Non-invasive minimal residual disease (MRD) analysis in diffuse large B-cell lymphoma (DLBCL)

Introduction: diffuse large B-cell lymphoma (DLBCL) is a heterogeneous disease with genetic diversity and variable outcomes. It arises from a mature clonal B cell population that exhibits clonal immunoglobulin (IG) gene rearrangements. The current standard for monitoring DLBCL response to therapy relies on the estimation of tumor reduction by CT or PET/CT scan. Clonality analysis using a next-generation sequencing (NGS)-based approach is a powerful tool not only to detect clonal B cells in lymph node tissue but also to track minimal residual disease (MRD). In this context, one major obstacle is the absence of circulating cells. cell-free DNA (cfDNA) might be the best analyte for MRD assessment, as it is easily accessible from peripheral blood (PB). MRD analysis is of great interest since it could help to identify patients at high risk of recurrence and to guide treatment decisions. Aims: to test IG heavy (IGH) and light (IGK) chain rearrangements as target of clonality using NGS on formalin-fixed paraffin embedded (FFPE) lymph node biopsies and on cfDNA extracted from PB, in newly diagnosed DLBCL patients; to explore if tracking IG clones by NGS on cfDNA samples during/after treatment can be a valid non-invasive way to study MRD; to study the correlation with radiologic assessment of early and final response and with clinical variables. Results and methods: NGS-based clonality testing was performed using the LymphoTrack assay (Invivoscribe Inc, San Diego, CA) in 53 patients provided with DNA from the tumor biopsies and with cfDNA extracted from PB samples. Tumor-specific clonotypes were detected in 88.5% of 52 evaluable FFPE samples and 80.5% of 46 cfDNA samples. Clonality identification rate on cfDNA at diagnosis correlated with disease stage and a trend for association with extra-nodal disease was observed. MRD was performed by tracing the disease-specific clonotypes in plasma samples collected at interim, at the end of treatment (EOT), and at follow-up. MRD at interim was positive in 8 of the 26 evaluated cases and 67% of them subsequently relapsed; it was negative in 18 patients and 12% of them relapsed (p<0.0001). MRD at interim allowed to better categorize the 15 cases with partial response assessed by CT scans: 8 patients were MRD negative and two relapsed, 7 were MRD positive and 5 relapsed (p=0.08). MRD at EOT proved promising for the identification of patients at high risk of relapse, as the best PFS was observed in patients who achieved MRD negativity at this time point (p=0.001). Indeed, MRD was positive in 7 cases and all relapsed; of these 7 patients, only one had a positive PET/CT at EOT. Nine patients were MRD negative and never experienced relapse; 2/9 showed a PET/CT false-positive result. Finally, we showed that cfDNA was detectable in the plasma before clinical progression in most cases. Conclusions: NGS-based assays are suitable for IG-based biomarker identification and MRD analysis in plasma samples of DLBCL patients, since this analysis identifies patients with a high risk of relapse. MRD evaluation both at interim and at EOT allows to better stratify DLBCL patients’ outcome. The role of cfDNA for MRD detection and its impact on prognosis has to be further explored in larger series of DLBCL patients, in order to validate these results and to explore the possibilities of MRD-adapted therapy regimens that can ultimately improve patients’ care.