Genomic profile at diagnosis by targeted high throughput sequencing, and real-world clinical outcomes in patients diagnosed with acute myeloid leukemia at a tertiary care cancer center.

Abstract

e19515 Background: Acute Myeloid Leukemia (AML) is a highly heterogeneous disease. High throughput sequencing (HTS) has led to the discovery of a number of recurrently mutated genes in AML & has provided a detailed molecular landscape of the disease. This information is now extensively utilised & interpreted in AML. Methods: Retrospective analysis of 47 patients with newly diagnosed AML for those who had both cytogenetic and targeted HTS information at diagnosis, [median age: 44y (12-80y); 24 females] & treated at Tata Medical Center, India between 2013-2019. HTS was performed using Illumina Trusight myeloid panel and/or Thermo Fisher Scientific Oncomine Myeloid panel, and mutation data from the common 28 genes were used. Patients were stratified by European Leukemia Net (ELN) 2010 and 2017 guidelines. Standard of care clinical outcome data was available for all patients. Results: 45 of 47 patients had mutations. The most frequently mutated gene was FLT3 (n = 13, 28.9%), followed by NPM1 (n = 8, 17.8%), and others [ IDH1, IDH2, TET2, CEBPA, and NRAS; n = 7, 15.6%). ELN-2010 classified 6 patients as Favourable (12.8%), 10 as Intermediate-I (21.3%), 25 as Intermediate-II (53.2%) and 6 as Adverse (12.8%) risk groups. One patient migrated to the adverse group from Intermediate-II when ELN-2017 was applied. ELN 2017 re-classification: Favorable (n = 6, 12.8%), intermediate (n = 34, 72.3%), and adverse risk groups (n = 7, 14.9%). Co-occurrence of driver gene mutations was found in NPM1 and FLT3 (8.2%) frequently. Patients with mutations in epigenetic modifiers ( ASXL1, IDH1, IDH2, DNMT3A) had high Variant Allele Frequency (VAF, 40-50%) while those with signaling pathway gene mutations ( NRAS, KRAS) had low VAF (9-25%). At a median follow up of 3.25 years, median overall survival in the three risk groups (favorable, intermediate, adverse) were 11.25 mons, 7 mons and 8 mons respectively, and relapse rates were 16%, 29% and 42% respectively. Conclusions: Early experience from a middle-income real-world setting suggests feasibility of integration of HTS information in practise, as additional prognostic/ predictive information. The pattern of a high VAF in epigenetic modifier gene mutations, potentially as a result of early acquisition during clonal development, needs further evaluation. This data set has helped us initiate a prospective multimodal omics study at our institution.