A targeted circulating tumor DNA landscape of copy number aberrations in large B-cell lymphomas

Leukemia ARTICLE www.nature.com/leu OPEN A targeted circulating tumor DNA landscape of copy number aberrations in large B-cell lymphomas Maare Arffman 1,2, Leo Meriranta1,2, Judit Jørgensen3, Marja-Liisa Karjalainen-Lindsberg4, Klaus Beiske5,6, Mette Pedersen7,8, ✉ Kristina Drott9, Øystein Fluge10, Sirkku Jyrkkiö11, Peter Brown 12, Harald Holte 13,14 and Sirpa Leppä 1,2 © The Author(s) 2026 1234567890();,: The utility of circulating tumor DNA (ctDNA) for mutational genotyping, pretreatment prognostication, and assessment of molecular response is well established in patients with aggressive large B-cell lymphoma (LBCL). Here, we have applied targeted panel and duplex sequencing of plasma ctDNA to study copy number aberrations (CNAs) along with mutational landscapes in 123 uniformly treated patients with high-risk LBCL. We find a robust correlation between targeted and whole-genome sequenced CNA landscapes (R = 0.81) and identify CNAs in the ctDNA in 76% of the patients above the limit of detection. We describe the most frequently affected genomic regions, their interactions with diagnostic and genetic subtypes, and associations with overall and progressionfree survival. Specifically, we show how ctDNA profiling of TP53 loss outperforms fluorescence in situ hybridization (FISH)-based TP53/17p analysis in risk assessment, independent of clinical risk factors and ctDNA concentration. We validate key findings of prognostic tumor fraction and TP53 loss in an independent LBCL cohort. Furthermore, we detect dynamic shifts between the fractions of lymphoma clones by assessing CNAs and mutations in the ctDNA at disease progression. These findings demonstrate the potential of minimally invasive, targeted CNA analysis in resolving the molecular heterogeneity of LBCLs. Leukemia; https://doi.org/10.1038/s41375-026-02955-w INTRODUCTION Large B-cell lymphomas (LBCLs) comprise a biologically and clinically heterogeneous group of the most common aggressive lymphoid cancers [1, 2]. Although patient outcomes are mainly favorable following anthracycline-based immunochemotherapy, 30% of patients will relapse after an initial response, with approximately 10% having primary refractory disease and being at high risk of dying from lymphoma [3]. On the other hand, many patients are overtreated and suffer from treatment-related toxicities [4]. Despite molecular characterization of LBCLs through the profiling of genetic landscapes [5–8], tumor microenvironment [9, 10], and host response [11], clinical and biological heterogeneity remains an obstacle to developing personalized treatments and improving survival. The genomic landscape of LBCL is complex and comprises a diverse set of genetic alterations, including point mutations, translocations, and copy number aberrations (CNAs), which define genetic subtypes and impact patient outcomes [6–8, 12]. Among other aberrations, CNAs have an important role in the pathogenesis of LBCLs [13–15]. Specifically, the loss of TP53 (17p13) is associated with increased resistance to chemotherapy and a poor prognosis [7, 8, 16]. In our trial, we also observed that, unlike BCL2 and MYC translocations, the adverse impact of TP53 deletion on survival could not be mitigated by dose-intensified treatment [17]. Other recurrent and potentially actionable aberrations, including loss of CDKN2A (9p21) or gain of CD274 (9p24.1), have been associated with avoidance of immune response and more aggressive behavior in LBCLs [18, 19]. Recently, the detection of plasma circulating tumor DNA (ctDNA) has revolutionized LBCL profiling [20–22], and is expected to reduce the need for tissue biopsies. Targeted sequencing panels are often preferred in ctDNA profiling [20, 21, 23] to identify genetic abnormalities, primarily single nucleotide variants (SNVs) and small insertions and deletions. For instance, quantification of ctDNA burden at baseline [24–26] and the detection of minimal residual disease (MRD) [27–29], refined by high sequencing coverage together with duplex sequencing [17] or the identification of phased variants [23], have enabled dynamic risk assessment in LBCLs. However, profiling of CNAs from the ctDNA has remained less studied. A study employing shallow wholegenome sequencing showed that diffuse LBCL (DLBCL) not otherwise specified (NOS) can be distinguished from Hodgkin’s lymphoma using CNA profiles [30]. Another study using a targeted ctDNA panel showed that DLBCL NOS patients could be classified according to SNVs and CNAs into predefined genetic clusters, some of which, together with high ctDNA burden, were associated 1 Research Programs Unit, Applied Tumor Genomics, University of Helsinki, Helsinki, Finland. 2Department of Oncology, Helsinki University Hospital Comprehensive Cancer Centre, Helsinki, Finland. 3Department of Hematology, Aarhus University Hospital, Aarhus, Denmark. 4Department of Pathology, Helsinki University Hospital, Helsinki, Finland. 5 Department of Pathology, Oslo University Hospital, Oslo, Norway. 6Institute of Clinical Medicine, Medical Faculty, University of Oslo, Oslo, Norway. 7Department of Pathology, Zealand University Hospital, Roskilde, Denmark. 8Department of Clinical Medicine, Faculty of Health and Medical Sciences, University of Copenhagen, Copenhagen, Denmark. 9 Department of Oncology, Skane University Hospital, Lund, Sweden. 10Department of Oncology, Haukeland University Hospital, Bergen, Norway. 11Department of Oncology, Turku University Hospital, Turku, Finland. 12Department of Hematology, Rigshospitalet, Copenhagen, Denmark. 13Department of Oncology, Oslo University Hospital, Oslo, Norway. 14KG Jebsen Centre for B-cell malignancies, Oslo, Norway. ✉email: sirpa.leppa@helsinki.fi Received: 13 November 2025 Revised: 3 March 2026 Accepted: 27 March 2026 M. Arffman et al. 2 with worse survival [31]. In addition to the first-line therapy setting, the enrichment of specific CNAs has been investigated in CAR-T cell-resistant relapsed/refractory (R/R) LBCL patients [32]. While significant advances have been made in the field of lymphoma liquid biopsy (LB), the translational utility of ctDNA CNAs has not been addressed comprehensively in a uniformly treated cohort of LBCL patients. The aim of this study was to identify clinically relevant CNAs in LBCL patients using a targeted panel designed for variant profiling and MRD testing, enabling multi-layer cfDNA profiling from the same LB source. To achieve this, we profiled plasma CNAs from 123 patients with primary LBCL treated in a Nordic phase II trial. We demonstrate the use of targeted ctDNA profiling for detecting CNAs, enhance risk assessment of LBCL patients, and validate key findings in another LBCL cohort. We describe subtype-specific molecular differences and report examples of the molecular background of expanding clones in patients with R/R DLBCL. Taken together, we provide clinically important CNAs that could be employed in the personalized treatment of LBCL patients. MATERIALS AND METHODS Pati (...truncated)