Risk Factors for Cytokine Release Syndrome in Patients Receiving Bispecific Antibodies for B-Cell Lymphoma: A Single-Center, Retrospective Cohort Study

Introduction: CD20xCD3 bispecific antibodies (BsAb) have been investigated and approved for use in B-cell non-Hodgkin lymphomas (B-NHLs) with manageable toxicities. The main adverse event of clinical interest is cytokine release syndrome (CRS) and immune effector cell-associated neurotoxicity syndrome (ICANS). The risk factors for CRS and ICANS are not well understood and are the focus of this study.

Methods: We performed a single center retrospective study of patients with B-NHL treated between September 2018 and April 2024 with investigational or commercial BsAb. We obtained baseline characteristics, safety and efficacy outcomes. Uni-and multivariate analysis were performed to identify associations between CRS and clinical variables of interest.

Results: 84 patients with B-NHLs were treated with BsAb (median age 66; 70.2% male; 92.8% Caucasian; 94% non-Hispanic; 87% ECOG 0-1). Histologic subtypes included diffuse large B-cell lymphoma (DLBCL) (n=46, 40.8%), follicular lymphoma (FL) (n=20, 23.8%), mantle cell lymphoma (MCL) (n=14, 16.7%), and Richter’s transformation-DLBCL (n=4, 4.7%). BsAb was given as a single agent (n= 58, 69.0%) or in combination with other agents (n=26, 31.0%). Patients were treated frontline (n=10, 11.0%) or in the relapsed refractory (R/R) setting (n= 74, 89%) with a median of 3 (range 1-10) prior lines of therapy. 20 patients (23.3%) were previously treated with chimeric antigen receptor (CAR) T-cell therapy, 11 patients within 6 months prior to BsAb. Patients were treated with epcoritamab (Epco) (n=39, 46.4%), glofitamab (Glofit) (n=11, 13.1%), plamotamab (Plamo) (n=18, 21.4%) and mosunetuzumab (Mosun) (n=16, 19.1%). Complete response (CR) rate was 30.9% and overall response rate (ORR) was 61.8%. At median follow-up of 9.7 months, the median progression free survival (PFS) and median overall survival (OS) was 8.7 and 13.0 months, respectively.

CRS of any grade occurred in 48 (57.1%) patients: grade 1 (n=19, 22.5%), grade 2 (n=27, 32.1%), and grade 3 (n=2, 2.4%). There was no instance of grade 4 CRS. The maximum grade CRS occurred most frequently after the 1^st dose (n= 28, 2 Epco, 7 Glofit, 12 Plamo, 6 Mosun) followed by the 3^rd dose (n=14, 13 Epco,1 Mosun). CRS was managed with additional steroids in 14 patients (29.2%) and tocilizumab in 15 patients (31.3%). 15 patients (31.3%) had recurrent CRS. BsAb was held or discontinued due to CRS in 6 (7.1%) and 2 (2.4%) patients, respectively. There were no deaths attributed to CRS. Only 4 patients (4.7%) experienced ICANS- grade 1 (n=3, 3.6%) and grade 3 (n=1, 1.2%).

In univariate analysis, patients with DLBCL had lower associated risk of CRS (OR 0.28, 95% CI 0.11-0.69, p=0.008). There is a trend towards an increased risk of CRS in patients with MCL(OR 3.27, 95% CI 0.84-12.74, p=0.14). Patients with CRS had higher rate of bone marrow involvement (OR 4.80, 95% CI 1.46-15.82, p=0.01). There was no difference in CRS risk when assessing baseline demographics, BsAb drug, presence of bulky disease, ECOG, weight or BMI, baseline blood pressure, history of diabetes mellitus, use of granulocyte colony-stimulating factor, and prior bendamustine exposure. There were similar rates of prior CAR-T therapy within 6 months prior to BsAb in patients with or without CRS (OR 0.89, 95% CI 0.33-2.45, p=1.00). The mean absolute neutrophil counts (ANC, k /μl ) and albumin (Alb, g/dL) level at baseline (BL) and 1^st full-dose BsAb were mildly but significantly lower in patients with CRS (ANC at BL: 3.6 vs 5.1, p =0.02; ANC at 1^st full-dose: 4.1 vs 5.8, p=0.04; Alb at BL: 3.6 vs 3.8, p= 0.01; Alb at 1^st full-dose: 3.5 vs. 3.9, p <0.001). Creatinine, liver enzymes, and inflammatory markers (ferritin, fibrinogen, C-reactive protein, d-dimer) did not differ between groups. CR, ORR, PFS and OS were not impacted by CRS. In multivariate analysis with histologic subtype, the association of bone marrow involvement of lymphoma and CRS risk had reduced significance (OR: 3.67, 95% CI 0.99-13.68, p=0.052).

Conclusion: We report a lower risk of CRS in patients with DLBCL and trend towards higher risk in patients with MCL or patients with bone marrow involvement. CRS group had slightly lower ANC and albumin level. Inflammatory marker levels, prior CAR-T, and prior bendamustine did not impact the risk of CRS. Further studies in larger data sets are needed to identify patients of high CRS risk to allow for early management, and of low risk to allow for treatment in the outpatient setting.