of normal hematopoiesis resulting in delayed recovery. To test this hypothesis, we performed duplex sequencing on peripheral blood DNA samples collected from a random subset of 43 patients during remission. The detection of DTAS mutations in remission above a VAF of 2% was strongly associated with delayed recovery (P=0.0004; Fig. 3). Conclusion:Delayed hematologic recovery in AML patients after induction chemotherapy is associated with inferior RFS and persistence of preleukemic mutations (i.e., DTAS mutations). Our results support a model in which progenitors harboring DTAS mutations have reduced repopulation capacity leading to delayed hematologic recovery after induction chemotherapy. Disclosures Gupta: Incyte: Research Funding; Novartis: Consultancy, Honoraria, Research Funding. Schimmer:Otsuka Pharmaceuticals: Consultancy; Jazz Pharmaceuticals: Consultancy; Novartis: Consultancy, Membership on an entity's Board of Directors or advisory committees; Medivir AB: Research Funding. Yee:Agensys, Astex, GSK, Onconova, Genentech/Roche: Research Funding; Celgene, Novartis, Otsuka: Membership on an entity's Board of Directors or advisory committees. Maze:Novartis: Consultancy, Honoraria. Bratman:Roche: Other: SVB is a co-inventor on a patent describing methods for circulating tumor DNA analysis, which has been licensed to Roche Molecular Diagnostics.. Schuh:Shire: Consultancy; Jazz: Consultancy; Novartis: Consultancy; Otsuka: Consultancy; Teva: Consultancy; Pfizer: Consultancy; Celgene: Consultancy; Amgen Inc.: Consultancy.