Bispecific antibodies in action: the reality of engagement Free

In this issue of Blood, Brooks et al¹ report the real-world experience (RWE) and clinical outcomes of patients receiving epcoritamab or glofitamab for relapsed large B-cell lymphoma (DLBCL). This is one of the first comprehensive analyses of RWE with bispecific antibodies (bsAbs) in DLBCL.

Epcoritamab and glofitamab are both bispecific CD20×CD3 T-cell engagers approved for the treatment of relapsed DLBCL.^2,3 BsAbs represent a transformative class of therapies for B-cell non-Hodgkin lymphomas, harnessing the cytotoxic potential of T cells through off-the-shelf agents.^4,5 The lymphoma field has closely tracked long-term trial data while awaiting RWE. Given that trials often enroll fitter patients, RWE is essential to assess adverse events and management variability in broader populations. Although based on academic centers, this report offers important insights into bsAb efficacy, safety, predictive factors, and logistical challenges.

The authors report on 245 patients who received epcoritamab (n = 156) or glofitamab (n = 89) as standard of care. As expected, this cohort comprised patients with high-risk features: a median of 3 prior lines of therapy, 60% with prior chimeric antigen receptor T-cell (CAR-T) therapy, 65% with an International Prognostic Index ≥2 at treatment initiation, one-quarter with bulky disease (>7.5 cm), and more than half with primary refractory disease. Notably, 70% of patients would have been ineligible for the registrational trials of these agents. Several findings from this report are important to highlight:

With a median follow-up of 6.3 months, 70% of patients discontinued therapy due to disease progression, and 20% due to adverse events. Median progression-free survival (PFS) was 2.5 months, and overall survival (OS) was 7.9 months and notably, the median number of full-dose antibody doses was only 4. Consistent with trial data, 70% to 80% of patients who achieved a complete response (CR) remained in response at the current follow-up. Although the overall response rate (51%) aligns with clinical trials, the CR rate (25%) appears lower. One explanation may be lower utilization of imaging in routine practice, leading to under-documentation of CRs. The shorter PFS and OS likely reflect the high-risk nature of this population with 70% of patients being ineligible for registrational trials. Although baseline characteristics offer some comparison, much remains unclear without access to true trial-screening denominators. Trial ineligibility was associated with shorter PFS and OS, as was receiving bsAbs within 90 days of prior CAR-T. In the multivariable model, primary refractory disease and low ECOG Performance Status Scale predicted a shorter PFS. Although median PFS appears shorter than in registrational trials, patients achieving CR derive durable benefit, consistent with trial experience. PFS may be more useful as a pretreatment estimate (but unfortunately not routinely reported for the entire population of the pivotal trials), whereas duration of response, particularly CR duration, can guide ongoing treatment decisions if assessments are done early after treatment initiation.

A practical and important finding is the association of CD20 expression with clinical outcomes. Among patients with available CD20 data, those with CD20-negative biopsies had significantly shorter PFS and OS. This highlights the importance of repeat biopsy when feasible and considering alternative treatments in CD20-negative disease confirmed by both immunohistochemistry and flow cytometry. Additionally, a potential mechanism of bsAb failure was suggested: 86% (30/35) of patients with progressive disease had undetectable CD20 expression and in 85% of those with paired biopsies, CD20 antigen loss was confirmed. These findings have implications for treatment sequencing, especially as CD20-targeting agents are used across different lines and bsAbs are being studied in frontline settings.

The overall incidence of cytokine release syndrome (CRS) was consistent with trials but incidence of grade ≥3 CRS was higher in epcoritamab-treated patients, both compared to pivotal trials and to the glofitamab cohort. Notably, 3 fatal CRS events and 1 fatal neurotoxicity occurred with epcoritamab. Whether this reflects a true signal or is related to older age, larger cohort, earlier use, longer follow-up of the epcoritamab cohort, or other unmeasured differences, remains unknown. We will continue to closely monitor future studies for evolving safety signals for both drugs. Differences in institutional toxicity management, which are not fully captured in retrospective data sets, may contribute to outcome variation underscoring the need for harmonized toxicity management. The published consensus guidelines for bsAb-related toxicity management in lymphoma represent one such effort.⁶ Standardized approaches not only benefit patient care but also enhance the interpretability of nonprospective data. Infections remain a concern and require longer-term follow-up especially given the differences in the approved label of these drugs regarding time-limited vs continuous treatment.

Although glofitamab showed numerically better outcomes—including reduced mortality in multivariate analysis—drug vs drug comparisons, should be interpreted cautiously due to imbalanced baseline characteristics and short follow-up. As noted by the authors, residual confounding limits these conclusions. Indirect comparison methods (eg, matching-adjusted indirect comparisons) using patient-level data from pivotal trials, ideally conducted by independent investigators, along with longer follow-up and nonoverlapping RWE reports, will be essential because a head-to-head trial is unlikely.

We commend the authors for this timely and comprehensive RWE report. It highlights several practice-informing insights. These findings support our emphasis on biopsy and CD20-expression reassessment. Although bsAbs remain a promising standard of care option post–CAR-T, their efficacy appears diminished when used early after CAR-T failure. Clinical trial participation remains the preferred option for patients with relapsed DLBCL, and this study further supports prioritizing novel therapies in the post–CAR-T setting. Although outcomes appear somewhat inferior to registrational trials, larger data sets and longer follow-up are needed to confirm this. We urge continued collaboration and inclusion of community practices in future reports to further enhance our understanding of bsAbs in real-world care.

Conflict-of-interest disclosure: M.S. reports consulting and serving on advisory boards, steering committees, or data safety monitoring committees for AbbVie, Genentech, AstraZeneca, Genmab, Janssen, BeOne Medicines, Bristol Myers Squibb (BMS), Morphosys/Incyte, Kite Pharma, Eli Lilly, Fate Therapeutics, Jannsen, Pfizer, Nurix, Pierre Fabre, and Merck; research funding from Mustang Bio, Genentech, AbbVie, BeOne Medicines, AstraZeneca, Genmab, Sana Biotherapeutics, and Morphosys/Incyte; employment at BMS (spouse); and stock options at Koi Biotherapeutics. A.K.G. reports research funding from Merck, I-Mab Biopharma, IGM Biosciences, Takeda, Gilead, AstraZeneca, Agios, Janssen, BMS, SeaGen, Teva, Genmab, BeiGene, Pfizer, and Umoja; consultancy/honoraria from Incyte, Kite, Morphosys/Incyte, ADC Therapeutics, Acrotech, Merck, Karyopharm, Servier, BeiGene, Cellectar, Janssen, Compliment, SeaGen, Epizyme, I-Mab Biopharma, Gilead, Genentech, Lilly, Caribou, Fresenius-Kabi, SciTech, Sana, and Compliment; and stock options at Compliment Corporation and SciTech.