Monitoring Minimal Residual Disease in Acute Myeloid Leukemia Using Genomic or cfDNA with MyMRD®, a Targeted NGS Panel

Acute myeloid leukemia (AML) is a genetically and phenotypically heterogeneous disorder. Precision therapies for AML have been developed that target specific driver mutations. The efficacies of these therapies are variable, making it critical to determine successful therapies prior to patient relapse. For patients achieving a first complete remission, minimum residual disease (MRD) is an important prognostic factor, as MRD may provide a powerful and timely tool to evaluate therapeutic efficacy. There is a growing demand that new and promising drugs are approved as quickly as possible and accelerated approvals will require biomarker surrogate endpoints, such as MRD, rather than long-term survival endpoints. We have developed a sensitive NGS gene panel (MyMRD®), which identifies pathogenic variants in AML. The MyMRD panel targets single nucleotide variants (SNVs), insertions and deletions (indels) in coding exons of hotspots of 21 genes (ASXL1 BRAF CALR CEBPA CSF3R DNMT3A FLT3 IDH1 IDH2 JAK2 KIT KRAS MPL NPM1 NRAS PTPN11 RUNX1 SF3B1 SRSF2 TP53 ZRSR2), and structural variants of potential genomic breakpoint hotspots within 3 somatic gene fusion partners (CBFB-MYH11 KMT2A RUNX1-RUNX1T1). This 23 gene targeted panel can identify driver mutations that cause relapse in >90% of all AML patients, as well as common drivers in other myeloid neoplasms and myelodysplasic syndromes. MyMRD panel validation included determining the limit of blank (LoB), limit of detection (LoD), and linearity. Validation samples were generated using known variant containing DNA from cell lines and clinical samples diluted into NA12878 (“Genome in a Bottle”) DNA. The average background, crossover, and carryover rates were determined to be 0.017%, 0.023%, and 0.016%, respectively. The LoB was determined to be 0.13% from calculating each of expected variants using the 95th percentile of all negative samples. Overall, we established an LoD of 0.5% for >95% of the targeted SNV and indel sites in the assay with lower LoDs for specific mutations of interest, such as 0.17% for FLT3 TKD and 0.34% for NPM1. The LoD for structure variants was determined to be 1.8%. The assay shows strong linearity with R2 = 0.969 - 0.994 of 11 selected targets in the entire range (0.17% - 50%) of variant allele frequencies (VAFs) tested. Linearity and LoD determined from clinical samples containing SNVs and indels were consistent with conclusions obtained from contrived cell line samples. Cell-free DNA (cfDNA) isola