Risk factors for failure to complete step-up-dosing of bispecific antibody therapy for B-cell lymphomas

Abstract Introduction: Bispecific CD20-directed/T cell-engaging antibodies (BsAbs) are FDA approved for treatment (tx) of patients (pts) with relapsed/refractory (r/r) large B-cell lymphomas (LBCL) and follicular lymphoma (FL). For LBCL, pivotal trials report durable remissions for those pts who achieve complete response (CR); however, progression-free survival (PFS) is short and real-world evidence suggests inferior CR rate, PFS, and overall survival (OS) compared to pivotal trials (Brooks et al. Blood 2025). We performed a retrospective, single-center analysis to better characterize the timing and clinical characteristics predictive of BsAb failure. Methods: This retrospective analysis included pts with histologically confirmed r/r LBCL or FL who received at least one dose of a commercially available BsAb at the University of Pennsylvania between February 2023 and January 2025. Results: As of January 13, 2025, we identified 109 pts who received BsAb meeting study criteria; 68 (62%) pts had LBCL. Among LBCL pts, 31 (45.6%) received glofitamab (glofit), 31 (45.6%) epcoritamab (epco), and 6 (8.8%) mosunetuzumab (mosun). Forty-one pts had FL of whom 40 pts (97.6%) received mosun and 1 pt (2.4%) epco. Fourteen pts (12.8%) did not complete BsAb step-up to dosing (SUD). Among SUD failures, 13/14 pts (92.9%) had LBCL and 1 pt (7.1%) FL; 7 pts received epco, 6 pts glofit and 1 pt mosun. Of 55 LBCL pts who completed SUD, the median age was 68 years (IQR 60.5-75), 36 of 55 pts (65%) had Ann Arbor stage III-IV (advanced stage) disease and median prior lines of tx (LOT) was 3 (range 1-10). The median number of cycles completed for epco was 4 (range 0.75–20), glofit 8 (range 1-17), and mosun 4 (range 2-8). Of 31 pts treated with epco, 4 (12.9%) continue on tx and 4 (12.9%) used epco as a bridge to CAR-T; other reasons for epco discontinuation included progressive disease (PD; 9 pts, 29.0%), infections (4 pts,12.9%), Grade 5 CRS/ICANS (1 pt, 3.2%), and secondary malignancy (1 pt, 3.2%). One pt was lost to follow-up (FU). Of 31 pts who received glofit, one (3.2%) continues tx; 8 pts (25.8%) completed prescribed course; 12 pts (38.7%) discontinued due to PD, 2 pts (6.5%) bridged to CAR-T, and 2 pts (6.5%) after CR prior to completing the prescribed course. Of 6 transformed FL pts treated with mosun, 3 pts (50%) discontinued after CR, 2 pts (33.3%) bridged to CAR-T, and 1 pt (16.7%) had PD. Of the 13 pts with LBCL who did not complete SUD, median age was 73 years (IQR: 67-80) and 12 pts (92%) were advanced stage. Median prior LOT was 2 (range 1-8). 12 of 13 pts (92.3%) are deceased at data cut, with a median time to death 17 days (range 6-39) after initiating BsAb. Causes of death included PD (n=8), multi-system organ failure (n=2), infection (n=1) and CRS/ICANS (n=1). Generalized linear model univariate analysis of LBCL pts (n=68) indicated the following variables were significantly associated with SUD failure: LDH > 2 x upper limit of normal [ULN] (p<0.001); high/high-intermediate IPI score and longest recorded lymph node diameter as a continuous variable (each p<0.01); age (continuous variable), elevated ECOG PS score (> 2) and stage III/IV disease (each p<0.1). In a multiple regression model, pts with LDH > 2x ULN were at higher risk of not completing SUD (p < 0.05). Of 13 LBCL pts who did not complete SUD, 11 pts (84.6%) had LDH > 2 x ULN, while 8 of 53 pts who completed SUD had LDH > 2 x ULN (15.1%; 2/55 pts had no LDH available). No FL pt had LDH > 2 x ULN. For the entire LBCL cohort, median FU was 8.7 months (mo); median PFS was 9.4 mo (95% CI, 5.9-not reached [NR]) with median OS NR (95% CI, 10.5-NR). The overall response rate (ORR) was 57.3%; CR 32.3%, stable disease (SD) 7.4%, and PD 25% (7 pts [10.3%] did not have response assessment). For the entire FL cohort, median FU was 16.2 mo, median PFS and OS NR (95% CI, 8.0-NR and 95% CI, 26.7 mo-NR, respectively). The ORR was 80.5%; CR 56.1%, PR 24.2%, SD 2.4%, and PD 17.1% Conclusion: Our findings suggest a substantial percentage (20%) of pts with LBCL do not make it to the full dose of BsAb therapy. Risk factors, such as elevated LDH and LBCL histology, characterize a subset of pts, at high-risk for not completing SUD and early mortality, which may necessitate an alternative tx strategy. Efficacious “bridging” or pre-BsAb therapy could be a reasonable approach to improving outcomes for these high-risk pts. A prospective clinical trial is planned to explore this approach.