Clinical predictors of cytokine release syndrome in non-single chain CD3 directed therapies among cancer patients

Background: Bispecific T-cell engager (BiTE) antibodies are novel agents in cancer therapeutics which work by simultaneously binding a cancer cell specific antigen and engaging T-Cell. Major safety concern with use of BiTE therapy based on CD3 platforms is cytokine release syndrome (CRS) from massive unchecked T-cell activation. Thus initial dosing of BiTE requires hospitalization for monitoring and treating CRS. This leads to intensive care requirement and mortality if CRS is not identified and intervened early. We assessed the prevalence of CRS in cancer patients receiving non-single chain CD3 based BiTE therapy and contrasted clinical parameters in patient who developed CRS against those who did not. Methods: A retrospective chart review was performed on patients from our institution who were admitted for the initial and step up dosing of the BiTE therapy for monitoring of CRS between September 2024 - September 2025. The patients were divided in two groups, those who developed CRS and those who did not. Demographic data, previous treatment history, clinical characteristics, laboratory data and duration of hospitalization at the time of admission were noted. Results: 20 patients were identified with median age of 67 years. 30% were male and 70% were female. 35% received tarlatamab, 30% got glofitamab, 10% got talquetamab and epcoritamab each, 1 patient got mosunetuzumab and linvoseltamab. All 20% patients who had grade two or higher neurotoxicity syndrome had intracranial disease. 35% (n=7) patients had CRS. Three patients had grade I, two patients had grade II and two had Grade IV CRS. Culture data was negative for all CRS patients up to 1 week after the dose. All CRS was seen after first dose of BITE therapy, except 1 patient who developed it after second dose of tarlatamab. Monocyte to lymphocyte (M:L) ratio 24 - 48 hours after the dose of BiTE (1.51 in CRS vs 0.59 in no CRS, p=0.028*), days between diagnosis and BITE therapy (median 864 days in CRS vs 448 days in no CRS patients, p=0.0218*), and previous lines of therapy (LOT)(4 vs 3, p=0.0224*) were noted to be significantly higher in patients with CRS. Logistical regression suggested that only M:L ratio was significantly related to CRS (odds ratio 5.64, p=0.0372*), while previous LOT (OR 6.93, p=0.060) and time of disease before therapy (OR 1.00, p=0.058) were not significantly related. Discussion: Despite high CRS prevalence in patients receiving BiTE therapy, the literature describing the mechanism of CRS specific to BiTE therapy is very limited. Monocytes get involved in the uncontrolled release of cytokines which may appear early in circulation. The patients who sustain malignancy for longer and had more treatment lines may develop immune build up against the tumor which may increase the chance of CRS in such patients. These factors need to be explored further in the clinical setting to predict CRS and triage clinical resources.