Liquid biopsy (LB)-based comprehensive genomic profiling (CGP) of circulating tumor DNA (ctDNA) for the evaluation of patients with myeloid neoplasms.
Douglas A Mata,Mina L. Xu,Vignesh Shanmugam,Hanna Tukachinsky,Alexa Betzig Schrock,Jeffrey S. Ross,Erik Abraham Williams,Meagan Montesion,Brennan J. Decker,Jo-Anne Vergilio,Geoffrey R. Oxnard,Jamal K. Benhamida
DOI: https://doi.org/10.1200/jco.2022.40.16_suppl.e19064
IF: 45.3
2022-06-01
Journal of Clinical Oncology
Abstract:e19064 Background: LB-based CGP of ctDNA can facilitate the diagnosis, molecular profiling, and monitoring of patients with myeloid neoplasia, particularly when standard CGP is not feasible due to a lack of circulating disease or insufficient tumor in the marrow or tissue. Methods: Retrospective study of LB samples from patients reported to have histiocytic neoplasms (HNs), myelodysplastic syndrome (MDS), myeloproliferative neoplasms (MPN), or acute myeloid leukemia (AML) tested using hybrid-capture next-generation sequencing using the FoundationOne Liquid CDx (targeting 324 genes), FoundationOne Liquid (70 genes), or FoundationACT (62 genes) assays during routine clinical care between 5/2016 and 9/2021. Results: Among 48,459 unique patient samples submitted for CGP, 83 were from myeloid neoplasia: 34 HN, 23 MDS, 15 MPN, and 11 AML. The median cell-free DNA yield was 74.9 ng (interquartile range [IQR]: 52.0-167.4), well above the minimum needed for analysis. Reportable pathogenic alterations were detected in 62.7% (52/83) of cases. The median maximum somatic allele frequency (MSAF) was 17.0% (IQR: 1.7-49.0%), with AML (45.7%), MDS (25.9%), and MPN (43.8%) having higher median MSAFs than HNs (3.2%), not unexpected given that the former naturally circulate. In all, 147 pathogenic short variants and 6 pathogenic rearrangements were detected (median variant allele frequency [VAF]: 2.1%, IQR: 0.3-20.5%, range: 0.11-99.8%). Sixty-one had VAFs < 1.0%, indicating excellent sensitivity. HNs exhibited activating RAS-RAF-MEK pathway alterations in BRAF (31.3% of cases in which pathogenic alterations were identified), NF1 (12.5%), MAP2K1 (6.3%), and NRAS (6.3%), detected at VAFs as low as 0.21%, 0.13%, 0.11%, and 4.0%, respectively. AML cases exhibited pathogenic alterations in TP53 (57.1%), FLT3 (28.6%), IDH2 (28.6%), NPM1 (28.6%), KRAS (14.3%), and NRAS (14.3%), among others. Subclonal heterogeneity was identified, including an AML case with KRAS G12A and G13R and NRAS G12A mutations detected at VAFs of 0.24%, 0.19%, and 0.2%, respectively. Altered genes in patients undergoing workup for MDS or MPN included TP53 (41.4%), JAK2 (37.9%), DNMT3A (13.8%), CHEK2 (10.3%), PTPN11 (10.3%), SF3B1 (10.3%), TET2 (10.3%), ASXL1 (6.9%), MPL (3.5%), and U2AF1 (3.5%), among others. Last, among 48,243 patients with a LB for a solid tumor diagnosis, 3,487 JAK2 V617F, 366 CALR C-terminal truncation, and 343 MPL W515X mutations were detected, potentially indicating a concurrent or occult MPN. Conclusions: LB identified clinically relevant genomic alterations in myeloid neoplasia, offering a powerful tool that may be used pre- or post-treatment when CGP of the buffy coat, marrow, or tissue is infeasible. Given its low limit of detection, LB can also identify low-level emerging or persistent subclones, which may facilitate monitoring and minimal residual disease testing.
oncology