Bcl-2 inhibition in DLBCL: “the times they are a-changing”?

In this issue of Blood, Morschhauser et al report the results of the phase 2 CAVALLI study in 206 patients with International Prognostic Index (IPI) 2-5 diffuse large B-cell lymphoma (DLBCL).¹  The authors demonstrate that the addition of venetoclax (a selective and potent B-cell lymphoma-2 [Bcl-2] inhibitor) to rituximab plus cyclophosphamide, doxorubicin, vincristine, and prednisone (R-CHOP) has the potential to improve the outcome of patients with Bcl-2 overexpressed DLBCL despite a higher incidence of adverse effects.

This study is particularly exciting because vaulting over the wall of R-CHOP results in DLBCL has been the dream of many researchers for years. The addition of rituximab improved the efficacy of the CHOP regimen in DLBCL with 55% to 60% of patients cured and is still the standard of care.²  However, treatment fails for 40% to 45% of patients, and this prompted many attempts to improve the efficacy of first-line treatment for DLBCL. Chemotherapy regimens were modified to use infusional, dose-adjusted drugs or consolidation after R-CHOP with high-dose chemotherapy and autologous stem cell transplantation. Despite promising results in phase 2 studies, randomized phase 3 trials failed to show any overall survival benefit over R-CHOP.^3,4 

The substitution of rituximab with obinutuzumab was assessed in the international, randomized phase 3 GOYA-BO21005 (hereafter GOYA) study in 1418 DLBCL patients. Although obinutuzumab was more active than rituximab in follicular lymphoma and chronic lymphocytic leukemia (CLL), this did not happen in DLBCL. The GOYA study did not show an advantage over R-CHOP.⁵  Nevertheless, the GOYA study has been considered a benchmark for other DLBCL studies, and indeed Morschhauser et al compare their results with those from the GOYA study.

Overall, these data emphasize that chemoimmunotherapy resistance cannot be easily overcome simply by modifying chemotherapy regimens or by substituting a new anti-CD20 antibody. This may be because of the heterogeneity of DLBCL, which has 3 major biologically distinct pathophysiologic entities based on cell-of-origin (COO): germinal center B-cell–like (GCB), activated B-cell–like (ABC), and unclassified, with the ABC subgroup having the worst prognosis when treated with R-CHOP.⁶  These data prompted a change in strategy for treating DLBCL and a shift to targeting the group with the worst prognosis (ie, the ABC group) with appropriate drugs. Two drugs, ibrutinib and lenalidomide, showed activity in non-GCB/ABC DLBCL subgroups in early phase 1/2 studies with positive clinical results. Randomized phase 3 studies tested the addition of ibrutinib or lenalidomide to R-CHOP vs placebo/R-CHOP (PHOENIX and ROBUST studies) in non-GCB/ABC IPI 2-5 DLBCL. Unfortunately, there were no differences in 2-year progression-free survival (PFS) rates between the arms.^7,8 

The additional biological complexity of DLBCL explains these results. Indeed, a recent molecular classification of DLBCL identified 5 distinct molecular subsets and reported wider heterogeneity within the COO subgroup whose patients had very different prognoses.⁹  However, further subgroup analyses of the PHOENIX and ROBUST studies suggested some benefits in patients with specific clinical or biological characteristics. Patients who were young and/or those with overexpression of Bcl-2 and Myc had a more favorable outcome with ibrutinib–R-CHOP in the PHOENIX trial, and clinically high-risk patients (ie, patients with IPI 3-5) had a more favorable outcome with lenalidomide–R-CHOP in the Robust study. These data open a window to a possible successful strategy to improve R-CHOP results if patients are better selected for new treatments.

The Morschhauser et al study addressed a specific DLBCL subgroup with poor prognosis: patients with overexpressed Bcl-2. Indeed, overexpression of Bcl-2 (an anti-apoptotic protein) identifies DLBCL patients with inferior outcomes who are resistant to R-CHOP. Moreover, patients with concurrent overexpression of Bcl-2 and Myc proteins (known as double expressors [DELs]) have a poor prognosis when treated with R-CHOP. Overall, these patients represent 25% to 45% of all cases of DLBCL. Moreover, data from GOYA and other studies suggest that Bcl-2 overexpression is the main driver of poor prognosis in DELs, and counteracting Bcl-2 may be a way to improve their outcome.

Morschhauser et al targeted these poor-prognosis subgroups. The authors added venetoclax to R-CHOP in 206 patients with advanced-stage DLBCL. That study was nonrandomized, but the authors made an effort to compare the results with matched R-CHOP controls from the GOYA study (564 patients). Positron emission tomography complete response (PET-CR) rates, the primary end point, were superimposable on those of the GOYA cohort (69% and 63%). Interestingly, the addition of venetoclax was able to significantly increase 2-year PFS rate (the secondary end point) compared with GOYA results, namely in Bcl-2–positive patients (78% vs 62%; hazard ratio, 0.55; 95% confidence interval, 0.34-0.89), suggesting that the addition of venetoclax may induce a much deeper response, as occurred in CLL, leading to a reduction in the rate of progression (see table).

Although treatment was manageable, grade 3 to 4 adverse event rates were higher in CAVALLI compared with GOYA for neutropenia (68% vs 39%) and infections (23% vs 16%), which shows the need for more stringent, well-defined prophylactic measures when a novel drug is added to R-CHOP, a policy that was not clearly adopted in the Morschhauser et al study. However, the costs of adverse effects of treatment, if manageable, are worth paying when they are balanced with a long-term benefit, as happened in the Morschhauser et al study.

Limitations of this study include the nonrandomized design, the lack of PET-CR differences (ie, the primary end point) compared with the GOYA study, and the unbalanced comparison between the CAVALLI and GOYA studies, the first conducted in 50 sites and the second in 207 sites in a larger worldwide setting with different level of expertise. Nevertheless, the difference in PFS for Bcl-2–positive patients is wide enough and is also a good signal for conducting further studies with this combination.

The emerging molecular complexity of DLBCL could lead to novel therapeutic strategies with the identification of different molecular pathways targeted by appropriate drugs, possibly added in multiple combinations to chemoimmunotherapy to overcome refractoriness to chemotherapy.⁹  An intriguing example is the SMART-START study that combines rituximab, lenalidomide, and ibrutinib as a chemotherapy-free approach followed by CHOP or etoposide, prednisone, vincristine, cyclophosphamide, and doxorubicin (EPOCH) in 60 patients with non-GCB DLBCL. The preliminary 92% rate for PET-CR is encouraging and merits further exploration.¹⁰ 

“The times they are a-changin’” as Bob Dylan used to sing, may also be true for the treatment of at least a subset of DLBCL patients, but the data presented here underline that our work still has a long way to go.