Abstract
Background/Aim: This study evaluated the possibility of clinical use of circulating-tumor DNA (ctDNA) as a biomarker to determine up-front autologous stem cell transplantation (auto-SCT) for patients with high-risk diffuse large B-cell lymphoma (DLBCL) in practice. Patients and Methods: To explore the dynamics of ctDNA in DLBCL, blood samples were collected sequentially before and after treatment from patients with newly diagnosed DLBCL who received rituximab, cyclophosphamide, doxorubicin, vincristine, and prednisone (R-CHOP) chemotherapy. To conduct ctDNA genotyping and ctDNA monitoring simultaneously, targeted sequencing by cancer personalized profiling using deep sequencing was used. Results: Ten patients between the ages of 50 and 60 years were enrolled. Based on the international prognostic index (IPI), seven patients were classified as high-IPI-risk group, and three patients were classified as low-IPI-risk group. The IPI risk group correlated with total metabolic tumor volume. All patients completed six cycles of R-CHOP chemotherapy, and seven patients achieved complete response. Changes in ctDNA mutation numbers did not correlate with changes in PET scan images and treatment response. In most high-risk patients, new mutations appeared in ctDNA after completion of chemotherapy that conceivably marked resistant clones. Notably, disease relapse did not occur in high-risk patients with poor prognostic mutations who underwent autologous SCT. Conclusion: ctDNA monitoring was meaningful in high-risk patients. Moreover, ctDNA and well-known prognostic factors should be considered in the decision making for auto-SCT. If a new genetic mutation in ctDNA with a negative prognosis would emerge during treatment, high-risk patients should consider auto-SCT.
In patients with newly diagnosed diffuse large B-cell lymphoma (DLBCL), the main treatment consists of rituximab and immunochemotherapy. Approximately 60%-70% of patients who received standard R-CHOP (rituximab, cyclophosphamide, doxorubicin, vincristine, and prednisone) chemotherapy achieved cure. However, more than 30%-40% of patients suffer from relapse/refractory disease after front-line therapy, and the majority of these patients exhibit extremely poor survival outcomes (1). For patients, the aim is to avoid relapse through the accurate prediction of disease progression by the attending physicians, given that novel therapeutic options for salvage therapy have emerged (1, 2). The addition of an appropriate consolidation therapy could improve long-term survival in patients at high risk for disease relapse. A retrospective study with 219 newly diagnosed patients with DLBCL has been reported by Kim et al. Patients who received up-front autologous stem cell transplantation (auto-SCT) revealed a significantly better overall survival (OS) and progression-free survival than the non-transplantation group, which suggested that an up-front auto-SCT may play a role in preventing disease progression, particularly in high-risk DLBCL (3).
The prognostic evaluation of DLBCL is principally based on data at the time of diagnosis. Treatment response during the chemotherapy is mainly dependent on the image tools such as fluorodeoxyglucose-positron emission tomography (PET) scans (4, 5). However, imaging tools available in clinical practice fail to detect remnant minimal disease during or after treatment, and tissue biopsy during treatment is practically impossible for various reasons. Recently, liquid biopsy has emerged as a minimally invasive approach to determine various cancers, and its advantages have been described in many previous studies (6, 7). An ongoing interest exists on whether circulating-tumor DNA (ctDNA) from the blood can reflect a minimal residual disease (MRD), which can influence the decision making during DLBCL treatment (8, 9). Previous studies have demonstrated that the burdens of ctDNA were correlated with the International Prognostic Index (IPI) and total metabolic tumor volume (TMTV) in PET scan (10, 11). Furthermore, ctDNA level was significantly associated with treatment outcomes. Interestingly, ctDNA remained in nonresponding patients, and new mutations that are conceivably resistant to chemotherapy were detected (11).
Currently, screening high-risk patients for treatment failure and predicting clinical outcomes early at diagnosis or during chemotherapy through liquid biopsy has become the core of lymphoma research. Accordingly, blood samples were collected sequentially before and after treatment from patients with newly diagnosed DLBCL who received R-CHOP chemotherapy, and we evaluated the possibility of clinical use of ctDNA as a biomarker to determine up-front auto-SCT after first-line chemotherapy for high-risk patients in practice, prior to clinical trial.
Patients and Methods
Patients and treatment. To explore the dynamics of ctDNA in DLBCL, patients who fulfilled the following criteria were enrolled between 2020 and 2021 in Kyungpook National University Hospital (KNUH): 1) age ≥20 years, 2) pathologically confirmed DLBCL according to the World Health Organization criteria by a hematopathology specialist (12), 3) patients scheduled to receive R-CHOP chemotherapy, and 4) patients without contraindications for auto-SCT. Patients with primary central nervous system lymphoma, primary mediastinal large B-cell lymphoma, or human immunodeficiency virus-associated lymphoma were excluded. Staging was assessed according to the Ann Arbor staging system (13). The IPI was adjusted to classify the at-risk groups (14). Patients received six cycles of R-CHOP chemotherapy every three weeks. According to the International Working Group criteria, treatment response was assessed (15). Patients with Ann Arbor stage III/IV disease at the time of diagnosis and patients with partial response at the end of treatment (EOT) assessment were prepared for auto-SCT. Patients who were to undergo auto-SCT received busulfan, cyclophosphamide, and etoposide conditioning regimen followed by stem cell transplantation. This study was approved by the Institutional Review Board (IRB) of KNUH (IRB No: KNUH 2020-06-003), and all patients provided a written informed consent.
Collection of samples and preparation of DNA. In the current pilot study, the serum samples from DLBCL patients were explored to identify the clinical role of ctDNA as a biomarker to determine up-front auto-SCT in DLBCL. Blood samples were collected serially from the time of diagnosis and after completion of chemotherapy. For an optimal sample collection, at least 10 ml of blood were obtained using ethylenediaminetetraacetic acid tubes (16, 17). Thereafter, plasma was immediately isolated to preserve sufficient ctDNA molecules (17). Each extraction was carried out using the QIAamp Circulating Nucleic Acid Kit (Qiagen, Valencia, CA, USA), based on the manufacturers’ instructions. Initial QC checks of eluted ctDNA were carried out using the Qubit dsDNA HS Assay Kit and the Qubit 2.0 fluorometer (Life Technologies, Carlsbad, CA, USA) and the 2100 Bioanalyzer with High-Sensitivity DNA chips (Agilent Technologies, Santa Clara, CA, USA), based on the manufacturers’ instructions.
Preparation of libraries. To conduct ctDNA genotyping and ctDNA monitoring simultaneously, targeted sequencing by cancer personalized profiling using deep sequencing (CAPP-seq) was used as previously described (18, 19). Libraries were constructed from 40 ng of cfDNA with the SureSelect XT low input protocol (Agilent Technologies). Prepared libraries captured by Custom Panel were developed by the Macrogen (Macrogen, Seoul, Republic of Korea). The libraries were indexed individually and molecular-barcoded for each sample to be sequenced. The quality of libraries was checked with the 2100 Bioanalyzer (Agilent). The size of the products was 200-400 bp. Then, the libraries were quantified using the Qubit dsDNA HS Assay Kit and the Qubit 2.0 fluorometer (Life Technologies). The libraries were sequenced paired-end (2×150 bp) on a NextSeq500 instrument (Illumina, San Diego, CA, USA) using sequencing-by-synthesis chemistry.
Bioinformatic pipeline for targeted genome sequencing. The quality of the raw FASTQ files obtained from the sequencing machine was evaluated using FastQC tool (20). Thereafter, the reads were processed as follows: 1) Aligned to the reference genome of the targeted genes using BWA-mem (21). 2) Sorted, marked duplicates, and recalibrated base quality scores using Picard toolkit and GATK4 (22). Upon generation of the BAM file, GATK4 Mutect2 was utilized for variation calling, filtering, and somatic variant identification (23). Variant annotation and filtering were carried out employing GATK Variantannotator, in conjunction with dbSNP, Clinvar, AnnoVar, and VEP. Variants were represented in the variant call format (VCF). Candidate variants with a variant allele frequency exceeding 0.5% in the sample, were selected. Following the generation of the Mutation Annotation Format (MAF) file, the subsequent analytical steps involved the maftools package within the R statistical software (24). This package facilitated a range of analyses and the creation of informative plots.
ctDNA analysis. We compared the changes of ctDNA level and genotyping by CAPP-seq before treatment and after the EOT. Somatic mutations were identified through paired analysis of either pretreated plasma or germline DNA according to previous studies (10). Statistical analyses were conducted using R statistical software 4.3.1 (the R Foundation for Statistical Computing, Vienna, Austria).
Results
Patient characteristics and clinical outcomes. This study was conducted based on a prospectively collected, consecutive serial blood samples from patients with newly diagnosed DLBCL. The enrolled 10 patients were between the ages of 50 and 60 years at the time of diagnosis, and eight of 10 were men. Five patients were diagnosed with germinal center B-cell type DLBCL. Based on the IPI, two patients were IPI 5, four patients were IPI 4, and one patient was IPI 3, and in our study, these seven patients were classified as high-IPI-risk group. The other three patients were IPI 1 and classified as a low-IPI-risk group. The IPI risk group correlated with TMTV, in terms of disease burden (Figure 1). All patients completed six cycles of R-CHOP chemotherapy, and seven patients achieved complete response (CR). Three patients showed progressive disease, and they all died during salvage chemotherapy because of worsening disease. In the high-IPI-risk group, three patients underwent auto-SCT and are surviving well without disease relapse for more than two years. Stem cell collection failed in patient 3 and the patient experienced relapse. Table I summarizes the patient characteristics and clinical outcomes.
Comparison of the total metabolic tumor volume in high-risk and low risk group (p=0.019).
Patient characteristics.
Circulating-tumor DNA at diagnosis. At the time of diagnosis, serum samples from 10 patients were collected. In the ctDNA analysis, the majority of the genetic abnormalities were missense mutations. Other significant mutations were in-frame insertions and deletions, splice site mutations, and frameshift variants. Single-nucleotide polymorphisms were the most common variant. The 10 most common mutated genes in serum included NOTCH1, KMT2D, PCLO, ARID1A, RET, DDX3X, ITPKB, H1-4, CREBBP, and CARD11. In the current data, the number of ctDNA mutations at the time of diagnosis was not significantly associated with the risk group (Figure 2).
ctDNA analysis during diagnosis. (A) Classification of the mutations and frequency. Missense mutation was the most commonly identified mutation. The most common type was the single-nucleotide polymorphism (B). (C) Types of a single-nucleotide variant. (D) Prevalence and molecular spectrum of somatic mutations discovered in ctDNA. The 10 most frequently mutated genes are summarized in (E).
Comparison of circulating-tumor DNA before and after treatment. Figure 3 shows the genetic alterations before and after treatment. Unfortunately, EOT serum samples of patients 2 and 5 were not properly analyzed, and thus these two samples were excluded from the EOT analysis. For tumor burden, changes in ctDNA mutation numbers did not correlate with changes in PET scan images and treatment response. Nevertheless, all three patients in the low-IPI-risk group showed increased missense mutations after chemotherapy, although evidence of disease progression was absent. In most patients, new mutations appeared in ctDNA after completion of chemotherapy. PIK3CA, BRAF, TGM7, and EZH2 mutations, which were not identified in the pretreatment analysis, were identified in the posttreatment samples.
Overview of the diffuse large B-cell lymphoma genotype identified in ctDNA before treatment (A) and after treatment (B).
Changes in circulating-tumor DNA and clinical course in high-IPI-risk group. We failed to find significant results related to changes in ctDNA mutation numbers and clinical course. Instead, we focused on newly identified ctDNA mutations in the EOT samples in patients with high IPI scores that conceivably marked resistant clones particularly. In the analysis of patient 3, SPEN and IKBKB genes were identified. Patient 3 achieved CR after six cycles of R-CHOP therapy. However, he could not receive auto-SCT due to inappropriate stem cell collection and he manifested with disease relapse. In patient 4, CHD2, KLHL6, TP53, and SF3B1 mutations were newly developed. Patient 7 showed many new mutations including CREBBP, EZH2, PRDM1, BCOP, SF3B1, and STAT6. Both patients 4 and 7 demonstrated resistance to front-line and salvage chemotherapies and died. SPEN and TP53 mutations were newly discovered in patient 1, and TP53, BCOR, IKBKB, and NOTCH2 mutations were noted in patient 6. Both patients 1 and 6 underwent consolidative auto-SCT after achieving CR, and CR was maintained for over two years (Figure 3 and Table I).
Discussion
This study aimed to identify the clinical role of ctDNA related to autologous transplantation in patients with newly diagnosed DLBCL. Measurement of ctDNA can be performed using several methods including polymerase chain reaction (PCR) and next-generation sequencing (NGS)-based technologies. While PCR assays can only target one single or a small number of recurrent somatic variants, NGS-based analysis allows broad noninvasive genotyping instantaneously (8, 25). We conducted NGS-based analysis using CAPP-seq sequentially to track multiple mutations simultaneously, and the results were interpreted using a bioinformatics algorithm that largely eliminates sequencing errors (26, 27). Although the main purpose of the current study was to find a marker for determining auto-SCT through ctDNA analysis, our results suggested that ctDNA should be interpreted by combining the patient’s clinical information such as IPI to determine transplantation. In the current study, ctDNA changes before and after chemotherapy did not significantly affect long-term survival in the low-risk group. In the high-risk group, new mutations associated with poor prognosis surfaced after chemotherapy, and patients who underwent auto-SCT maintained CR without relapse. However, high-risk patients who did not receive auto-SCT experienced disease progression although they showed metabolic CR after completing six cycles of R-CHOP therapy. Although no statistical significance was noted owing to the small number of patients enrolled in this study, consolidative auto-SCT could improve survival in patients who showed new genetic mutations with poor prognosis such as TP53 alterations during chemotherapy.
Several studies have been published on the clinical role of ctDNA in patients with DLBCL. Similar to our study results, Davide et al. analyzed the clinical significance of the changes in ctDNA genotype during R-CHOP therapy in which new mutations appeared in the ctDNA in patients who were primarily refractory to R-CHOP or relapsed after treatment in the longitudinal monitoring of DLBCL genotype using ctDNA (10). Other studies mainly focused on the quantitative level of ctDNA as a tumor volume (11, 26, 28). Kurtz DM et al. measured ctDNA level during treatment in 217 patients with DLBCL, and in most patients, ctDNA was detectable. They showed that the dynamics of ctDNA during therapy correlated with disease response, and molecular response was significantly associated with event-free and OS in the multivariate analysis (26). In another study with 73 patients with DLBCL, posttreatment ctDNA was a sensitive indicator for detecting MRD in high-risk patients, and ctDNA monitoring after treatment can help predict early relapse (11).
CAR-T cell therapy has been emerging as an alternative for relapsed patients and it is showing outstanding survival outcomes compared to the conventional salvage strategies (29). However, access to CAR-T cell therapy remains limited because of the barriers associated with the facility and complexity of the CAR-T administration process. According to recent real-world data, only approximately 25% of patients who required CAR-T cell therapy received the regimen with a median time of six months in the waitlist (29). Although consolidative up-front autologous transplantation is faced with obstacles in conducting appropriate stem cell collection, there have been studies emphasizing the positive role of auto-SCT (3, 30). Preventing disease progression for patients is of utmost importance. Meanwhile, a high tumor burden in ctDNA was associated with a negative prognostic factor in patients who proceeded with CAR-T cell therapy (8, 11). Autologous transplantation followed by CAR-T therapy demonstrated better long-term outcomes in patients with TP53 mutations (31). Still, little is known regarding the association between specific gene mutations that developed during chemotherapy and clinical benefits of stem cell transplantation (32, 33). Autologous transplantation may improve survival in selected patients in a consolidative or salvage setting, and ctDNA monitoring could be used to select patients who would benefit from auto-SCT, even with CAR-T treatment.
Conclusion
ctDNA has the advantages of allowing a noninvasive evaluation of chemotherapy response and prognosis not only during diagnosis but also during treatment. MRD can be detected through dynamic monitoring of ctDNA, and overt relapse can be prevented through additional therapeutic options. As a pilot study, we conducted NGS-based ctDNA analysis to assist in the decision making for auto-SCT, prior to a larger-sized clinical trial. Our results suggested that ctDNA monitoring was meaningful in high-risk patients. Moreover, ctDNA and well-known prognostic factors should be considered in the decision making for auto-SCT. If a new genetic mutation with a negative prognosis would emerge during treatment in ctDNA, high-risk patients should consider auto-SCT.
Acknowledgements
This work was supported by Biomedical Research Institute grant, Kyungpook National University Hospital (2020).
Footnotes
Authors’ Contributions
J.K., J.H.M., H.S.H., and D.W.B. designed the study, contributed to patient enrollment and data collection. J.K., J.H.M., and D.W.B. drafted the manuscript. L.T.M., L.D., and N.H.D.T. performed genetic analysis, and bioinformatic analysis. H.J.C., S.K.S., J.G.K, S.Y.J., J.Y.H., and J.J.Y. contributed to the study design and data interpretation and revised the manuscript. All Authors provided final approval of the version to be submitted.
Conflicts of Interest
The Authors declare that they have no conflicts of interest in relation to this study.
- Received August 31, 2023.
- Revision received October 18, 2023.
- Accepted October 19, 2023.
- Copyright © 2024 The Author(s). Published by the International Institute of Anticancer Research.
This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY-NC-ND) 4.0 international license (https://creativecommons.org/licenses/by-nc-nd/4.0).
References
In this issue
Jump to section
Related Articles
Cited By...
- No citing articles found.