Impact of Time-to-relapse on Biology, Evolutionary Dynamics, and Chemosensitivity in Diffuse Large B-cell Lymphoma

Diffuse large B-cell lymphoma (DLBCL) is the most common form of lymphoma diagnosed in the United States and is potentially curable in at least 60% of patients who receive standard frontline R-CHOP chemo-immunotherapy (i.e., rituximab plus cyclophosphamide, doxorubicin, vincristine, and prednisone).¹  In patients with DLBCL who experience relapsed or refractory (r/r) disease following R-CHOP treatment, survival after second-line chemotherapy and consolidative autologous stem cell transplantation (ASCT) is strongly dependent on the time-to-relapse. For example, in the pivotal CORAL study, patients with r/r DLBCL had significantly inferior three-year event-free survival with second-line chemotherapy and consolidative ASCT if they had relapsed within one year of initial diagnosis (20% vs. 45%).²  While time-to-relapse is therefore a clinically significant milestone, the biological basis of the divergent outcomes among patients with early-versus late-relapsing DLBCL following second-line chemotherapy remains undefined.

In the referenced article,³  Laura Hilton, PhD, and colleagues confirmed significant survival differences based on time-to-relapse in a population-based cohort of 221 patients undergoing second-line chemotherapy and consolidative ASCT. The authors analyzed paired diagnostic and relapsed biopsy samples in patients with refractory (relapse <9 months from diagnosis), early-relapsing (9-24 months from diagnosis), and late-relapsing (>24 months from diagnosis) DLBCL to compare the genetic and biologic evolutionary dynamics driving relapse in these different subgroups. In the refractory and early-relapse groups, the authors identified a high degree of concordance in mutational variants between diagnostic and relapse samples, indicating a pattern of tumor evolution reflective of intrinsic chemoresistance. In contrast, a branching pattern of evolution was observed in the late-relapse group, with only a limited set of shared mutational variants between the specimens obtained at diagnosis and relapse. These shared variants are considered to derive from a putative common precursor cell (CPC), and the authors hypothesize that for late-relapsing DLBCL, while the initial chemotherapy effectively eradicates the aggressive clone, the underlying CPC persists and eventually gives rise to a genetically distinct secondary transformation. Therefore, because late-relapsing DLBCLs are often genetically distinct from their diagnostic counterparts, they can be viewed as “chemotherapy-naïve”, which likely explains their more favorable response to second-line salvage chemotherapy.

Collectively, these findings are timely given the evolving landscape of second-line treatment for DLBCL. For one, they provide confirmation of the dismal outcomes of chemotherapy-based regimens in the second-line setting for early-relapsing DLBCL and provide a scientific rationale for the intrinsic chemoresistance of these tumors. Moreover, with the recent overall survival advantage identified with CD19-directed chimeric antigen receptor (CAR) T-cell therapy versus ASCT in the second-line treatment of early-relapsing DLBCL,⁴  these data support the paradigm shift towards the adoption of cellular immunotherapies for the treatment of these patients. The findings also suggest that, for patients who require bridging treatment during the CAR T-cell manufacturing period, non-chemotherapy-based immunotherapies such as bispecific antibodies should be investigated since many refractory DLBCLs are intrinsically chemoresistant. In contrast, for patients with late-relapsing DLBCL, the two-year progression-free survival after second-line chemotherapy and ASCT was more favorable at 60% in the report by Dr. Hilton and colleagues, which highlights the need for additional comparative studies before CAR T-cell therapy is deemed superior to ASCT for the second-line treatment of late-relapsing DLBCL.

The results from this study also shed new light on the biological stability of DLBCL at the time of relapse. Even for late-relapsing DLBCL, in which significant genetic divergence occurs, canonical DLBCL classification systems such as cell-of-origin⁵  and LympGen⁶  are concordant between diagnostic and relapsed samples in approximately 90% of cases due to convergent evolution. R-CHOP therefore applies minimal pressure to the biologic evolution of DLBCL, which may have important implications for clinical decision-making.^7,8  For example, if at diagnosis a DLBCL expresses an activated B-cell-like (ABC) cell-of-origin or is classified by LymphGen as belonging to the MCD subtype, there is a high probability these classifications will persist at relapse, which may inform response to treatments such as Bruton tyrosine kinase inhibitors.^7,8 

Finally, this study inspires the broader application of evolutionary dynamics in lymphoma. Future studies should evaluate the evolutionary dynamics of DLBCL after novel treatments (e.g., polatuzumab vedotin and CAR T-cell therapy) and integrate evolutionary dynamics with other factors such as the tumor microenvironment. Moreover, the idea of comparative evolutionary dynamics based on time-to-relapse should be applied to other lymphomas, as time-to-relapse after frontline chemotherapy is also strongly associated with prognosis and second-line chemosensitivity in patients with follicular lymphoma (FL).⁹  A deeper understanding of the evolutionary dynamics of FL after frontline chemotherapy may therefore reveal a scientific basis for incorporating non-chemotherapy-based approaches like CAR-T for second-line treatment of early-relapsing FL.