Although the treatment outcomes of diffuse large B-cell lymphoma (DLBCL) have improved with the development of new drugs, relapse is frequent and associated with dismal outcomes (1, 2). Bone marrow (BM) involvement is classified as extranodal and stage 4, which increases the international prognostic index (IPI) and is directly linked to shorter survival (3–5). The reported incidence of BM involvement is 11%-36% and the classic definition of BM involvement is abnormal lymphoma cells in BM aspirates or biopsies (3, 4, 6). However, minimal BM involvement of malignant lymphoma cells often generates false negative results because a histological diagnosis is very difficult in the absence of significant morphological changes (7–9). Bone marrow involvement can be diagnosed using 18F-FDG PET, but only within a limited range, and diagnostic rates vary depending on the lymphoma subtype (10–13). These problems have been addressed using the polymerase chain reaction (PCR) to detect immunoglobulin gene rearrangement (IgR) in BM samples because B-cell non-Hodgkin lymphoma (NHL) undergoes clonal IgR (9, 14). Clonal immunoglobulin heavy chain (IGH) and kappa chain (IGK) gene rearrangement could help the diagnostic process when histological findings are inconclusive. Moreover, gene rearrangement can be a helpful indicator during follow-up, as well as for diagnoses (15, 16). Patients with negative histological BM can be classified based on whether they test positive for IgR and negative for BM histology which indicates a more accurately determined advanced stage and a poorer prognosis (14, 17). Immunoglobulin gene rearrangement has mostly been detected using PCR; however, next-generation sequencing (NGS) has also been recently used (18). Detecting BM involvement in patients newly diagnosed with DLBCL indicates poor prognosis; to that end, IgR tests have been applied in a few studies to detect BM involvement in patients DLBCL treated with R-CHOP (rituximab, cyclophosphamide, doxorubicin, vincristine, and prednisolone) chemotherapy (14, 19–21).

Patients newly diagnosed with DLBCL accompanied by negative BM histology and poor outcomes of current standard treatment should be tested for IgR to precisely diagnose BM involvement. New treatment approaches should also be applied such as high-intensity chemotherapy to overcome BM involvement as a poor prognostic factor (5). Here, we investigated the clinical characteristics and treatment outcomes of upfront consolidative ASCT as part of a high-intensity chemotherapeutic regimen in patients with DLBCL and BM involvement determined by histological or molecular biological methods.

This study enrolled patients from nine institutions in Korea who were newly diagnosed with DLBCL based on the World Health Organization classification (22) and histological BM involvement between 2010 and 2019. The study was conducted according to the guidelines of the Declaration of Helsinki, and approved by the Institutional Review Board of Severance Hospital (4-2019-0579 Aug 5, 2019) and each institution.

The control group comprised patients from Severance Hospital who were newly diagnosed with DLBCL and were tested for IgR regardless of BM involvement status within the same period. The exclusion criteria were disease transformation from indolent follicular lymphoma, primary central nervous system lymphoma, cutaneous DLBCL, primary mediastinal B-cell lymphoma, and human immunodeficiency virus-associated DLBCL. All patients were administered with R-CHOP as first-line chemotherapy. Upfront consolidative ASCT was considered for patients with Ann Arbor stages III or IV and elevated lactic dehydrogenase (LDH) levels who achieved complete (CR) or partial (PR) remission after R-CHOP chemotherapy. The international prognostic index (IPI) score was calculated as described (23). Responses were assessed based on the Cheson criteria (24).

The present study findings revealed that the clinical characteristics of patients with DLBCL and positive IgR in BM samples (besides those with traditional histological BM involvement) who received R-CHOP chemotherapy, resembled those of patients with advanced-stage lymphoma. Moreover, these patients did not respond well to R-CHOP first-line treatment and had poor OS and PFS. Therefore, tests for IgR should be applied to precisely predict the prognosis of patients with negative BM histology.

BM involvement of DLBCL cells showed an unfavorable gene signature, which was related to tumor cell proliferation, migration, and immune escape. These could explain high-risk clinical features and poor prognosis (5). However, differentiating the histological diagnosis of BM involvement of malignant lymphoma cells can be challenging particularly in patients with small amount of lymphoma cells. The IgR test could be helpful under such circumstances. The IgR results were positive in 13%–16% of patients with DLBCL who were diagnosed with histologically normal BM and these patients did not survive for long (14, 20). Here, we found positive IgR in 17.5% of patients with negative histological BM involvement, which was similar to previous findings. Without IgR tests, these patients would have been classified as having no BM involvement and the disease stage would have been lowered. Therefore, routine IgR tests of BM samples should be recommended to evaluate the molecular BM involvement of DLBCL cells.

A higher proportion of patients with histological BM involvement had a more advanced disease stage, more frequent extranodal involvement, and more elevated LDH than patients with positive IgR and negative BM histology. Patients with histological BM involvement were classified as having stage 4 disease or a high IPI score at the time of diagnosis. However, patients with positive IgR might not be classified as having an advanced disease stage and might have been down-staged because histological BM involvement was not found. Therefore, the clinical characteristics of patients with histological BM involvement differed from those with only positive IgR with negative BM histology. Nevertheless, we found that the differences in OS and PFS between patients with histological BM involvement and those with positive IgR and negative BM histology were not significant. In addition, the multivariate analysis identified positive IgR as an important prognostic factor associated with poor OS and PFS in patients who were not treated with upfront ASCT.

The most useful tool for assessing clonality in patients with NHL until recently was BIOMED-2 PCR assays. These had been widely used as they were standardized and deemed suitable for technically routine test environments (26). However, PCR is limited by being unsuitable for samples with poor DNA quality, such as formalin-fixed paraffin-embedded (FFPE) samples, which could produce false negative results (15, 18). However, small amplicons and FFPE samples can be analyzed using NGS (18). We compared the ability of PCR and NGS to detect clonality and found no significant differences.

Although histological BM involvement is considered a poor prognostic factor, a standardized treatment approach has not yet been established. Furthermore, patients with positive IgR have not been studied. High-intensity chemotherapy, such as fractionated cyclophosphamide, vincristine, doxorubicin, and dexamethasone alternating with high-dose methotrexate and cytarabine (rituximab-hyper-CVAD/MA) or dose-adjusted etoposide, prednisone, vincristine, cyclophosphamide, doxorubicin, and rituximab (EPOCH-R) might overcome poor prognoses, and treatment outcomes are better than those of R-CHOP in high-risk patients with DLCBL (5). Consolidative upfront ASCT might also be considered as a different approach to high-intensity chemotherapy for DLBCL because it can eradicate PCR-detectable NHL cells and consequently reduce recurrence (27). Upfront ASCT in the rituximab era improves PFS in high-risk patients with DLBCL (28, 29). Based on this, we investigated whether upfront ASCT could mitigate the poor prognosis of patients with DLBCL and BM involvement. According to Korean reimbursement guidelines, consolidative upfront ASCT in clinical practice can be recommended for patients with elevated LDH and stage III/IV DLBCL at the time of diagnosis who respond to front-line R-CHOP chemotherapy. However, the present study was retrospective, and as a result, patients with good treatment response and performance might have been selected to receive upfront ASCT. Accordingly, the patients were divided into groups with and without upfront ASCT when analyzing the prognostic factors associated with survival to minimize bias associated with the intensity of treatment. Analysis of all enrolled patients showed that survival was shorter for patients with histological BM involvement or positive IgR with negative BM histology than for those without histological BM involvement and negative IgR. The results of the multivariate analysis showed that patients with poor performance status, elevated LDH, or positive IgR who did not receive upfront ASCT tended to have poor OS and PFS. These results indicated that upfront ASCT plays an important role in overcoming a poor prognosis due to histological BM involvement or positive IgR with negative BM histology. The routine application of upfront ASCT consolidation after R-CHOP is not considered standard care in all countries. However, we suggest that upfront ASCT for high-risk patients with DLBCL and BM involvement should be considered at least in those countries with access to novel target agents.

This study had the following limitations. First, this was a retrospective study and not a prospective randomized study. To overcome this limitation, we registered as many patients as possible from nine institutions in Korea. Another limitation was the absence of regular follow-up data for IgR tests, although they were applied at the time of diagnosis. Based on the concept of minimal residual disease, follow-up tests for IgR are underway and will be examined through further follow-up studies. Although IgR tests could not discriminate infiltration by a high- or low-grade component, the poor prognostic impact of IgR positivity for patients with DLBCL nevertheless generated meaningful information. Despite these limitations, our findings were meaningful insofar as we used PCR and NGS to investigate the role of IgR, in addition to histological BM involvement, in a large cohort of patients with DLBCL and analyzed their clinical characteristics and treatment outcomes.

In conclusion, tests to detect IgR BM allowed a more detailed classification of the prognosis of patients who were negative for histological BM involvement. Patients who did not receive upfront ASCT could not overcome the poor prognosis associated with BM involvement. Our results suggested that ASCT could mitigate the poor prognosis of not only patients with histological BM involvement but also those with positive IgR and negative BM histology. Accordingly, these findings require validation through future prospective studies.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

The studies involving humans were approved by Ethical Review Committee of Severance Hospital. The studies were conducted in accordance with the local legislation and institutional requirements. The ethics committee/institutional review board waived the requirement of written informed consent for participation from the participants or the participants’ legal guardians/next of kin because this is a retrospective cohort study.

YK: Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, Methodology, Resources, Software, Validation, Visualization, Writing – original draft, Writing – review & editing. HS: Data curation, Resources, Writing – original draft, Writing – review & editing. H-YY: Data curation, Resources, Writing – original draft, Writing – review & editing. D-HY: Data curation, Resources, Writing – original draft, Writing – review & editing. YP: Data curation, Resources, Writing – original draft, Writing – review & editing. JL: Data curation, Resources, Writing – original draft, Writing – review & editing. W-SL: Data curation, Resources, Writing – original draft, Writing – review & editing. YD: Data curation, Resources, Writing – original draft, Writing – review & editing. Y-CM: Data curation, Resources, Writing – original draft, Writing – review & editing. DK: Data curation, Resources, Writing – original draft, Writing – review & editing. JK: Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, Methodology, Project administration, Resources, Software, Supervision, Validation, Visualization, Writing – original draft, Writing – review & editing.