Molecular Surveillance of Aggressive Large B-cell Lymphoma using Circulating Tumor DNA Duplexes

Molecular Surveillance of Aggressive Large B-cell Lymphoma using Circulating Tumor DNA Duplexes | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Brief Communication Molecular Surveillance of Aggressive Large B-cell Lymphoma using Circulating Tumor DNA Duplexes Sirpa Leppä, Leo Meriranta, Maare Arffman, Matias Autio, Judit Jørgensen, and 8 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8007791/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Treatment failure affects one-third of patients with aggressive large B-cell lymphoma (LBCL), and current tools are inadequate in determining remission. We investigated the added value of longitudinal ctDNA monitoring in plasma by duplex sequencing. Clearance of ctDNA after 2 or 4 cycles of immunochemotherapy identified patients with excellent survival, whereas elevated interim ctDNA levels among those with measurable residual disease (MRD) predicted refractory disease. Two-year progression-free survival rates based on end-of-therapy MRD varied greatly (94.3% vs 18.2%, negative and positive, respectively), and ctDNA analysis resolved most false radiological response assessments and revealed poor targeting of post-treatment interventions to MRD-positive cases. In surveillance, ctDNA was detectable at least 3 months before clinical relapse, while patients with isolated central nervous system relapses had the shortest plasma ctDNA lead times. Overall, ctDNA monitoring with ultra-sensitive duplex sequencing offers a dynamic molecular tool that can guide clinical decision-making in LBCL. Health sciences/Medical research/Outcomes research Biological sciences/Cancer/Haematological cancer/Lymphoma/Non-hodgkin lymphoma/B-cell lymphoma Health sciences/Biomarkers/Prognostic markers Circulating tumor DNA duplex sequencing measurable residual disease B-cell lymphoma surveillance relapse longitudinal Figures Figure 1 Figure 2 Figure 3 Main Disease recurrence affects 30% of patients with aggressive large B-cell lymphoma (LBCL) and predicts poor outcomes 1 . Treating LBCL relapse is challenging 2 , and high tumor volumes predict poor survival 3 . Early detection of recurrence could improve outcomes; therefore, treatment responses are monitored, and patients undergo imaging with computed tomography (CT) or positron-emission tomography (PET), although their utility in surveillance is limited 4,5 . Repeated imaging exposes patients to radiation, the sensitivity of these methods is modest, and the poor specificity leads to invasive and unnecessary biopsies, which potentially harm patients 6,7 . Consequently, more accurate methods are needed for monitoring response and recurrence. Liquid biopsies using plasma circulating tumor DNA (ctDNA) complement diagnostics, predict treatment response, and inform survival in LBCL 8 . These tools are expected to transform decision-making 9-12 . The presence of ctDNA in plasma at diagnosis and recurrence is well-established; however, detecting it before clinical relapse requires sensitive methods 8 . Early studies utilized immunoglobulin high-throughput sequencing (Ig-HTS) to detect ctDNA, showing promising results in response evaluation and surveillance 9,10 . However, Ig-HTS lacks sensitivity, tracking only single or a few genomic events, resulting in a disappointing 56% detection rate before or at relapse in a prospective study 13 . Hybrid-capture panel techniques detect ctDNA more sensitively by tracing more mutations, but sequencing error rates limit their performance 11,14 . Phased variants are unlikely to be detected by error, and can be harnessed for sensitive disease monitoring 14,15 . Still, phased events are rare, often found in hypermutable loci affected by clonal heterogeneity, and disease detection mainly depends on the number of informative molecules 16 . Duplex sequencing, which recovers both strands of the original DNA for error correction 17 , achieves an extremely low error rate and can detect mutations down to parts per million (ppm) 18 . We reasoned that longitudinal duplex sequencing of lymphoma ctDNA could demonstrate the potential of detecting measurable residual disease (MRD) over time in LBCL. We aimed to determine the added clinical value of ctDNA detection by duplex sequencing in serial plasma samples from LBCL patients. We applied a previously established targeted duplex sequencing-based ctDNA assay 19 , 20 on 447 longitudinal plasma cfDNA samples from 123 patients with primary LBCL at various timepoints according to a prospective protocol (Figure 1A) 20 . The patients were young (aged 18-65 years), had high-risk LBCL, and were treated with a curative intent with early high-dose-methotrexate, dose-intensive immunochemotherapy (DA-EPOCH-R or R-CHOEP), and high-dose-cytarabine in a Nordic phase II trial 20 . The lowest ctDNA fractions detected were in the ppm range, and the median limit-of-detection was 1.69*10 -5 with 26% of the samples having a limit-of-detection in the ppm range. Statistical analyses were performed in the R environment. Additional details are described in the Supplementary methods and Tables S1-S4. During an updated median follow-up of 60.4 months, 22 patients experienced progression, including 17 relapses and 13 deaths. Pretreatment ctDNA levels declined in all evaluable patients after two cycles of immunochemotherapy, with 67% (75/112) reaching molecular remission (MRD CYC2–, Figure 1B, S1A). MRD CYC2 was prognostic; yet, 68% (25/37) of patients with MRD CYC2+ did not relapse (Figure 1C). However, the negative predictive value (NPV) of MRD CYC2 evaluation for relapse was high (95%, Table S2-S3). Notably, among the 37 MRD CYC2+ cases, elevated ctDNA levels after cycle 2 were predictive of refractory disease and worse survival (Figure 1D). Major molecular response 21 (MMR, ≥2.5 log reduction in ctDNA) was achieved by 85% of patients (94/111), and failure to reach MMR identified patients with poor survival; however, its sensitivity was lower than that of MRD CYC2 (Figure 1E, Table S2). Overall, clearance of ctDNA after 2 cycles predicts excellent survival, and in cases where ctDNA is present, high levels of ctDNA indicate worse survival. Twenty-seven patients with MRD CYC2+ were evaluable for ctDNA after 4 cycles (MRD CYC4 ). Of these, 14 (52%) had become MRD-negative, while 13 (48%) remained MRD-positive, which was associated with poor survival (Figure 1F, S1B). In patients with MRD CYC4+ , high ctDNA levels predicted poor survival similarly to MRD CYC2+ ,and primary refractory cases were identified by ctDNA levels above 10 hGE/ml (Figure 1G-H). After the last cycles of therapy (MRD EOT ), 7 of the 13 patients (54%) with MRD CYC4+ became MRD-negative, while 6 patients (46%) remained MRD-positive. These findings reveal different response kinetics that update prognostic information on therap 22 , and interim ctDNA quantification identifies patients unlikely to respond. Among all evaluable patients, MRD EOT identified a dramatic difference in outcomes, with all patients who died within two years of follow-up being MRD EOT+ (Figure 1I, S1C, Table S2-3). MRD EOT+ detected all 4 patients with primary refractory disease, and 42% (5/12) of those progressed later during follow-up. The median time to progression for MRD EOT+ patients was 8.4 months, compared to 31.2 months for MRD EOT– (P=0.0008, Mann-Whitney-U), and the NPVs of MRD EOT– for 1-, 2-, and 4-year PFS were 98%, 94%, and 91%, respectively. MRD EOT was associated with Deauville scores (DS), showing a 5.6% (4/71) MRD EOT+ rate for DS 1-3 and 46% (6/13) for DS 4-5 (Fisher’s P=0.0051, Figure 1J, S1D). MRD EOT was positive in 50% (3/6) and 71% (5/7) of patients with relapse in the DS 1-3 and DS 4-5 groups, respectively, with 1 false positive in each category, demonstrating a significant prognostic impact (Figure 1J-K, S1E-F). These findings highlight the excellent prognostic value of molecular response measurement in LBCL and are highly consistent with recent results reported using ctDNA assays with a similar error rate as duplex sequencing 16,23 . Seven patients underwent tissue biopsies from PET-positive lesions after therapy, with 2 confirmed lymphoma progressions (29%). MRD EOT+ was detected in 3 of these 7 patients, including both confirmed progressions. Twenty percent (20/99) of the evaluable patients for MRD EOT received radiotherapy (RT), which was not associated with MRD EOT status (Fisher’s exact, P=0.69, Figure S1G). Six patients received RT due to tumor bulk at diagnosis, and all were MRD EOT– .Three of the 11 MRD EOT+ patients received RT for PET-avid lesions, and in two of those with available surveillance samples, MRD remained positive after RT, with the patients relapsing during follow-up. These results highlight the limitations of current clinical tools in providing appropriate interventions for patients at higher risk of progression after immunochemotherapy, a gap that could be addressed with ctDNA analysis. On post-treatment surveillance, all evaluable patients with MRD EOT+ remained MRD-positive in subsequent plasma samples, and 86% (6/7) of the relapsing patients with MRD EOT– became MRD-positive before clinical relapse (Figure 2A-B). Only one patient with late relapse was MRD-negative for over a year of follow-up until turning MRD-positive. Once ctDNA was detected, 93% (14/15) of subsequent tests remained positive, whereas the specificity remained 100% in patients without relapse, including 1 patient with non-lymphoma-related death (22 tests, Figure 2B). Surveillance provided lead time before relapse in 92% (11/12) of patients (median=6 months). A lead time of 3 months was achieved in 83% of cases (10/12, Figure 2C). Three patients had a lead time of over 12 months, including one with ctDNA detected throughout measurements (Figure 3A). The patient developed a symptomatic parailiacal tumor 3 months after metabolic CR and responded to corticosteroids; however, tissue biopsies showed an unspecific histiocytic reaction twice before a conclusive diagnosis was made using genomic techniques, as tumor cells had lost B-cell phenotypic markers (Figure 3A). Finally, patients with short ctDNA lead times to recurrence (0 and 1.6 months) were the only cases with isolated CNS recurrence (n=2, Figure 2C). Since mutations in primary CNS lymphomas can be more easily detected in cerebrospinal fluid (CSF) 24,25 , we analyzed cfDNA from CSF samples available for one of the patients, who had CSF flow cytometry positivity at inclusion. Exploratory analysis revealed lymphoma reporter mutations in the pretreatment CSF, all of which became undetectable after two cycles of therapy, reflecting ctDNA response in plasma (Figure 3B). Despite flow cytometry clearance after therapy, however, lymphoma reporters were again detectable in the CSF at EOT, indicating CNS-restricted resistance with an additional three months’ lead time compared to plasma-based detection (Figure 3B). These findings demonstrate significant lead times to systemic LBCL relapses with ctDNA surveillance, while CSF analysis could help in early detection of CNS recurrence. Here, we showed that longitudinal plasma ctDNA detection accurately predicts survival, determines therapy responses, and clarifies various clinical scenarios upon LBCL recurrence. Treatment could be personalized earlier based on on-therapy ctDNA measurements, and clinical interventions could be more precisely directed at treatment-refractory patients using ctDNA analysis. The similar prognostic significance observed in patients with either radiological remission or residual positivity, along with the high NPV of ctDNA clearance, supports incorporating molecular remission into LBCL response criteria to complement imaging 23 . Late recurrences could be detected earlier through repeated post-therapy surveillance and intercepted before a high tumor burden develops. Surveillance could be tailored according to dynamic risk profiling, and CSF analysis could supplement surveillance in patients with suspected CNS recurrence. Prospective studies are warranted to confirm the feasibility and potential benefits of ctDNA-guided therapy. The comprehensive clinical data and sampling, long median follow-up, and ctDNA detection down to ppm with duplex sequencing are strengths of our study. Further enhancements in sequencing design, analysis of larger plasma volumes, and inclusion of phased variants could further improve the practical sensitivity of ctDNA detection. Online Methods Clinical data 123 patients treated in Nordic phase II protocol (NCT03293173) were included in the study with 119 patients with samples evaluable for circulating tumor DNA (ctDNA) analysis. The clinical trial results, study treatment and patient characteristics have been previously reported 20 . Briefly, young patients (≥65 years of age) with de novo aggressive large B-cell lymphoma (LBCL) were recruited and treated with a curative intent. The treatment consisted of early central nervous system (CNS) prophylaxis and dose-dense or dose-adjusted chemoimmunotherapy (R-CHOEP or DA-EPOCH-R) according to tumor biomarker profile. The follow-up data of the patients was updated for the current study. Lead times to recurrence were calculated as the difference between the date of first ctDNA positive plasma sample draw and the date of recurrence. Samples, sample processing and sequencing The reagents and resources used in the study are listed in Table S1. Peripheral blood drawn in K2-EDTA tubes at inclusion was used as source of matched normal DNA and was processed as previously described 20 . Residual tumor DNA derived from diagnostic fresh frozen biopsies or formalin-fixed paraffin-embedded tissue were analyzed when available. Plasma samples used for cell-free DNA (cfDNA) extraction were collected according to study protocol at pretreatment, on-therapy after second and fourth cycles (interim computer tomography [CT] stagings), and at the end-of-therapy (EOT) at the time of PET-CT end-staging (Figure 1A). Additionally, surveillance samples were collected on follow-up at 3, 6, 9, 12, 18 and 24 months and at relapse if present. Ten milliliters of peripheral blood was drawn to standard K2-EDTA or white blood cell preserving (cell-free DNA BCT, Streck) tubes, and the plasma was immediately separated, aliquoted and stored. In addition, cerebrospinal-fluid (CSF) samples were collected from spinal taps at pretreatment, and if tumor cells were detected by flow cytometry, CSF samples were collected after 2 cycles and at EOT (one patient in the study reported in Figure 3B, 7 CSF samples studied in total). 2-4 ml of cerebrospinal fluid was be collected in a sterile test tube which was centrifuged at 2,000g for 10 minutes at +4 C to remove cell debris and the supernatant was transferred to vials and stored at -70 C. Extraction of cfDNA from CSF was performed similarly to plasma samples. In total, 119 pretreatment and 324 on-therapy or post-therapy surveillance samples from 119 patients were analyzed. Sample selection of pretreatment, after 2 cycles (n = 112) and EOT (n = 99) plasma were based on sample availability. Plasma samples drawn after cycle 4 (n=55) were enriched for patients with follow-up events with all available samples from patients with events and all available samples from patients with measurable residual disease (MRD) after 2 cycles of therapy (MRD CYC2+ ) were analyzed (n=27/55). All available post-therapy surveillance samples from patients with events (progression; recurrence or death from any cause) were analyzed with a control series of patients with no follow-up events based on baseline ctDNA positivity and sample availability (n=54). Additionally, 4 plasma samples drawn upon recurrence were included in the study. ctDNA analysis Previously established platform using targeted panel of ~750 kilobases in size with exhaustive duplex sequencing was used for ctDNA analysis 19 . Centralized cfDNA and DNA extraction, quantification and quality control steps and library preparation, target enrichment, sequencing and raw data processing were performed as previously described 19,20 . Sequencing metrics of cfDNA samples are reported in table S4A and analysis metrics are shown in table S4B. Presence of MRD was estimated using adapted software 26 with concepts described previously 27 . All tests were run and the samples analyzed blinded to the clinical data using predetermined test parameters. Reporter mutations were single nucleotide variants not in predetermined clonal hematopoiesis target genes identified in genotyping diagnostic tumor tissues with ≥10% variant allele frequency (VAF) and/or in pretreatment plasma cfDNA with ≥1% VAF. Patients with below 20 reporters (n=4/119) were considered genotyping failures and regarded as MRD-negative for all the tested samples. Median number of reporters in evaluable patients was 252 (range 24–1316). Only duplex corrected reads with mapping quality of 60 and bases with base quality of 60 were considered in the MRD analysis. Sample specific background error rate was measured as any mismatch from reference and the median error rate was 2.74e-06 (range 3.18e-07–1.09e-05). 128 (40%) plasma cfDNA samples were drawn in Streck tubes and 196 (60%) in tubes with EDTA as anticoagulant, and no difference in background rate or number of informative duplex reads was detected between the starting materials (Mann-Whitney U tests, P > 0.05, data in Table S4B). Results from Monte-Carlo based framework were interpreted with a predetermined cut-off for empirical p-value calibrated to 95% specificity among withheld controls determined previously 20 . The ctDNA levels at interim time-points were measured as haploid genomic equivalents (hGE/ml), and the median levels were 2.19 hGE/ml after 2 cycles and 6.31 hGE/ml after 4 cycles of therapy. The analytical sensitivity of duplex sequencing allows mutation detection below parts per million 17 . In the context of cfDNA with limited input material, the sample specific sensitivity depends on the number of informative molecules, and therefore, the sample-wise limit of detection (LOD) was estimated with a binomial-based detection model using the informative duplex depth assuming 95% sensitivity and 5% false-positive rate (LOD95) similarly as previously described 16 . Informative duplex depth per patient was defined as the sum of altered and reference counts from reporter mutation positions meeting MRD test criteria above. 96% (301/313) of the samples from evaluable patients reached below 10^-4 LOD95, the median LOD95 was 1.69e-5, and 26% (82/313) of the samples reached LOD95 in 10^-6 range (Table S4B). In line with these models, the lowest fractions of ctDNA molecules in the study were detected in the parts per million range (Table S4B). Immunohistochemical images Slides of immunohistochemical stainings from routine diagnostics were performed in Helsinki University Hospital, department of Pathology and were available for the current study. Microscopy images were acquired with Leica DM LB brightfield microscope (Leica Microsystems GmbH) with Olympus DP50 camera. Statistical analysis Statistical analyses were performed in the R environment (R foundation for Statistical Computing, Vienna, Austria, Table S1). Statistical details accompany the results in the text and/or related figure legends. The used tests were non-parametric and two-sided, and P values < 0.05 were considered statistically significant. The Kaplan-Meier method with log rank test was used to estimate survival rates between different patient groups and Cox regression was used to estimate survival in univariate models which are available in Table S3 (R package survival, 3.8-3). Survival was estimated and the results interpreted using the follow-up times without considering time-point of the ctDNA test. Test performance metrics were evaluated and the cross-tabulation and confusion matrixes generated with the R package caret (version 7.0-1). Declarations Acknowledgements We thank all participating study sites for patient enrolment and sample collection, Laura Hakala for trial coordination, the staff at the Clinical Trial Office in Aarhus, Denmark for trial management, Anne Aarnio and Emma Saarinen for meticulous laboratory work, Henrikki Almusa for sequencing data processing, and Sari Hannula for guiding the design and coordination of sequencing. Library preparation and sequencing data processing were performed by the FIMM Genomics NGS Sequencing unit at the University of Helsinki, supported by HiLIFE and Biocenter Finland. CSC IT (Center for Science, Finland) is acknowledged for providing the computational resources for this study. Conflicts of interest H.H*: Genmab: Safety Committee; Incyte: Consultancy; Novartis: Research Funding; Elicera Therapeutics: Safety Committee; Serb SA: Advisory Board. J.J*: Abbvie: consultancy; BMS/Celgene: Consultancy; Gilead: Consultancy; Incyte: Consultancy; Novartis: Consultancy; Caribou Bioscience: Consultancy; Roche: Consultancy; SOBI: consultancy. P.B*: BMS: Consultancy; Gilead: Consultancy; Novartis: Consultancy; Roche: Consultancy. Ø.F.*: Roche: Consultancy; Bayer: Consultancy. S.L*: AbbVie: Honoraria, Consultation; GENMAB: Research Funding**; GILEAD: Honoraria; Incyte, Consultation, Honoraria; Novartis: Research Funding*; Roche: Honoraria, Consultancy, Research Funding*; BMS/Celgene: Consultancy, Research Funding**, Sobi: Honorary. Other authors declare no conflicts of interest. *All outside of the submitted work. **To the institute Ethics approval and trial registration The study was conducted in accordance with the Guidelines on Good Clinical Practice from the International Conference on Harmonization and the principles outlined in the Declaration of Helsinki. The protocol was approved by the medical agencies and ethics committees in Finland, Denmark, Norway, and Sweden, and the trial was registered at ClinicalTrials.gov, number NCT01325194. All patients signed informed consent forms before participating in the study. Author contributions Conceptualization: LM, SL Sample and clinical data collection and processing: MAr, JJ, MLK, KB, MP, KD, ØF, SJ, PB, HH, SL Investigation: LM, MAu, SL Formal analysis: LM Data interpretation: All authors Funding acquisition: SL Writing - Original Draft Preparation: LM Writing - Review & Editing: SL Final approval of manuscript: All authors Funding This study was funded by the Research Council of Finland (SL), Finnish Cancer Organizations, iCAN Flagship (SL), Nordic Cancer Union (SL), Sigrid Juselius Foundation (SL), and Helsinki and Uusimaa Hospital District (SL). LM was supported by personal grants from the Finnish Medical Foundation and Sigrid Juselius Foundation. References Sehn LH, Salles G (2021) Diffuse Large B-Cell Lymphoma. N Engl J Med 384:842–858 Chaganti S et al (2025) Management of relapsed or refractory large B-cell lymphoma: A British Society for Haematology Guideline. Br J Haematol Dean EA et al (2020) High metabolic tumor volume is associated with decreased efficacy of axicabtagene ciloleucel in large B-cell lymphoma. Blood Adv 4:3268–3276 Cheson BD et al (2014) Recommendations for initial evaluation, staging, and response assessment of Hodgkin and non-Hodgkin lymphoma: the Lugano classification. J Clin Oncol 32:3059–3068 Thompson CA et al (2014) Utility of routine post-therapy surveillance imaging in diffuse large B-cell lymphoma. J Clin Oncol 32:3506–3512 Huntington SF, Svoboda J, Doshi JA (2015) Cost-effectiveness analysis of routine surveillance imaging of patients with diffuse large B-cell lymphoma in first remission. J Clin Oncol 33:1467–1474 Cheson BD, Meignan M (2021) Current Role of Functional Imaging in the Management of Lymphoma. Curr Oncol Rep 23:144 Lauer EM, Mutter J, Scherer F (2022) Circulating tumor DNA in B-cell lymphoma: technical advances, clinical applications, and perspectives for translational research. Leukemia 36:2151–2164 Roschewski M et al (2015) Circulating tumour DNA and CT monitoring in patients with untreated diffuse large B-cell lymphoma: a correlative biomarker study. Lancet Oncol 16:541–549 Kurtz DM et al (2015) Noninvasive monitoring of diffuse large B-cell lymphoma by immunoglobulin high-throughput sequencing. Blood 125:3679–3687 Scherer F et al (2016) Distinct biological subtypes and patterns of genome evolution in lymphoma revealed by circulating tumor DNA. Sci Transl Med 8:364ra155 Rossi D et al (2017) Diffuse large B-cell lymphoma genotyping on the liquid biopsy. Blood 129:1947–1957 Kumar A et al (2020) Interim Analysis from a Prospective Multicenter Study of Next-Generation Sequencing Minimal Residual Disease Assessment and CT Monitoring for Surveillance after Frontline Treatment in Diffuse Large B-Cell Lymphoma. Blood 136:46–47 Meriranta L et al (2022) Molecular features encoded in the ctDNA reveal heterogeneity and predict outcome in high-risk aggressive B-cell lymphoma. Blood 139:1863–1877 Kurtz DM et al (2021) Enhanced detection of minimal residual disease by targeted sequencing of phased variants in circulating tumor DNA. Nat Biotechnol 39:1537–1547 Krupka JA et al (2025), The DIRECT study: A roadmap for ctDNA-based risk prediction, molecular profiling and MRD detection in Diffuse Large B Cell Lymphoma. medRxiv, 2025.2004.2014.25325806 Schmitt MW et al (2012) Detection of ultra-rare mutations by next-generation sequencing. Proc Natl Acad Sci U S A 109:14508–14513 Kennedy SR et al (2014) Detecting ultralow-frequency mutations by Duplex Sequencing. Nat Protoc 9:2586–2606 Meriranta L et al (2025) Circulating Tumor DNA Determinants of Response and Outcome in Relapsed/Refractory Mantle Cell Lymphoma. Blood Adv Leppa S et al (2025) Biomarker-adapted treatment in high-risk large B-cell lymphoma. Hemasphere 9:e70139 Kurtz DM et al (2018) Circulating Tumor DNA Measurements As Early Outcome Predictors in Diffuse Large B-Cell Lymphoma. J Clin Oncol 36:2845–2853 Kurtz DM et al (2019) Dynamic Risk Profiling Using Serial Tumor Biomarkers for Personalized Outcome Prediction. Cell 178:699–713e619 Roschewski M et al (2025) Remission Assessment by Circulating Tumor DNA in Large B-cell Lymphoma. J Clin Oncol, 101200JCO2501534 Mutter JA et al (2023) Circulating Tumor DNA Profiling for Detection, Risk Stratification, and Classification of Brain Lymphomas. J Clin Oncol 41:1684–1694 Heger JM et al (2024) Entirely noninvasive outcome prediction in central nervous system lymphomas using circulating tumor DNA. Blood 143:522–534 Alkodsi A, Meriranta L, Pasanen A, Leppä S, ctDNAtools (2020) An R package to work with sequencing data of circulating tumor DNA. bioRxiv , 2020.2001.2027.912790 Newman AM et al (2016) Integrated digital error suppression for improved detection of circulating tumor DNA. Nat Biotechnol 34:547–555 Additional Declarations Yes there is potential Competing Interest. H.H*: Genmab: Safety Committee; Incyte: Consultancy; Novartis: Research Funding; Elicera Therapeutics: Safety Committee; Serb SA: Advisory Board. J.J*: AbbVie: consultancy; BMS/Celgene: Consultancy; Gilead: Consultancy; Incyte: Consultancy; Novartis: Consultancy; Caribou Bioscience: Consultancy; Roche: Consultancy; SOBI: consultancy. P.B*: BMS: Consultancy; Gilead: Consultancy; Novartis: Consultancy; Roche: Consultancy. Ø.F. : Roche: Consultancy; Bayer: Consultancy. S.L : AbbVie: Honoraria, Consultation; GENMAB: Research Funding**; GILEAD: Honoraria; Incyte, Consultation, Honoraria; Novartis: Research Funding*; Roche: Honoraria, Consultancy, Research Funding*; BMS/Celgene: Consultancy, Research Funding**, Sobi: Honorary. Other authors declare no conflicts of interest. *All outside of the submitted work. **To the institute Supplementary Files SupplementaryTables.xlsx Supplementary Tables S1-S4 SupplementaryTablesFigures.pdf Supplementary Figure S1 , Supplementary Table and Figure legends Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-8007791","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Brief Communication","associatedPublications":[],"authors":[{"id":540465049,"identity":"3ab791d9-f0aa-4689-b626-83a235daa2c3","order_by":0,"name":"Sirpa Leppä","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAABCklEQVRIiWNgGAWjYDACZjBpASIYH4MYbEB4gMFGgpAWsAJmYxADoiUNjxYGhBY2aQiDDYjTcKuVb+d99uAHg4ScuXTzs+qCCgl7Pum2xAMMCRY4tRgcZjc37GGQMLacc8zs9owzEoltMscOALXgdpgBMxubBA+DROKGGwlmt3nbJBLYJNIbDjD+wK1FvpmNTfIPg0T9hhvp34p5/0nYg7Xgs4XhMBubNNCWBIMbOWbMvA0SjG0SaQQcdpiN3VjGQMJw55wzxdI8x4B+kUhLOJCAR4t8/zG2h28qbOTNpds3fuapsbGXn5Fm/OFDQh1uh4GjwQCIUIxNwKcBEnPoWkbBKBgFo2AUIAEAMpJF5CvWAlIAAAAASUVORK5CYII=","orcid":"https://orcid.org/0000-0002-8265-511X","institution":"University of Helsinki, Helsinki University Hospital Comprehensive Cancer Center and iCAN Digital Precision Medicine Flagship","correspondingAuthor":true,"prefix":"","firstName":"Sirpa","middleName":"","lastName":"Leppä","suffix":""},{"id":540465050,"identity":"b2c89f9f-c346-4ebe-bf06-4983f4b0ebb3","order_by":1,"name":"Leo Meriranta","email":"","orcid":"","institution":"University of Helsinki and Helsinki University Hospital Comprehensive Cancer Centre","correspondingAuthor":false,"prefix":"","firstName":"Leo","middleName":"","lastName":"Meriranta","suffix":""},{"id":540465051,"identity":"a77b703d-746e-4b08-b85c-d8cb6248f0d8","order_by":2,"name":"Maare Arffman","email":"","orcid":"","institution":"University of Helsinki, Helsinki University Hospital Comprehensive Cancer Center and iCAN Digital Precision Medicine Flagship","correspondingAuthor":false,"prefix":"","firstName":"Maare","middleName":"","lastName":"Arffman","suffix":""},{"id":540465052,"identity":"9724ff59-56d5-4cae-b9be-1c1b15a04dc9","order_by":3,"name":"Matias Autio","email":"","orcid":"","institution":"University of Helsinki and Helsinki University Hospital Comprehensive Cancer Centre","correspondingAuthor":false,"prefix":"","firstName":"Matias","middleName":"","lastName":"Autio","suffix":""},{"id":540465053,"identity":"ea819dac-9988-481a-a2fc-6e58f7278815","order_by":4,"name":"Judit Jørgensen","email":"","orcid":"","institution":"Aarhus University Hospital","correspondingAuthor":false,"prefix":"","firstName":"Judit","middleName":"","lastName":"Jørgensen","suffix":""},{"id":540465054,"identity":"244362b7-0eb7-4a67-8302-b3c6b66de996","order_by":5,"name":"Marja-Liisa Karjalainen-Lindsberg","email":"","orcid":"","institution":"Helsinki University Hospital","correspondingAuthor":false,"prefix":"","firstName":"Marja-Liisa","middleName":"","lastName":"Karjalainen-Lindsberg","suffix":""},{"id":540465055,"identity":"4d6c5303-2d18-4364-84a1-323a22e3fdf5","order_by":6,"name":"Klaus Beiske","email":"","orcid":"","institution":"Oslo University Hospital and University of Oslo","correspondingAuthor":false,"prefix":"","firstName":"Klaus","middleName":"","lastName":"Beiske","suffix":""},{"id":540465056,"identity":"0d3d7373-9b82-4015-b09a-0d850850efec","order_by":7,"name":"Mette Pedersen","email":"","orcid":"","institution":"Zealand University Hospital and University of Copenhagen","correspondingAuthor":false,"prefix":"","firstName":"Mette","middleName":"","lastName":"Pedersen","suffix":""},{"id":540465057,"identity":"4d4de65a-59aa-4ed9-b35d-80c995a5316b","order_by":8,"name":"Kristina Drott","email":"","orcid":"","institution":"Skane University Hospital","correspondingAuthor":false,"prefix":"","firstName":"Kristina","middleName":"","lastName":"Drott","suffix":""},{"id":540465058,"identity":"a2e49863-083f-4352-ae3c-8ce2bab3794f","order_by":9,"name":"Øystein Fluge","email":"","orcid":"","institution":"Haukeland University Hospital","correspondingAuthor":false,"prefix":"","firstName":"Øystein","middleName":"","lastName":"Fluge","suffix":""},{"id":540465059,"identity":"e1fac538-a492-45a5-b509-0e3c57579194","order_by":10,"name":"Sirkku Jyrkkiö","email":"","orcid":"","institution":"Turku University Hospital","correspondingAuthor":false,"prefix":"","firstName":"Sirkku","middleName":"","lastName":"Jyrkkiö","suffix":""},{"id":540465060,"identity":"2b7f62d8-b870-4c23-aaa6-42a1640369b2","order_by":11,"name":"Peter Brown","email":"","orcid":"https://orcid.org/0000-0002-6522-4086","institution":"Copenhagen University Hospital, Rigshospitalet","correspondingAuthor":false,"prefix":"","firstName":"Peter","middleName":"","lastName":"Brown","suffix":""},{"id":540465061,"identity":"93a85727-2e23-42b8-9a2e-66824dd4f432","order_by":12,"name":"Harald Holte","email":"","orcid":"https://orcid.org/0000-0001-9799-9428","institution":"Oslo University Hospital","correspondingAuthor":false,"prefix":"","firstName":"Harald","middleName":"","lastName":"Holte","suffix":""}],"badges":[],"createdAt":"2025-11-01 20:25:07","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-8007791/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8007791/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":95808425,"identity":"41c9a0c3-cc31-473f-a5c1-dc597b1e4aba","added_by":"auto","created_at":"2025-11-13 08:49:27","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":188578,"visible":true,"origin":"","legend":"\u003cp\u003eConcentration, clearance, and prognostic significance of circulating tumor DNA (ctDNA) in serial samples collected during chemoimmunotherapy.\u003c/p\u003e\n\u003cp\u003eA)\u0026nbsp;\u0026nbsp; Collection of plasma samples according to the study protocol.\u003c/p\u003e\n\u003cp\u003eB)\u0026nbsp;\u0026nbsp; Dendrogram of ctDNA detection and levels per patient. Colors according to end-of-therapy (EOT) response by PET-CT. Red, progressive disease (PD); Orange, partial response (PR); Blue, complete response (CR).\u003c/p\u003e\n\u003cp\u003eC)\u0026nbsp;\u0026nbsp; Kaplan-Meier (KM) survival estimate for overall survival (OS) according to presence of measurable residual disease (MRD) after two cycles of therapy.\u003c/p\u003e\n\u003cp\u003eD)\u0026nbsp;\u0026nbsp; KM survival estimate for OS among MRD\u003csup\u003eCYC2+\u003c/sup\u003e patients (n=37) according to ctDNA concentration after two cycles (categorized according to median concentration of 2.19 haploid genome equivalents per milliliter of plasma [GE/ml]).\u003c/p\u003e\n\u003cp\u003eE)\u0026nbsp;\u0026nbsp; KM survival estimate for OS according to major molecular response (MMR).\u003c/p\u003e\n\u003cp\u003eF)\u0026nbsp;\u0026nbsp;\u0026nbsp; KM survival estimate for OS among evaluable MRD\u003csup\u003eCYC2+\u003c/sup\u003e patients according to MRD after four cycles (n=27).\u003c/p\u003e\n\u003cp\u003eG)\u0026nbsp;\u0026nbsp; KM survival estimate for OS among available MRD\u003csup\u003eCYC4+\u003c/sup\u003e patients (n=15) according to ctDNA concentration after four cycles, categorized according to median concentration (6.31 GE/ml).\u003c/p\u003e\n\u003cp\u003eH)\u0026nbsp;\u0026nbsp; Box and dot plot of ctDNA concentration after four cycles (y-axis, GE/ml) according to EOT PET-CT responses.\u003c/p\u003e\n\u003cp\u003eI)\u0026nbsp;\u0026nbsp;\u0026nbsp;\u0026nbsp;\u0026nbsp; KM survival estimate for OS according to EOT MRD.\u003c/p\u003e\n\u003cp\u003eJ)\u0026nbsp;\u0026nbsp;\u0026nbsp; KM survival estimate for OS according to EOT MRD among the patients with complete metabolic response (Deauville score [D-S] 1-3).\u003c/p\u003e\n\u003cp\u003eK)\u0026nbsp;\u0026nbsp; KM survival estimate for OS according to EOT MRD among the patients with incomplete metabolic response (D-S 4-5).\u003c/p\u003e","description":"","filename":"Figure1.png","url":"https://assets-eu.researchsquare.com/files/rs-8007791/v1/7c4307f3612be942db207eff.png"},{"id":95808748,"identity":"40512c16-c6f4-45a6-86c3-70f562ac8c07","added_by":"auto","created_at":"2025-11-13 08:49:35","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":82543,"visible":true,"origin":"","legend":"\u003cp\u003eCirculating tumor DNA (ctDNA) detection on surveillance provides lead-time to clinical recurrence.\u003c/p\u003e\n\u003cp\u003eA) Dendrogram of ctDNA detection and levels per patient during treatment and surveillance. Patients with ctDNA detected are shown. MRD, measurable residual disease.\u003c/p\u003e\n\u003cp\u003eB) Swimmer’s plot of the patients (rows) and their follow-up (x-axis). MRD test results from surveillance samples drawn at protocol-specified time points are indicated in red (positive) and blue (negative) rectangles. Lead time to recurrence visualized with green bar.\u003c/p\u003e\n\u003cp\u003eC) Bar plot of lead times in evaluable patients with late relapse during surveillance. Grey bars represent cases with isolated recurrence in the central nervous system.\u003c/p\u003e","description":"","filename":"Figure2.png","url":"https://assets-eu.researchsquare.com/files/rs-8007791/v1/fbf1545c77c57be793d05f24.png"},{"id":95808688,"identity":"24efc112-0b60-4be9-94b0-c5838eaa013e","added_by":"auto","created_at":"2025-11-13 08:49:34","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":1615434,"visible":true,"origin":"","legend":"\u003cp\u003eAnalysis of circulating tumor DNA (ctDNA) in plasma and cerebrospinal fluid resolves clinical challenges.\u003c/p\u003e\n\u003cp\u003eA) Longitudinal analysis of plasma ctDNA, radiological imaging and tumor biopsies in a patient presenting with recurrence. Array comparative genomic hybridization was performed in clinical routine on primary and relapse tissues, revealing a shared copy number profile and regions of loss of heterozygosity between the biopsies, which confirmed a clonal relationship between the primary and relapse tumors.\u003c/p\u003e\n\u003cp\u003eB) Dot and line graph of reporter variants (dots) variant allele fractions (y-axis) in serially collected (x-axis) plasma (top) and cerebrospinal fluid (bottom) samples in a case with isolated central nervous system recurrence on follow-up. Illustrative graphics created using biorender.com.\u003c/p\u003e","description":"","filename":"Figure3.png","url":"https://assets-eu.researchsquare.com/files/rs-8007791/v1/bf0e6724977f5ba4fc60be0a.png"},{"id":95810617,"identity":"7c3a3be7-fde2-4f13-b164-8fb15391af29","added_by":"auto","created_at":"2025-11-13 08:53:14","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":2306904,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8007791/v1/7592ed40-ebcc-479e-ac20-75329c44699b.pdf"},{"id":95808931,"identity":"f35860fd-54a6-4d70-9b57-ae9a26579695","added_by":"auto","created_at":"2025-11-13 08:49:43","extension":"xlsx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":116186,"visible":true,"origin":"","legend":"Supplementary Tables S1-S4","description":"","filename":"SupplementaryTables.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-8007791/v1/2908d382b585641e6c072ea4.xlsx"},{"id":95808439,"identity":"6aaa33dd-5076-420d-8129-5534c2c0f91d","added_by":"auto","created_at":"2025-11-13 08:49:28","extension":"pdf","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":627261,"visible":true,"origin":"","legend":"Supplementary Figure S1 , Supplementary Table and Figure legends","description":"","filename":"SupplementaryTablesFigures.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8007791/v1/c7d59b2339b7fda6702afede.pdf"}],"financialInterests":"\u003cb\u003eYes\u003c/b\u003e there is potential Competing Interest.\nH.H*: Genmab: Safety Committee; Incyte: Consultancy; Novartis: Research Funding; Elicera Therapeutics: Safety Committee; Serb SA: Advisory Board. J.J*: AbbVie: consultancy; BMS/Celgene: Consultancy; Gilead: Consultancy; Incyte: Consultancy; Novartis: Consultancy; Caribou Bioscience: Consultancy; Roche: Consultancy; SOBI: consultancy. P.B*: BMS: Consultancy; Gilead: Consultancy; Novartis: Consultancy; Roche: Consultancy. Ø.F.*: Roche: Consultancy; Bayer: Consultancy. S.L*: AbbVie: Honoraria, Consultation; GENMAB: Research Funding**; GILEAD: Honoraria; Incyte, Consultation, Honoraria; Novartis: Research Funding*; Roche: Honoraria, Consultancy, Research Funding*; BMS/Celgene: Consultancy, Research Funding**, Sobi: Honorary. Other authors declare no conflicts of interest. \r\n*All outside of the submitted work. \r\n**To the institute","formattedTitle":"Molecular Surveillance of Aggressive Large B-cell Lymphoma using Circulating Tumor DNA Duplexes","fulltext":[{"header":"Main","content":"\u003cp\u003eDisease recurrence affects 30% of patients with aggressive large B-cell lymphoma (LBCL) and predicts poor outcomes\u003csup\u003e1\u003c/sup\u003e. Treating LBCL relapse is challenging\u003csup\u003e2\u003c/sup\u003e, and high tumor volumes predict poor survival\u003csup\u003e3\u003c/sup\u003e. Early detection of recurrence could improve outcomes; therefore, treatment responses are monitored, and patients undergo imaging with computed tomography (CT) or positron-emission tomography (PET), although their utility in surveillance is limited\u003csup\u003e4,5\u003c/sup\u003e. Repeated imaging exposes patients to radiation, the sensitivity of these methods is modest, and the poor specificity leads to invasive and unnecessary biopsies, which potentially harm patients\u003csup\u003e6,7\u003c/sup\u003e. Consequently, more accurate methods are needed for monitoring response and recurrence.\u003c/p\u003e\n\u003cp\u003eLiquid biopsies using plasma circulating tumor DNA (ctDNA) complement diagnostics, predict treatment response, and inform survival in LBCL\u003csup\u003e8\u003c/sup\u003e. These tools are expected to transform decision-making\u003csup\u003e9-12\u003c/sup\u003e. The presence of ctDNA in plasma at diagnosis and recurrence is well-established; however, detecting it before clinical relapse requires sensitive methods\u003csup\u003e8\u003c/sup\u003e. Early studies utilized immunoglobulin high-throughput sequencing (Ig-HTS) to detect ctDNA, showing promising results in response evaluation and surveillance\u003csup\u003e9,10\u003c/sup\u003e. However, Ig-HTS lacks sensitivity, tracking only single or a few genomic events, resulting in a disappointing 56% detection rate before or at relapse in a prospective study\u003csup\u003e13\u003c/sup\u003e. Hybrid-capture panel techniques detect ctDNA more sensitively by tracing more mutations, but sequencing error rates limit their performance\u003csup\u003e11,14\u003c/sup\u003e. Phased variants are unlikely to be detected by error, and can be harnessed for sensitive disease monitoring\u003csup\u003e14,15\u003c/sup\u003e. Still, phased events are rare, often found in hypermutable loci affected by clonal heterogeneity, and disease detection mainly depends on the number of informative molecules\u003csup\u003e16\u003c/sup\u003e. Duplex sequencing, which recovers both strands of the original DNA for error correction\u003csup\u003e17\u003c/sup\u003e, achieves an extremely low error rate and can detect mutations down to parts per million (ppm)\u003csup\u003e18\u003c/sup\u003e. We reasoned that longitudinal duplex sequencing of lymphoma ctDNA could demonstrate the potential of detecting measurable residual disease (MRD) over time in LBCL.\u003c/p\u003e\n\u003cp\u003eWe aimed to determine the added clinical value of ctDNA detection by duplex sequencing in serial plasma samples from LBCL patients. We applied a previously established targeted duplex sequencing-based ctDNA assay\u003csup\u003e19\u003c/sup\u003e\u003csup\u003e,\u003c/sup\u003e\u003csup\u003e20\u003c/sup\u003e on 447 longitudinal plasma cfDNA samples from 123 patients with primary LBCL at various timepoints according to a prospective protocol (Figure 1A)\u003csup\u003e20\u003c/sup\u003e. The patients were young (aged 18-65 years), had high-risk LBCL, and were treated with a curative intent with early high-dose-methotrexate, dose-intensive immunochemotherapy (DA-EPOCH-R or R-CHOEP), and high-dose-cytarabine in a Nordic phase II trial\u003csup\u003e20\u003c/sup\u003e. The lowest ctDNA fractions detected were in the ppm range, and the median limit-of-detection was 1.69*10\u003csup\u003e-5\u003c/sup\u003e with 26% of the samples having a limit-of-detection in the ppm range. Statistical analyses were performed in the R environment.\u0026nbsp;Additional details are described in the Supplementary methods and Tables S1-S4.\u003c/p\u003e\n\u003cp\u003eDuring an updated median follow-up of 60.4 months, 22 patients experienced progression, including 17 relapses and 13 deaths. Pretreatment ctDNA levels declined in all evaluable patients after two cycles of immunochemotherapy, with 67% (75/112) reaching molecular remission (MRD\u003csup\u003eCYC2–,\u0026nbsp;\u003c/sup\u003eFigure 1B, S1A). MRD\u003csup\u003eCYC2\u003c/sup\u003e was prognostic; yet, 68% (25/37) of patients with MRD\u003csup\u003eCYC2+\u0026nbsp;\u003c/sup\u003edid not relapse (Figure 1C). However, the negative predictive value (NPV) of MRD\u003csup\u003eCYC2\u003c/sup\u003e evaluation for relapse was high (95%, Table S2-S3). Notably, among the 37 MRD\u003csup\u003eCYC2+\u0026nbsp;\u003c/sup\u003ecases, elevated ctDNA levels after cycle 2 were predictive of refractory disease and worse survival (Figure 1D). Major molecular response\u003csup\u003e21\u003c/sup\u003e (MMR, ≥2.5 log reduction in ctDNA) was achieved by 85% of patients (94/111), and failure to reach MMR identified patients with poor survival; however, its sensitivity was lower than that of MRD\u003csup\u003eCYC2\u003c/sup\u003e (Figure 1E, Table S2). Overall, clearance of ctDNA after 2 cycles predicts excellent survival, and in cases where ctDNA is present, high levels of ctDNA indicate worse survival.\u003c/p\u003e\n\u003cp\u003eTwenty-seven patients with MRD\u003csup\u003eCYC2+\u003c/sup\u003e were evaluable for ctDNA after 4 cycles (MRD\u003csup\u003eCYC4\u003c/sup\u003e). Of these, 14 (52%) had become MRD-negative, while 13 (48%) remained MRD-positive, which was associated with poor survival (Figure 1F, S1B). In patients with MRD\u003csup\u003eCYC4+\u003c/sup\u003e, high ctDNA levels predicted poor survival similarly to MRD\u003csup\u003eCYC2+\u003c/sup\u003e,and primary refractory cases were identified by ctDNA levels above 10 hGE/ml (Figure 1G-H). After the last cycles of therapy (MRD\u003csup\u003eEOT\u003c/sup\u003e), 7 of the 13 patients (54%) with MRD\u003csup\u003eCYC4+\u0026nbsp;\u003c/sup\u003ebecame MRD-negative, while 6 patients (46%) remained MRD-positive. These findings reveal different response kinetics that update prognostic information on therap\u003csup\u003e22\u003c/sup\u003e, and interim ctDNA quantification identifies patients unlikely to respond.\u003c/p\u003e\n\u003cp\u003eAmong all evaluable patients, MRD\u003csup\u003eEOT\u003c/sup\u003e identified a dramatic difference in outcomes, with all patients who died within two years of follow-up being MRD\u003csup\u003eEOT+\u0026nbsp;\u003c/sup\u003e(Figure 1I, \u0026nbsp;S1C, Table S2-3). MRD\u003csup\u003eEOT+\u0026nbsp;\u003c/sup\u003edetected all 4 patients with primary refractory disease, and 42% (5/12) of those progressed later during follow-up. The median time to progression for MRD\u003csup\u003eEOT+\u0026nbsp;\u003c/sup\u003epatients was 8.4 months, compared to 31.2 months for MRD\u003csup\u003eEOT–\u0026nbsp;\u003c/sup\u003e(P=0.0008, Mann-Whitney-U), and the NPVs of MRD\u003csup\u003eEOT–\u0026nbsp;\u003c/sup\u003efor 1-, 2-, and 4-year PFS were 98%, 94%, and 91%, respectively. MRD\u003csup\u003eEOT\u003c/sup\u003e was associated with Deauville scores (DS), showing a 5.6% (4/71) MRD\u003csup\u003eEOT+\u0026nbsp;\u003c/sup\u003erate for DS 1-3 and 46% (6/13) for DS 4-5 (Fisher’s P=0.0051, Figure 1J, S1D). MRD\u003csup\u003eEOT\u003c/sup\u003e was positive in 50% (3/6) and 71% (5/7) of patients with relapse in the DS 1-3 and DS 4-5 groups, respectively, with 1 false positive in each category, demonstrating a significant prognostic impact (Figure 1J-K, S1E-F). These findings highlight the excellent prognostic value of molecular response measurement in LBCL and are highly consistent with recent results reported using ctDNA assays with a similar error rate as duplex sequencing\u003csup\u003e16,23\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003eSeven patients underwent tissue biopsies from PET-positive lesions after therapy, with 2 confirmed lymphoma progressions (29%). MRD\u003csup\u003eEOT+\u003c/sup\u003e was detected in 3 of these 7 patients, including both confirmed progressions. Twenty percent (20/99) of the evaluable patients for MRD\u003csup\u003eEOT\u003c/sup\u003e received radiotherapy (RT), which was not associated with MRD\u003csup\u003eEOT\u003c/sup\u003e status (Fisher’s exact, P=0.69, Figure S1G). Six patients received RT due to tumor bulk at diagnosis, and all were MRD\u003csup\u003eEOT–\u003c/sup\u003e.Three of the 11 MRD\u003csup\u003eEOT+\u0026nbsp;\u003c/sup\u003epatients received RT for PET-avid lesions, and in two of those with available surveillance samples, MRD remained positive after RT, with the patients relapsing during follow-up. These results highlight the limitations of current clinical tools in providing appropriate interventions for patients at higher risk of progression after immunochemotherapy, a gap that could be addressed with ctDNA analysis.\u003c/p\u003e\n\u003cp\u003eOn post-treatment surveillance, all evaluable patients with MRD\u003csup\u003eEOT+\u0026nbsp;\u003c/sup\u003eremained MRD-positive in subsequent plasma samples, and 86% (6/7) of the relapsing patients with MRD\u003csup\u003eEOT–\u0026nbsp;\u003c/sup\u003ebecame MRD-positive before clinical relapse (Figure 2A-B). Only one patient with late relapse was MRD-negative for over a year of follow-up until turning MRD-positive. Once ctDNA was detected, 93% (14/15) of subsequent tests remained positive, whereas the specificity remained 100% in patients without relapse, including 1 patient with non-lymphoma-related death (22 tests, Figure 2B).\u003c/p\u003e\n\u003cp\u003eSurveillance provided lead time before relapse in 92% (11/12) of patients (median=6 months). A lead time of 3 months was achieved in 83% of cases (10/12, Figure 2C). Three patients had a lead time of over 12 months, including one with ctDNA detected throughout measurements (Figure 3A). The patient developed a symptomatic parailiacal tumor 3 months after metabolic CR and responded to corticosteroids; however, tissue biopsies showed an unspecific histiocytic reaction twice before a conclusive diagnosis was made using genomic techniques, as tumor cells had lost B-cell phenotypic markers (Figure 3A).\u003c/p\u003e\n\u003cp\u003eFinally, patients with short ctDNA lead times to recurrence (0 and 1.6 months) were the only cases with isolated CNS recurrence (n=2, Figure 2C). Since mutations in primary CNS lymphomas can be more easily detected in cerebrospinal fluid (CSF)\u003csup\u003e24,25\u003c/sup\u003e, we analyzed cfDNA from CSF samples available for one of the patients, who had CSF flow cytometry positivity at inclusion. Exploratory analysis revealed lymphoma reporter mutations in the pretreatment CSF, all of which became undetectable after two cycles of therapy, reflecting ctDNA response in plasma (Figure 3B). Despite flow cytometry clearance after therapy, however, lymphoma reporters were again detectable in the CSF at EOT, indicating CNS-restricted resistance with an additional three months’ lead time compared to plasma-based detection (Figure 3B). These findings demonstrate significant lead times to systemic LBCL relapses with ctDNA surveillance, while CSF analysis could help in early detection of CNS recurrence.\u003c/p\u003e\n\u003cp\u003eHere, we showed that longitudinal plasma ctDNA detection accurately predicts survival, determines therapy responses, and clarifies various clinical scenarios upon LBCL recurrence. Treatment could be personalized earlier based on on-therapy ctDNA measurements, and clinical interventions could be more precisely directed at treatment-refractory patients using ctDNA analysis. The similar prognostic significance observed in patients with either radiological remission or residual positivity, along with the high NPV of ctDNA clearance, supports incorporating molecular remission into LBCL response criteria to complement imaging\u003csup\u003e23\u003c/sup\u003e. Late recurrences could be detected earlier through repeated post-therapy surveillance and intercepted before a high tumor burden develops. Surveillance could be tailored according to dynamic risk profiling, and CSF analysis could supplement surveillance in patients with suspected CNS recurrence. Prospective studies are warranted to confirm the feasibility and potential benefits of ctDNA-guided therapy. The comprehensive clinical data and sampling, long median follow-up, and ctDNA detection down to ppm with duplex sequencing are strengths of our study. Further enhancements in sequencing design, analysis of larger plasma volumes, and inclusion of phased variants could further improve the practical sensitivity of ctDNA detection.\u003c/p\u003e"},{"header":"Online Methods","content":"\u003cp\u003e\u003cem\u003eClinical data\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003e123 patients treated in Nordic phase II protocol (NCT03293173) were included in the study with 119 patients with samples evaluable for circulating tumor DNA (ctDNA) analysis. The clinical trial results, study treatment and patient characteristics have been previously reported\u003csup\u003e20\u003c/sup\u003e. Briefly, young patients (≥65 years of age) with de novo aggressive large B-cell lymphoma (LBCL) were recruited and treated with a curative intent. The treatment consisted of early central nervous system (CNS) prophylaxis and dose-dense or dose-adjusted chemoimmunotherapy (R-CHOEP or DA-EPOCH-R) according to tumor biomarker profile. The follow-up data of the patients was updated for the current study. Lead times to recurrence were calculated as the difference between the date of first ctDNA positive plasma sample draw and the date of recurrence.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eSamples, sample processing and sequencing\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eThe reagents and resources used in the study are listed in Table S1. Peripheral blood drawn in K2-EDTA tubes at inclusion was used as source of matched normal DNA and was processed as previously described\u003csup\u003e20\u003c/sup\u003e. Residual tumor DNA derived from diagnostic fresh frozen biopsies or formalin-fixed paraffin-embedded tissue were analyzed when available.\u003c/p\u003e\n\u003cp\u003ePlasma samples used for cell-free DNA (cfDNA) extraction were collected according to study protocol at pretreatment, on-therapy after second and fourth cycles (interim computer tomography [CT] stagings), and at the end-of-therapy (EOT) at the time of PET-CT end-staging (Figure 1A). Additionally, surveillance samples were collected on follow-up at 3, 6, 9, 12, 18 and 24 months and at relapse if present. Ten milliliters of peripheral blood was drawn to standard K2-EDTA or white blood cell preserving (cell-free DNA BCT, Streck) tubes, and the plasma was immediately separated, aliquoted and stored. In addition, cerebrospinal-fluid (CSF) samples were collected from spinal taps at pretreatment, and if tumor cells were detected by flow cytometry, CSF samples were collected after 2 cycles and at EOT (one patient in the study reported in Figure 3B, 7 CSF samples studied in total). 2-4 ml of cerebrospinal fluid was be collected in a sterile test tube which was centrifuged at 2,000g for 10 minutes at +4 C to remove cell debris and the supernatant was transferred to vials and stored at -70 C. Extraction of cfDNA from CSF was performed similarly to plasma samples.\u003c/p\u003e\n\u003cp\u003eIn total, 119 pretreatment and 324 on-therapy or post-therapy surveillance samples from 119 patients were analyzed. Sample selection of pretreatment, after 2 cycles (n = 112) and EOT (n = 99) plasma were based on sample availability. Plasma samples drawn after cycle 4 (n=55) were enriched for patients with follow-up events with all available samples from patients with events and all available samples from patients with measurable residual disease (MRD) after 2 cycles of therapy (MRD\u003csup\u003eCYC2+\u003c/sup\u003e) were analyzed (n=27/55). All available post-therapy surveillance samples from patients with events (progression; recurrence or death from any cause) were analyzed with a control series of patients with no follow-up events based on baseline ctDNA positivity and sample availability (n=54). Additionally, 4 plasma samples drawn upon recurrence were included in the study.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003ectDNA analysis\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003ePreviously established platform using targeted panel of ~750 kilobases in size with exhaustive duplex sequencing was used for ctDNA analysis\u003csup\u003e19\u003c/sup\u003e. Centralized cfDNA and DNA extraction, quantification and quality control steps and library preparation, target enrichment, sequencing and raw data processing were performed as previously described\u003csup\u003e19,20\u003c/sup\u003e. Sequencing metrics of cfDNA samples are reported in table S4A and analysis metrics are shown in table S4B.\u003c/p\u003e\n\u003cp\u003ePresence of MRD was estimated using adapted software\u003csup\u003e26\u003c/sup\u003e with concepts described previously\u003csup\u003e27\u003c/sup\u003e. All tests were run and the samples analyzed blinded to the clinical data using predetermined test parameters. Reporter mutations were single nucleotide variants not in predetermined clonal hematopoiesis target genes identified in genotyping diagnostic tumor tissues with ≥10% variant allele frequency (VAF) and/or in pretreatment plasma cfDNA with ≥1% VAF. Patients with below 20 reporters (n=4/119) were considered genotyping failures and regarded as MRD-negative for all the tested samples. Median number of reporters in evaluable patients was 252 (range 24–1316). Only duplex corrected reads with mapping quality of 60 and bases with base quality of 60 were considered in the MRD analysis. Sample specific background error rate was measured as any mismatch from reference and the median error rate was 2.74e-06 (range 3.18e-07–1.09e-05). 128 (40%) plasma cfDNA samples were drawn in Streck tubes and 196 (60%) in tubes with EDTA as anticoagulant, and no difference in background rate or number of informative duplex reads was detected between the starting materials (Mann-Whitney U tests, \u003cem\u003eP\u003c/em\u003e \u0026gt; 0.05, data in Table S4B). Results from Monte-Carlo based framework were interpreted with a predetermined cut-off for empirical p-value calibrated to 95% specificity among withheld controls determined previously\u003csup\u003e20\u003c/sup\u003e. The ctDNA levels at interim time-points were measured as haploid genomic equivalents (hGE/ml), and the median levels were 2.19 hGE/ml after 2 cycles and 6.31 hGE/ml after 4 cycles of therapy.\u003c/p\u003e\n\u003cp\u003eThe analytical sensitivity of duplex sequencing allows mutation detection below parts per million\u003csup\u003e17\u003c/sup\u003e. In the context of cfDNA with limited input material, the sample specific sensitivity depends on the number of informative molecules, and therefore, the sample-wise limit of detection (LOD) was estimated with a binomial-based detection model using the informative duplex depth assuming 95% sensitivity and 5% false-positive rate (LOD95) similarly as previously described\u003csup\u003e16\u003c/sup\u003e. Informative duplex depth per patient was defined as the sum of altered and reference counts from reporter mutation positions meeting MRD test criteria above. 96% (301/313) of the samples from evaluable patients reached below 10^-4 LOD95, the median LOD95 was 1.69e-5, and 26% (82/313) of the samples reached LOD95 in 10^-6 range (Table S4B). In line with these models, the lowest fractions of ctDNA molecules in the study were detected in the parts per million range (Table S4B).\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eImmunohistochemical images\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eSlides of immunohistochemical stainings from routine diagnostics were performed in Helsinki University Hospital, department of Pathology and were available for the current study. Microscopy images were acquired with Leica\u0026nbsp;DM\u0026nbsp;LB brightfield microscope (Leica Microsystems\u0026nbsp;GmbH) with Olympus\u0026nbsp;DP50 camera.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eStatistical analysis\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eStatistical analyses were performed in the R environment (R foundation for Statistical Computing, Vienna, Austria, Table S1). Statistical details accompany the results in the text and/or related figure legends. The used tests were non-parametric and two-sided, and \u003cem\u003eP\u0026nbsp;\u003c/em\u003evalues \u0026lt; 0.05 were considered statistically significant. The Kaplan-Meier method with log rank test was used to estimate survival rates between different patient groups and Cox regression was used to estimate survival in univariate models which are available in Table S3 (R package survival, 3.8-3). Survival was estimated and the results interpreted using the follow-up times without considering time-point of the ctDNA test. Test performance metrics were evaluated and the cross-tabulation and confusion matrixes generated with the R package caret (version 7.0-1).\u003c/p\u003e"},{"header":"Declarations","content":"\u003ch2\u003e\u003cstrong\u003eAcknowledgements\u003c/strong\u003e\u003c/h2\u003e\n\u003cp\u003eWe thank all participating study sites for patient enrolment and sample collection, Laura Hakala for trial coordination, the staff at the Clinical Trial Office in Aarhus, Denmark for trial management,\u0026nbsp;Anne Aarnio and Emma Saarinen for meticulous laboratory work, Henrikki Almusa for sequencing data processing, and Sari Hannula for guiding the design and coordination of sequencing. Library preparation and sequencing data processing were performed by the FIMM Genomics NGS Sequencing unit at the University of Helsinki, supported by HiLIFE and Biocenter Finland. CSC IT (Center for Science, Finland) is acknowledged for providing the computational resources for this study.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConflicts of interest\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eH.H*: Genmab: Safety Committee; Incyte: Consultancy; Novartis: Research Funding; Elicera Therapeutics: Safety Committee; Serb SA: Advisory Board. J.J*: Abbvie: consultancy; BMS/Celgene: Consultancy; Gilead: Consultancy; Incyte: Consultancy; Novartis: Consultancy; Caribou Bioscience: Consultancy; Roche: Consultancy; \u0026nbsp;SOBI: consultancy. \u0026nbsp;P.B*: BMS: Consultancy; Gilead: Consultancy; Novartis: Consultancy; Roche: Consultancy. Ø.F.*: Roche: Consultancy; Bayer: Consultancy. S.L*: AbbVie: Honoraria, Consultation; GENMAB: Research Funding**; GILEAD: Honoraria; Incyte, Consultation, Honoraria; Novartis: Research Funding*; Roche: Honoraria, Consultancy, Research Funding*; BMS/Celgene: Consultancy, Research Funding**, Sobi: Honorary. Other authors declare no conflicts of interest.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e*All outside of the submitted work.\u0026nbsp;**To the institute\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEthics approval and trial registration\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe study was conducted in accordance with the Guidelines on Good Clinical Practice from the International Conference on Harmonization and the principles outlined in the Declaration of Helsinki. The protocol was approved by the medical agencies and ethics committees in Finland, Denmark, Norway, and Sweden, and the trial was registered at ClinicalTrials.gov, number NCT01325194. All patients signed informed consent forms before participating in the study.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor contributions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eConceptualization: LM, SL\u003c/p\u003e\n\u003cp\u003eSample and clinical data collection\u0026nbsp;and processing: MAr, JJ, MLK, KB, MP, KD, ØF, SJ, PB, HH, SL\u003c/p\u003e\n\u003cp\u003eInvestigation: LM, MAu, SL\u003c/p\u003e\n\u003cp\u003eFormal analysis: LM\u003c/p\u003e\n\u003cp\u003eData interpretation: All authors\u003c/p\u003e\n\u003cp\u003eFunding acquisition: SL\u003c/p\u003e\n\u003cp\u003eWriting - Original Draft Preparation: LM\u003c/p\u003e\n\u003cp\u003eWriting - Review \u0026amp; Editing: SL\u003c/p\u003e\n\u003cp\u003eFinal approval of manuscript: All authors\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis study was funded by the Research Council of Finland (SL), Finnish Cancer Organizations, iCAN Flagship (SL), Nordic Cancer Union (SL), Sigrid Juselius Foundation (SL), and Helsinki and Uusimaa Hospital District (SL). LM was supported by personal grants from the Finnish Medical Foundation and Sigrid Juselius Foundation.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eSehn LH, Salles G (2021) Diffuse Large B-Cell Lymphoma. N Engl J Med 384:842\u0026ndash;858\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eChaganti S et al (2025) Management of relapsed or refractory large B-cell lymphoma: A British Society for Haematology Guideline. Br J Haematol\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eDean EA et al (2020) High metabolic tumor volume is associated with decreased efficacy of axicabtagene ciloleucel in large B-cell lymphoma. Blood Adv 4:3268\u0026ndash;3276\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eCheson BD et al (2014) Recommendations for initial evaluation, staging, and response assessment of Hodgkin and non-Hodgkin lymphoma: the Lugano classification. J Clin Oncol 32:3059\u0026ndash;3068\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eThompson CA et al (2014) Utility of routine post-therapy surveillance imaging in diffuse large B-cell lymphoma. J Clin Oncol 32:3506\u0026ndash;3512\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHuntington SF, Svoboda J, Doshi JA (2015) Cost-effectiveness analysis of routine surveillance imaging of patients with diffuse large B-cell lymphoma in first remission. J Clin Oncol 33:1467\u0026ndash;1474\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eCheson BD, Meignan M (2021) Current Role of Functional Imaging in the Management of Lymphoma. Curr Oncol Rep 23:144\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLauer EM, Mutter J, Scherer F (2022) Circulating tumor DNA in B-cell lymphoma: technical advances, clinical applications, and perspectives for translational research. Leukemia 36:2151\u0026ndash;2164\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eRoschewski M et al (2015) Circulating tumour DNA and CT monitoring in patients with untreated diffuse large B-cell lymphoma: a correlative biomarker study. Lancet Oncol 16:541\u0026ndash;549\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKurtz DM et al (2015) Noninvasive monitoring of diffuse large B-cell lymphoma by immunoglobulin high-throughput sequencing. Blood 125:3679\u0026ndash;3687\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eScherer F et al (2016) Distinct biological subtypes and patterns of genome evolution in lymphoma revealed by circulating tumor DNA. Sci Transl Med 8:364ra155\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eRossi D et al (2017) Diffuse large B-cell lymphoma genotyping on the liquid biopsy. Blood 129:1947\u0026ndash;1957\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKumar A et al (2020) Interim Analysis from a Prospective Multicenter Study of Next-Generation Sequencing Minimal Residual Disease Assessment and CT Monitoring for Surveillance after Frontline Treatment in Diffuse Large B-Cell Lymphoma. Blood 136:46\u0026ndash;47\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMeriranta L et al (2022) Molecular features encoded in the ctDNA reveal heterogeneity and predict outcome in high-risk aggressive B-cell lymphoma. Blood 139:1863\u0026ndash;1877\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKurtz DM et al (2021) Enhanced detection of minimal residual disease by targeted sequencing of phased variants in circulating tumor DNA. Nat Biotechnol 39:1537\u0026ndash;1547\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKrupka JA et al (2025), \u003cem\u003eThe DIRECT study: A roadmap for ctDNA-based risk prediction, molecular profiling and MRD detection in Diffuse Large B Cell Lymphoma. medRxiv, 2025.2004.2014.25325806\u003c/em\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSchmitt MW et al (2012) Detection of ultra-rare mutations by next-generation sequencing. Proc Natl Acad Sci U S A 109:14508\u0026ndash;14513\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKennedy SR et al (2014) Detecting ultralow-frequency mutations by Duplex Sequencing. Nat Protoc 9:2586\u0026ndash;2606\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMeriranta L et al (2025) Circulating Tumor DNA Determinants of Response and Outcome in Relapsed/Refractory Mantle Cell Lymphoma. Blood Adv\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLeppa S et al (2025) Biomarker-adapted treatment in high-risk large B-cell lymphoma. Hemasphere 9:e70139\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKurtz DM et al (2018) Circulating Tumor DNA Measurements As Early Outcome Predictors in Diffuse Large B-Cell Lymphoma. J Clin Oncol 36:2845\u0026ndash;2853\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKurtz DM et al (2019) Dynamic Risk Profiling Using Serial Tumor Biomarkers for Personalized Outcome Prediction. Cell 178:699\u0026ndash;713e619\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eRoschewski M et al (2025) Remission Assessment by Circulating Tumor DNA in Large B-cell Lymphoma. J Clin Oncol, 101200JCO2501534\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMutter JA et al (2023) Circulating Tumor DNA Profiling for Detection, Risk Stratification, and Classification of Brain Lymphomas. J Clin Oncol 41:1684\u0026ndash;1694\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHeger JM et al (2024) Entirely noninvasive outcome prediction in central nervous system lymphomas using circulating tumor DNA. Blood 143:522\u0026ndash;534\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eAlkodsi A, Meriranta L, Pasanen A, Lepp\u0026auml; S, ctDNAtools (2020) An R package to work with sequencing data of circulating tumor DNA. \u003cem\u003ebioRxiv\u003c/em\u003e, 2020.2001.2027.912790\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eNewman AM et al (2016) Integrated digital error suppression for improved detection of circulating tumor DNA. Nat Biotechnol 34:547\u0026ndash;555\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"Circulating tumor DNA, duplex sequencing, measurable residual disease, B-cell lymphoma, surveillance, relapse, longitudinal","lastPublishedDoi":"10.21203/rs.3.rs-8007791/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8007791/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"Treatment failure affects one-third of patients with aggressive large B-cell lymphoma (LBCL), and current tools are inadequate in determining remission. We investigated the added value of longitudinal ctDNA monitoring in plasma by duplex sequencing. Clearance of ctDNA after 2 or 4 cycles of immunochemotherapy identified patients with excellent survival, whereas elevated interim ctDNA levels among those with measurable residual disease (MRD) predicted refractory disease. Two-year progression-free survival rates based on end-of-therapy MRD varied greatly (94.3% vs 18.2%, negative and positive, respectively), and ctDNA analysis resolved most false radiological response assessments and revealed poor targeting of post-treatment interventions to MRD-positive cases. In surveillance, ctDNA was detectable at least 3 months before clinical relapse, while patients with isolated central nervous system relapses had the shortest plasma ctDNA lead times. Overall, ctDNA monitoring with ultra-sensitive duplex sequencing offers a dynamic molecular tool that can guide clinical decision-making in LBCL.","manuscriptTitle":"Molecular Surveillance of Aggressive Large B-cell Lymphoma using Circulating Tumor DNA Duplexes","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-11-13 08:20:30","doi":"10.21203/rs.3.rs-8007791/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"787c18c7-f4e4-421b-8be9-4a5150819560","owner":[],"postedDate":"November 13th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":57492504,"name":"Health sciences/Medical research/Outcomes research"},{"id":57492505,"name":"Biological sciences/Cancer/Haematological cancer/Lymphoma/Non-hodgkin lymphoma/B-cell lymphoma"},{"id":57492506,"name":"Health sciences/Biomarkers/Prognostic markers"}],"tags":[],"updatedAt":"2026-02-23T22:45:43+00:00","versionOfRecord":[],"versionCreatedAt":"2025-11-13 08:20:30","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-8007791","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8007791","identity":"rs-8007791","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}