Efficacy, Toxicity, and Predictors of Outcomes with CD3-CD20 Bi-Specific Antibodies Post CAR T-Cell Failure for Aggressive B-Cell Lymphoma

Introduction: Chimeric antigen receptor T-cell therapy (CAR T) achieves durable remissions in 30-40% of patients (pts) with relapsed or refractory B-cell lymphoma (R/R BCL). For those with CAR T failure, few effective therapeutics are available. Although clinical trials with CD3-CD20 bispecific antibodies (BsAbs) in this population have demonstrated CR rates of approximately 40%, only small numbers of pts previously treated with CAR T were included with short follow-up. Furthermore, little is known of real-world practices and efficacy of BsAbs post CART failure. Methods: In a cohort of 830 pts (pts) with R/R BCL treated with CAR T between 2015 - 2024 across 14 academic institutions, we identified pts treated with BsAbs post CART failure. Wilcoxon rank-sum test and pooled t-test were utilized (two-sided P value of <.05) to determine the statistical significance of differences between variables. Time-to-event curves were estimated using Kaplan-Meier. Cox proportional hazards models were used to determine the impact of clinical variables in pts treated with BsAbs. Results: 429 pts in our cohort had CART failure. 64 pts were treated with BsAbs at a median of 218 days (range 32-2008) post CAR T failure (30= Epcoritamab, 21=Glofitamab, 7= Mosunetuzumab, and 6= Other) for a median of 4 cycles (4-12). 12 pts (19%) received BsAbs in combination with an adjunctive therapy (6= Revlimid, 2= XRT, 1= Zanubrutinib, 1= BV, 1= Polatuzumab, 1= Azacitidine). Clinical characteristics at time of BsAb administration included median age of 62 years (range 24-84), 59% male, 73% Caucasians, 17% African American, 39% with primary refractory disease with frontline therapy (PRD), 70% with DLBCL followed by 6% with tFL, 19% with double hit (DHL), and 30% with double expressor (DEL) lymphoma. Overall response rate (ORR) to BsAbs post CAR T was 54% (29/54 evaluable pts) with a CR rate of 33% (18/54). At a median follow-up of 400 days (d) after initiation of BsAbs, 66.7% pts remained in CR. 23.4% (15/64) developed grade 1-2 cytokine release syndrome (CRS) with BsAbs; no pts developed grade 3 CRS. 4 pts developed grade 1 neurotoxicity (NT); no pts developed higher grade NT. 6 pts received tocilizumab and 7 pts received steroids for CRS. Median PFS (mPFS) and median OS (mOS) for pts treated with BsAbs were 145 and 227 d respectively. After BsAbs failure, mOS was 75 d. Prior bendamustine exposure, use of adjunctive therapy with BsAbs, BsAbs construct, onset of CRS or use of steroids for toxicity management with BsAbs, did not impact PFS and OS with BsAbs treatment. Elevated LDH at time of BsAb administration impacted PFS (85 d vs NR, p =0.01), but not OS. Pts who developed NT had decreased mOS as compared to pts without NT (75 vs 532 d, p = 0.007). Pts with DHL (n= 11, 17%) vs non-DHL had inferior mPFS (80 d vs 168 d, p = 0.06) and significantly inferior mOS (115 vs 654 d, p =0.01). Pts with PRD vs not had both inferior mPFS (85 days vs 184 days, p = 0.05) and mOS (161 days vs 532 days, p < 0.05). 45% (29/64) of pts received BsAbs as first line salvage post CAR T and had improved mPFS when compared to those who received BsAbs ≥ 2nd line post CART (148 vs 103 d, p = 0.07); however, there was no difference in mOS between these two groups (p=0.2). Although there was no difference in ORR or CR between pts who experienced early (< 90 days post) vs late (> 90 days) post CART failure/relapse, mOS was markedly reduced in pts with early relapse vs late relapse post CART (150 vs 1045 d, p=0.02). In a multivariable analysis, inferior PFS was associated with elevated LDH (HR 9.0, 95% CI 3.0-36.4) and bulky disease (HR 4.4, 95% CI 1.7-11.6). Inferior OS was associated with DHL (HR 3.0, 95% CI 1.04-8.4), elevated LDH at time of BsAbs treatment (HR 3.3, 95% CI 1.1 - 9.5), early CART failure/relapse (HR 2.8 95% CI 1.1-7.0). Conclusions: In this large real-world multi-center cohort of pts treated with BsAbs after CAR T failure, response rates and survival were diminished as compared to results previously reported in clinical trials. Although there did not appear to be a benefit to the concurrent administration of additional therapies with BsAbs, the small sample size of this pt subset limits interpretability. In pts with prior CAR T, elevated LDH, DHL, and early CART failure within 90 days appear to be negative predictors of survival for treatment with BsAbs with dismal outcomes after BsAb failure. These pt subsets should be prioritized for clinical trials evaluating alternative novel therapies.