Utility of Combined DNA and RNA Next Generation Sequencing in Leukemias for Identification of Prognostic and Therapeutically Relevant Genomic Alterations in Clinical Practice

Genome and exome sequencing have provided important insights into the biology of hematologic malignancies, and have led to the identification of prognostically relevant mutations and therapeutically targetable genomic alterations. However, the utility of genomic profiling in routine clinical practice remains an area of active investigation, and there is a need for evidence-based approaches to application of these data. Although most adults with acute leukemia relapse after initial therapy, only 25% of acute leukemia patients are enrolled onto clinical trials. Reasons for limited enrollment are multifold, but include a limited understanding by clinicians of genomic alterations, and lack of access to broad-based analytically validated clinical genomic profiling, both hampering the ability of clinicians to enroll patients on genotype-specific, mechanism-based clinical trials. As such, the development of high throughput, state-of-the-art genomic profiling assays for clinical practice, such as FoundationOneHeme, provide an unprecedented opportunity. Here we analyzed a consecutive series of 62 leukemia patients at Memorial Sloan Kettering Cancer Center for whom FoundationOneHeme was performed to determine the incidence and applicability of genomic findings on clinical care.

The clinical sample cohort included 17 patients with myelodysplastic syndromes (MDS), 14 patients with acute myeloid leukemia (AML), 14 patients with acute lymphoblastic leukemia (ALL), 6 patients with myeloproliferative neoplasms (MPN), 5 patients with chronic lymphocytic leukemia (CLL), 4 patients with aplastic anemia (AA), and 2 patients with chronic myeloid leukemias (CML). DNA and RNA were successfully extracted from 58/62 samples (94%). Adaptor ligated sequencing libraries were captured by solution hybridization using a custom bait-set targeting 405 cancer-related genes and 31 frequently rearranged genes by DNA-seq, and 265 frequently-rearranged genes by RNA-seq. Samples were sequenced to high depth (Illumina HiSeq) in a CLIA-certified CAP-accredited laboratory (Foundation Medicine), averaging 590x for DNA and >20M total pairs for RNA, to enable the sensitive and specific detection of short variants (substitutions and indels), CNAs and gene fusions. A genomic alteration was characterized as “therapeutic” if it has been shown to determine response or resistance to an available therapy, or mechanism-based trial and “prognostic” if it has been shown to be predictive of outcome by individual disease states.

Of the 58 patients with informative genomic data, 84% had at least one pathogenic variant identified. A total of 154 such alterations were identified (2.7 alterations per sample), including 81 base substitutions, 39 indels, 6 splice mutations, 10 CNAs and 18 fusions/rearrangements. 69% of patients had variants identified that were classified as clinically significant. 53% of patients had potential therapeutically relevant genomic alterations, and 43% of patients had prognostically relevant alterations identified. 36% of patients had both therapeutic and prognostic alterations identified. The most common alterations identified in our cohort were TP53 (n= 8), NRAS (n= 6), KMT2A, KRAS, RUNX1 (n= 5), CDKN2A, CDKN2B, SF3B1, TET2 (n= 4). In sum, 54 prognostic and 47 therapeutic genomic alterations were identified. 7 MDS cases, 4 AA cases, 2 AML cases, 2 ALL cases, 1 CML case, and 1 CLL case did not have a pathogenic or actionable allele, demonstrating that almost all patients in our cohort had mutations with biologic and clinical relevance. We also noted a set of variants of unknown significance, including 37 deletions and 26 insertions, which are under investigation in regard to prognostic and therapeutic significance. However, given that at least one genomic variant was identified in the majority of patients, our data demonstrate the utility of this approach in the identification of somatic genomic alterations for prognostic and therapeutic value, and to identify clonal markers which can be used to track molecular response during anti-leukemic therapy.

In summary, we analyzed the mutational profiles of leukemia patients using an analytic and clinically validated assay, which allowed us to identify both prognostic and therapeutically relevant mutations in the majority of patients, confirming the utility of comprehensive next-generation sequencing in clinical practice.