Background: All immune therapies that rapidly activate T cells, including T cell engagers, can induce cytokine release syndrome (CRS). Tebentafusp (tebe), a T cell receptor bispecific (gp100 x CD3) can also induce skin adverse events (AEs), due to gp100+ cutaneous melanocytes. CRS and skin AEs may require management with short term corticosteroids, which may also be used as premedication for subsequent tebe doses. Here we report the first analysis of systemic corticosteroid use and correlation with efficacy from a Phase (Ph) 3 trial for any T cell engager. Methods: Post hoc analyses were performed on the tebe arm of the Ph3 [NCT03070392] study in previously untreated HLA-A*02:01+ metastatic uveal melanoma (mUM) (N = 245). Due to the low rate of severe AEs in Ph1 trials, prophylactic corticosteroids were not mandated. The association between overall survival (OS) and corticosteroid use (new start within 30 days of first tebe dose) was investigated using landmark analyses in the safety population. Multivariate analyses were adjusted for key patient characteristics and AEs of special interest: CRS, rash, and liver function test (LFT) elevation. Steroid type (hydrocortisone vs. others) and treatment duration (1 vs. > 1 day) were also investigated. Results: In the Ph3 trial, 64/245 (26%) patients received new systemic corticosteroid within 30 days after the first dose of tebe, mostly for treatment of AEs (56/64, 88%) or pre-medication due to previous AE (14/64, 22%). 25 of the 64 patients received corticosteroids only for a single day. The most frequent AEs (≥15%) were rash (18/64, 28%), CRS (15/64, 23%), and hypotension (12/64, 19%). In a logistic regression model, elevated baseline LDH, the dominant prognostic marker, was most strongly associated with use of corticosteroids (p = 0.01). In the multivariate analysis, corticosteroids were not associated with any significant OS difference (HR 1.41, 95% CI 0.83-2.4, p = 0.2) and this effect did not differ in patients with or without CRS, rash or LFT elevation (all interaction tests p > 0.2). There was no difference in OS according to corticosteroid type or whether administered for 1 vs > 1 day. Conclusions: This is the first analysis from a phase 3 trial of the impact of systemic corticosteroids on survival for a T cell engaging cancer therapy. The vast majority of tebe-treated patients (84%) either did not require corticosteroids (74%) or only received them on a single day (10%). The most frequent reason for corticosteroid use was an emergent AE, including CRS and rash. Corticosteroid use following the pre-specified AE guidelines was not associated with any significant impact on OS. Clinical trial information: NCT03070392.