Utilization of next-generation sequencing (NGS) in clinical hematologic oncology practice is rapidly rising and may help further our knowledge in the diagnosis, treatment, and prognosis of these complex diseases.
An estimated combined total of 174,250 people in the U.S. are expected to be diagnosed with leukemia, lymphoma, or myeloma in 2018 (Cancer Facts and Figures [American Cancer Society; 2018]). These new cases are expected to represent 10 percent of all new U.S. cancer cases diagnosed in 2018.
In addition to advances in treatment, including chemotherapy, immunotherapy, and stem cell transplantation, genomics can be helpful in characterizing these diseases further, as well as potentially identifying other targeted treatment options. Though the role of NGS in hematologic malignancies can be expected to keep evolving, this overview will examine the current relevance of NGS in hematologic malignancies.
Genomics in Acute Myeloid Leukemia
NGS offers the ability to measure somatic allele frequencies from the complete coding sequences of many genes in the same assay, which is more comprehensive than traditional molecular assays that test only a relatively small panel of commonly mutated sites (Clin Transl Sci 2016;9(6):283-292).
Previous studies have shown that acute myeloid leukemia (AML) is a complex disease characterized by multiple somatically acquired driver mutations, coexisting competing clones, and evolution over time (Nature 2012;481(7382):506-510, N Engl J Med 2012;366(12):1079-1089, Cell 2012;150(2):264-278).
NGS has helped identify the number of known mutations in patients with hematologic malignancies. The Cancer Genome Atlas (TCGA) project included data identified by either whole-genome or whole-exome sequencing from 200 patients with de novo AML. Twenty-three genes were found to be recurrently mutated, including such well-known genes as NPM1, FLT3, CEBPA, DNMT3A, IDH1, and IDH2 (N Engl J Med 2013;368(22):2059-2074).
Further studies have confirmed multiple recurrent somatic mutations in genes involved in chromatin modification (EZH2, ASXL1), DNA methylation (DNMT3A, TET2, IDH1/2), RNA splicing (SF3B1, U2AF1, SRSF2, and ZRSR2), signal transduction (JAK2, KRAS, CBL, FLT3), and transcriptional regulation (EVI1, RUNX1, GATA2), and the mutations have been the subject of study by investigators interested in many myeloid and lymphoid malignancies (Crit Rev Oncol Hematol 2018;126:64-79).
Genomic profiling has also helped describe clonal evolution in these myeloid malignancies. The TCGA project researchers demonstrated that more than half of the 200 patients exhibited at least one subclone in addition to a founding leukemia clone (N Engl J Med 2013;368(22):2059-2074). This helps support the concept that mutations in genes involved in epigenetic regulation (DNMT3A, TET2, and ASXL1) occur as early founder events in preleukemic progenitor cells before leukemogenic events (NPM1 or signaling molecules) (J Clin Oncol 2017;35(9):934-946).
Another concept that has evolved from the NGS era is the “clonal hematopoiesis of indeterminate potential (CHIP)” (Blood 2015;126(1):9-16). This phrase describes the presence of hematologic malignancy-associated somatic mutations in blood or bone marrow in the absence of conventional diagnostic criteria for a hematologic malignancy (J Clin Oncol 2017;35(9):934-946, Blood 2015;126(1):9-16). This may be found in 10 percent of individuals age 70 years old or older and up to 20 percent of those individuals age 90 or older (N Engl J Med 2014;371(26):2488-2498). Early data suggest the transformation rate of CHIP into a hematologic disease is 0.5-1 percent per year, similar to the transformation rate of monoclonal gammopathy of undetermined significance to multiple myeloma (J Clin Oncol 2017;35(9):934-946).
NGS studies have helped develop a broader understanding of the biology of leukemias. In addition to using chromosomal translocations and inversions for classifications, gene mutations can also help inform disease classification. In a study of 1,540 patients with AML, analysis by targeted resequencing of 111 myeloid cancer genes along with cytogenetic profiles segregated patients with AML into 11 nonoverlapping classes, each with a distinct clinical phenotype and outcome (N Engl J Med 2016;374(23):2209-2221). In addition, NGS has also helped monitor minimal residual disease at early time points to help identify patients at high risk for relapse (Leukemia 2014;28(1):129-137).
Genomics in Lymphoma
Diffuse large B-cell lymphoma (DLBCL) is the most common form of adult lymphoma worldwide, accounting for 30-40 percent of newly diagnosed non-Hodgkin lymphoma (WHO Classification of Tumours of Haematopoietic and Lymphoid Tissues [Lyon, France: IARC;2008]).
Collaborators on a Lymphoma Study Association project developed a Lymphopanel NGS assay. Designed to identify mutations in 34 genes important for lymphoma genesis, it was based on a literature review of whole-exome sequencing studies in DLBCL as well as relapsed or refractory DLBCL cases (Clin Cancer Res 2016;22(12):2919-2928). This study sequenced genes of 215 patients with de novo DLBCL and confirmed heterogeneity among activated B-cell-like (ABC), germinal center B-cell-like, and primary mediastinal B-cell lymphoma (PMBL). These lymphomas were frequently affected by mutations in NF-kB, epigenetic, and JAK-STAT pathways, respectively.
Novel pathways were identified in PMBL. Also, TNFAIP3 and GNA13 mutations detected in patients with ABC lymphoma treated with cyclophosphamide, hydroxydaunorubicin (doxorubicin), vincristine, and prednisone plus rituximab were associated with prognoses that were significantly less promising than the prognoses of these patients without the mutations (Clin Cancer Res 2016;22(12):2919-2928).
Another evolving role for NGS is detecting circulating tumor DNA (ctDNA) in the peripheral immunoglobulin receptor loci. Malignant B cells have a unique DNA sequence that encodes its rearranged immunoglobulin variable, diversity, and joining, or VDJ, genes. This unique sequence can be detected with a ctDNA test and serve as a quantitative biomarker in these lymphomas (Oncology (Williston Park) 2016;30(8):731-738, 744). Detection can aid in molecular monitoring before, during, and after therapy. In addition, ctDNA testing may also help identify treatment-resistant clones. Similar to the way positron-emission tomography scan findings can serve as predictors for treatment response, ctDNA may be able to serve as an adjunct biomarker as well.
Aside from being used in common lymphomas, NGS may also be utilized in rare lymphomas. The World Health Organization has defined approximately 100 subtypes of lymphomas (Blood 2016;127(20):2375-2390). The majority of lymphomas diagnosed in the U.S. are DLBCL, follicular B-cell lymphoma, and peripheral T-cell lymphoma–not otherwise specified. Almost half of the lymphomas fall into the nearly 100 lymphoma subtypes (Curr Opin Hematol 2018;25(4):307-314).
NGS has helped identify common mutations in these rarer diseases as well as offered potential targets for treatment. Identification of the BRAF V600E mutation in hairy cell leukemia through exome sequencing has led to treatment with BRAF inhibitors (Blood 2016;127(23):2847-2855). Mutations activating various areas of the B-cell or T-cell receptor signaling pathways occur frequently in rare lymphoma subtypes. Recurrent mutations in PLCG1 have been found in cutaneous T-cell lymphoma, adult T-cell lymphoma, and Sézary syndrome. CD28 mutations have also been seen in these lymphomas and are involved with downstream NF-kB signaling (Blood 2016;128(11):1490-1502, Leukemia 2016;30(5):1062-1070).
Other pathways relevant in lymphoma progression include the mitogen-activated protein kinase pathway (known as MAPK), Janus kinase–signal transducer and activator of transcription (JAK-STAT), and Notch signaling (Curr Opin Hematol 2018;25(4):307-314). These mutations and pathways may help serve as biomarkers to assess response, prognosis, and targets for therapy.
The success of imatinib in revolutionizing chronic myeloid leukemia treatment almost 20 years ago generated enthusiasm and motivation to discover and develop other targeted therapies and the hope of similar outcomes. Unfortunately, similar successes have been difficult to achieve.
Most oncogenic processes are driven by a series of mutations as opposed to a single molecular mutation (Nature 2013;502(7471):333-339). Data on the efficacy of using genomics-based therapy in hematologic malignancies remains limited; however, investigators recently published a meta-analysis comparing biomarker-based treatment strategies with other approaches. In trials of hematologic malignancies, they identified a higher relative response rate of 24.5 versus 13.5 percent (p<0.001) and a higher progression-free survival of 13.6 months versus 4 months in patients undergoing biomarker-based treatment, although the latter difference was not statistically significant (JAMA Oncol 2016;2(11):1452-1459).
One challenge to obtaining more robust data on hematologic malignancies may be lower clinical trial enrollment in biomarker-driven clinical trials. One report described the initial clinical trial accrual experience of three affiliated cancer programs, which did not bear NCI designation, after NGS and demonstrated only four of 200 (2%) patients had hematologic malignancies (J Oncol Pract 2016:12(4):e396-e404). Low accrual may be due to the lower incidence of hematologic malignancies in general, as well as the number of other clinical trial options available to those patients in particular.
Despite the paucity of data regarding treatment response based on genomics-based therapy, the application of genomics to hematologic malignancies, nevertheless, can help stratify diseases better and potentially identify specific patient populations more likely to benefit from genomics-based therapy.
As we start gathering more information and understanding the complexities in the genomic makeup of various hematologic malignancies, we will be able to design clinical trials better and gain more insight regarding mutations and biomarkers and their roles in the diagnosis, prognosis, treatment, and potential relapse of these diseases. Basket trials, which are based on molecular alterations or biomarkers rather than tumor histology, are currently enrolling patients with specific mutations, regardless of diagnosis; however, the lymphoma and leukemia patient population is small, and clinical trial accrual is low.
Nonetheless, in the future, through multi-institutional trials that include community hospitals, data sharing, and other collaborations—disease-specific advocacy groups, artificial intelligence platforms, and tumor registries—investigators are expected to be able to collect sufficient data on these cancers to make data interpretation valid, overcoming the low accrual that has undercut the feasibility of single-site or small multi-site clinical trials.