Early toxicity and temporal mortality patterns following bispecific T-cell engager therapy: A real-world multicenter analysis.

e23336 Background: Bispecific T-cell engagers are increasingly used across hematologic malignancies and, more recently, solid tumors. Cytokine release syndrome (CRS) and immune effector cell–associated neurotoxicity syndrome (ICANS) are hallmark toxicities that typically occur early after treatment initiation, yet real-world data describing severe events and downstream outcomes across clinically relevant time windows remain limited. Methods: We conducted a retrospective cohort study using the TriNetX US Collaborative Network. Adult patients (≥18 years) with small cell carcinoma, B-cell non-Hodgkin lymphoma, multiple myeloma, or acute lymphoblastic leukemia who received at least one bispecific T-cell engager (tarlatamab, teclistamab, epcoritamab, mosunetuzumab, glofitamab, or blinatumomab) were included. The index date was defined as first bispecific exposure. Patients with myeloid malignancies, non–small cell lung cancer, selected solid tumors without approved bispecific therapies, or prior chimeric antigen receptor T-cell (CAR-T) therapy were excluded. Outcomes included all-cause mortality, grade ≥3 CRS, grade ≥3 ICANS, Tocilizumab use and early high-acuity supportive care utilization (ICU admission, vasopressors, mechanical ventilation, shock). Outcomes were assessed at 7-, 30-, and 90-day post-index. For supportive care outcomes, patients with the outcome prior to the time window were excluded to capture incident events. Results: A total of 3,642 patients were included. All-cause mortality occurred in 0.7% at 7 days, 3.5% at 30 days, and 9% at 90 days, corresponding to survival probabilities of 99.3%, 96.5%, and 91%, respectively. Severe CRS (grade ≥3) was infrequent and occurred predominantly early, with incidences of 0.5% at 7 days, 0.8% at 30 days, and 1.0% at 90 days. Similarly, severe ICANS (grade ≥3) occurred in 0.4% at 7 days, 0.9% at 30 days, and 1.3% at 90 days. Tocilizumab was administered in 6.7% at 7 days, 9.0% at 30 days, and 9.8% at 90 days. Incident high-acuity supportive care utilization was observed in 4.3% at 7 days, 9.8% at 30 days, and 17.0% at 90 days among patients at risk, whereas hospitalization status was common and reported at 53.0%, 58.4%, and 61.9%, respectively. Conclusions: In this large real-world cohort of patients treated with bispecific T-cell engagers, coded severe CRS and ICANS events were uncommon and plateaued early, whereas mortality and healthcare utilization continued to accrue over time. Given the known under-capture of immune-related toxicity in real-world data, these findings suggest that ongoing risk over follow-up likely reflects a combination of treatment-related toxicity and underlying disease progression, highlighting the need for standardized toxicity ascertainment and strategies to identify patients at highest risk for early clinical deterioration.