Multifaceted cfDNA profiling in mantle cell lymphoma Open Access

In this issue of Blood Advances, Meriranta et al¹ introduce a multimodal cell-free DNA (cfDNA) based approach to characterize disease biology, identify determinants of therapeutic outcomes and detect minimal residual disease (MRD) in patients with relapsed/refractory mantle cell lymphoma (R/R MCL). To date, the predominant MRD assays utilized in the clinic and in clinical trials for MCL have focused on circulating tumor cells (CTCs). These are immunoglobulin high-throughput sequencing (IgHTS) and real-time quantitative polymerase chain reaction (RT-qPCR), both of which follow patient-specific immunoglobulin rearrangements.^2,3 However, these methods have limited sensitivity and a narrow scope, in that they are only capable quantifying disease burden/MRD.

In this study, the authors used hybrid capture–based duplex sequencing, a method that minimizes error rates by sequencing both DNA strands and reporting only variants detected on both, which enables accurate genotyping of variants at very low variant allele frequencies (VAFs).⁴ This methodology was applied to profile cfDNA, a subset of which includes circulating tumor DNA (ctDNA), as well as tissue-derived DNA from 58 patients enrolled in the MRD driven phase 1b/2 VALERIA clinical trial evaluating venetoclax, lenalidomide, and rituximab.⁵ This approach allowed the authors to simultaneously genotype tumor mutations, quantify tumor burden, detect MRD, and assess clonal hematopoiesis (CH) mutations.

To quantify disease burden and detect MRD, tumor-specific single-nucleotide variants (SNVs) were first identified using a pretreatment tumor sample. These SNVs were then used to quantify ctDNA levels and detect MRD in serial plasma samples. Multiple studies have demonstrated that both pretreatment and dynamic ctDNA levels are predictive of patient outcomes in B-cell non-Hodgkin lymphoma, with detectable MRD and higher pretreatment ctDNA levels consistently associated with inferior prognosis.^6,7 However, the majority of this prior work has focused on other B-cell non-Hodgkin lymphoma subytes, and this study represents one of the most comprehensive to date focused on MCL. In this study, higher pretreatment ctDNA levels were associated with poor prognosis and were correlated with established clinical risk factors such as LDH and the Mantle Cell Lymphoma International Prognostic Index. Among patients that had an available plasma sample following 6 cycles of therapy, 4 patients had detectable ctDNA (MRD-positive), whereas 25 had no detectable ctDNA (MRD-negative). MRD status was strongly predictive of outcomes, and all 4 patients who were MRD-positive relapsed, as compared with only 4 of the 25 patients who were MRD-negative.

Although this approach offers several advantages over CTC based methods, its limit of detection (LOD), at least 1:10 000 across samples, is comparable to IgHTS and RT-qPCR. As noted by the authors, 4 patients who were MRD-negative by ctDNA ultimately relapsed, indicating a need for increased assay sensitivity. Two critical technical determinants of the LOD are the number of monitored reporter mutations, and the inherent error rate of the detection method. The approach commonly used in large B-cell lymphoma, wherein an off-the-shelf hybrid capture panel targeting stereotyped mutations is used across patients, may not translate effectively to MCL due to the lower mutation burden in this subtype. Consequently, achieving sufficient assay sensitivity in MCL may require a personalized approach, targeting patient-specific mutations. Additionally, next-generation sequencing methods that utilize SNV-based tracking with reduced technical error rates, such as approaches monitoring phased variants, have emerged as promising tools.⁸ It could also be of interest to determine if a combination of ctDNA and CTC-based MRD methods might augment each other and improve MRD detection.

An important advantage of ctDNA is that it can be used to genotype genomic alterations, including SNVs, small insertions and deletions, and copy number alterations.^9,10 In this study, the authors found that SMARCA4 mutations were associated with superior outcomes, while TP53 mutations, which have an established association with poor prognosis, were indeed associated with poor outcomes. When comparing VAFs of individual mutations between matched tissue biopsy, ctDNA and peripheral blood (PB) samples, there were comparative differences in VAFs, suggesting that ctDNA is capable of profiling tumor spatial heterogeneity. Additionally, the authors detected subclonal BTK mutations in ctDNA from 2 patients previously treated with ibrutinib, highlighting the potential of this tool to profile clonal evolution and identify therapeutic resistance mutations. Notably, PB and tissue samples were prioritized for mutational genotyping in this study, which was likely necessary given that MCL often presents with leukemic involvement. Many existing pipelines designed to perform de novo genotyping from ctDNA use PB mononuclear cells to filter out germ line variants; however, this approach would be problematic in MCL due to contamination from CTCs in a significant fraction of cases. This challenge could be overcome by developing improved computational algorithms for germ line variant identification, or by using alternative sources of germ line DNA, to make ctDNA a reliable genotyping source in lymphoma subtypes with leukemic involvement, such as MCL.

Several studies have demonstrated that multimodal approaches to cfDNA/ctDNA profiling can look beyond mutational genotyping and MRD assessments in order to characterize additional features.¹¹ In this study, the authors also assessed clonal hematopoiesis mutations from cfDNA, and noted the positive selection of TP53 and PPM1D mutated CH clones was associated with inferior outcomes and hematologic toxicity. Further, the authors present a method to differentiate tumor specific mutations from CH mutations, by leveraging the fragment length difference between ctDNA molecules (shorter) and cfDNA molecules derived from circulating blood cells (longer). While compelling, further validation of these methods is warranted given the frequent presence of circulating lymphoma cells in MCL, which could complicate the ability to clearly delineate CH from tumor-specific mutations.

In summary, Meriranta et al present a framework for a multimodal cfDNA platform capable of performing mutational genotyping, quantifying ctDNA, defining MRD status and assessing CH in R/R MCL. The study is an important advance as few studies utilizing cfDNA based analyses in the context of MCL have been described. However, validation in larger prospective cohorts utilizing more sensitive profiling methods is needed before these methods can be translated clinically for patients with MCL.

Conflict-of-interest disclosure: A.F.H. reports research funding from Bristol Myers Squibb, Genentech, Merck, AstraZeneca, and Pfizer; and consultancy with Bristol Myers Squibb, Genentech, Merck, Takeda, Genmab, Pfizer, AbbVie, Allogene Therapeutics. B.J.S. reports consultancy from Foresight Diagnostics; stock ownership in CARGO Therapeutics and Allogene Therapeutics; and honoraria/advisory board involvement in ADC Therapeutics.