Impact of Prior Bendamustine Exposure on Bispecific Antibody Treatment Outcomes in Patients with Relapsed/Refractory Follicular Lymphoma

Introduction: Bispecific antibodies (BsAbs) are approved for patients with relapsed/refractory (R/R) follicular lymphoma (FL) based on single-arm clinical trials. These T-cell engaging agents achieve approximately 60% complete response rate (CRR), but information regarding real-world prognostic factors is scarce. In light of the deleterious impact of pre-treatment bendamustine (benda) exposure on chimeric antigen receptor (CAR) T-cell outcomes, and given the similar mechanism of action, we aimed to assess the impact that this alkylating agent has on BsAbs outcomes in R/R FL patients. Methods: We conducted a multicenter, international study including patients with R/R FL treated at 13 sites with CD20/CD3-targeted BsAbs from November 2017 until March 2024. We analyzed response rates and survival outcomes according to prior benda exposure in the full cohort and according to treatment partner (BsAb monotherapy vs combination). Then, we carried out an inverse probability of treatment weighting (IPTW) analysis in the monotherapy group to compare benda-exposed and naïve patients with matched key baseline characteristics. Finally, we examined the impact of benda dose and washout prior to BsAbs on efficacy outcomes. Results: We included 169 patients with R/R FL who received BsAb therapy, 73 (43%) of whom had been exposed to benda at a median of 42 months (range 1-141) prior to BsAb start; in 31 (42%) patients, it was the last line prior to BsAbs. Most patients (N=114, 67%), received the BsAb in monotherapy. Median follow-up for the full dataset was 21.8 months (95%CI 17.7-31.4). At time of BsAb start, median age was 64 years (IQR 55-72), 89 (53%) patients were male and most had advanced stage of disease (84%), high FLIPI score (54%) and 2 or more prior lines of therapy (77%, median 2 [range 1-9]); 49% were POD24 and 4% had received CAR T-cells. The benda-exposed cohort had more prior lines (>2, 58% vs 29% [p<0.01]) and worse performance status (ECOG >0, 46% vs 29%) in comparison with naïve patients. First, we analyzed the impact of benda on several parameters. Patients with prior benda exposure had lower median CD4+ cells (270 vs 387/mm3, p=0.02) at time of BsAb start, compared to naïve patients. In terms of early toxicity, there were no differences in cytokine release syndrome (42% vs 53% any grade [p=0.2], mainly grade 1-2), neurotoxicity (1% vs 4% any grade [p=0.4]) or infections (69% vs 73% [p=0.7]) between exposed and naïve patients. Regarding efficacy, overall and complete response rates (ORR [CRR]) were also comparable (83% [66%] vs 86% [70%], respectively). For the full cohort, median progression-free survival (PFS) and overall survival (OS) from BsAb start were 30 months and not reached (NR), respectively. The benda-exposed subset had a trend towards shorter PFS (18.0 months vs NR [p=0.06]) and had a shorter OS (72.0 months vs NR [p=0.03]) in comparison with naïve patients. Taking into account the presence of a treatment partner with BsAbs, the combination cohort (n=55) had comparable ORR, CR, PFS and OS between those with (N=22, 40%) and without (N=33, 60%) prior benda exposure. Focusing on the BsAb-monotherapy cohort (N=114), we observed a shorter PFS (p=0.01), duration of response (DoR, p=0.03) and OS (p=0.03) for benda-exposed patients. In light of the different baseline characteristics between exposed and naïve patients, we carried out an IPTW analysis balancing for 15 variables (including age, prior lines of therapy, ECOG and FLIPI, among others). In the balanced analysis, none of the efficacy outcomes were significantly different between benda-exposed and naïve patients, including CRR (61% vs 65% [p=0.47]), PFS (median 18 months vs NR [p=0.14]) and OS (NR vs NR [p=0.3]). Given the ample range in washout from last benda dose to BsAb start, we analyzed the impact of this interval on efficacy outcomes. We did not identify any difference according to this time interval; of note, only 11 (15%) and 21 (29%) patients had a washout shorter than 12 and 24 months, respectively. In terms of benda dose, median number of cycles was 6 (range 1-10). We did not identify an impact of number of benda cycles on response rate or survival. Conclusions: Bendamustine-containing regimens do not seem to confer a negative prognostic impact before BsAb therapy in follicular lymphoma. Longer follow-up and larger cohorts, particularly of patients exposed to benda within 12 months of BsAb treatment, are warranted to confirm these results.