Background: Liquid biopsy (LB)-based next-generation sequencing (NGS) of circulating tumor DNA (ctDNA) fragments can facilitate the molecular profiling of hematopoietic neoplasms (HNs), particularly in patients for whom conventional NGS is infeasible due to insufficient tumor in the blood, bone marrow, or tissue, or when tissue biopsy cannot be performed. Because HNs shed ctDNA into the plasma at levels comparable to those shed by solid tumors, LB assays initially designed for solid tumors can be adapted for use in hematology. We evaluated a LB-based ctDNA approach in patients with HNs. Methods: Retrospective study of LB samples from patients reported to have HNs tested using hybrid-capture NGS using the FoundationOne Liquid CDx (targeting 324 genes), FoundationOne Liquid (70 genes), or FoundationACT (62 genes) assays between 7/2016 and 3/2022. Paired buffy coat, bone marrow aspirate, or tissue samples undergoing conventional NGS with the FoundationOne Heme (406 genes), FoundationOne CDx (324 genes), or FoundationOne (322 genes) assays were sought where available, and genomic alterations baited across both the LB- and tissue-based assays were assessed for concordance. Results: Among 67,500 unique patient samples submitted for LB-based NGS, 271 were from HNs, including 89 non-Hodgkin lymphoma (NHL), 43 plasma-cell neoplasm (PCN), 41 histiocytosis, 27 myelodysplastic syndrome (MDS), 25 diffuse large B-cell lymphoma (DLBCL), 22 myeloproliferative neoplasm (MPN), 14 Hodgkin lymphoma (HL), and 10 acute myeloid leukemia (AML) cases. The median (interquartile range [IQR]) patient age was 68.0 (55.5-78.0) years, and 121 (44.6%) patients were female. Samples were sequenced to a median exon coverage of 8,564× and pathogenic alterations were detected in 73.4% (199/271) of cases. The median maximum somatic allele frequency (MSAF) among all cases was 7.3% (IQR, 1.1-42.7%), with higher median MSAFs in MPN (45.8%), AML (30.9%), and MDS (19.7%) than in NHL (8.4%), DLBCL (7.1%), histiocytosis (3.1%), HL (2.1%), and PCN (2.0%) cases (P = 0.02, Kruskal-Wallis test), not unexpected given that myeloid neoplasms naturally circulate (Figure 1). LB detected characteristic genomic alterations across HNs, including in TP53, KRAS, MYD88, and BTK in NHLs; TP53, KRAS, NRAS, and BRAF in PCNs; IGH in DLBCL; TP53, ATM, and PDCD1LG2 in HL; BRAF and MAP2K1 in histiocytoses; TP53, SF3B1, DNMT3A, TET2, and ASXL1 in MDS; JAK2 in MPNs; and FLT3, IDH2, and NPM1 in AML, among others (Figure 2). Paired buffy coat, bone marrow aspirate, or tissue specimens collected a median of 118 (IQR 36-346) days apart from the LB specimens sequenced to a median exon coverage of 810× were available for 42 patients. Of 95 variants in 20 HN-associated genes equally baited across the pairs, 31 (32.6%) were detected in both specimens. Of 42 variants detected on buffy coat, bone marrow aspirate, or tissue sequencing, 31 were also detected by LB (positive percent agreement = 73.8%). Eight pairs harbored variants detected in tissue only, all of which were explainable by either low ctDNA shed (e.g., a BRAF fusion was not detected in a LB sample with MSAF = 0.1%) or heterogeneity related to the disease course (e.g., a FLT3 tyrosine kinase domain variant identified by conventional tissue NGS was not detected on follow-up LB two years later, post-treatment). Conversely, 22 samples had ≥1 variant detected in the LB sample only, including in NF1 (n=4), KRAS (n=3), and BRAF (n=2), likely reflecting the enhanced sensitivity of LB for detecting low-level disease clones. Conclusions: LB detected relevant genomic alterations in the majority of patients with a variety of HNs. Given its analytical sensitivity, LB may also detect low-level residual or emerging therapy-resistant clones unable to be detected in a single tissue specimen and may play a role in the evaluation of the patient's disease course and in associated treatment decisions. This pilot study suggests the potential value of LB in the clinical management of patients; however, further prospective evidence would be required to quantify the utility of LB for patients with HNs. Figure 1View largeDownload PPTFigure 1View largeDownload PPT Close modal
Read full abstract