Background: Tislelizumab (tisle) monotherapy or plus chemotherapy demonstrated antitumor activity in patients (pts) with solid tumors, including esophageal squamous cell carcinoma (ESCC) (NCT03469557 and CTR20160872). Methods: In this global phase 3 study (NCT03430843), adults with histologically confirmed advanced/unresectable or metastatic ESCC whose disease progressed following prior systemic therapy with ≥1 evaluable lesion per RECIST v1.1 and an Eastern Cooperative Oncology Group performance score (ECOG PS) of ≤1 were included. Pts were randomized (1:1) to receive tisle 200 mg intravenously every 3 weeks or investigator-chosen standard chemotherapy ([ICC]; paclitaxel, docetaxel, or irinotecan) and treated until disease progression, unacceptable toxicity, or withdrawal. Stratification factors included ICC option, region, and ECOG PS. The primary endpoint was overall survival (OS) in the intent-to-treat (ITT) population. The key secondary endpoint was OS in the programmed death-ligand 1 (PD-L1)+ population (visually-estimated combined positive score [vCPS] ≥10%, by VENTANA PD-L1 SP263 assay). Other secondary endpoints included (by RECIST v1.1) progression-free survival, overall response rate (ORR), duration of response (DoR), and safety. Results: Overall, 512 pts (median age: 62 years; range 35-86 years) from 132 sites in 10 countries in Asia (404 pts [79%]), Europe, and North America (108 pts [21%]) were randomized to tisle (n=256) or ICC (n=256) (ITT population). Of these, 157 pts (tisle [n=89], ICC [n=68]) had vCPS ≥10% (PD-L1+ population). On 1 Dec 2020 (data cut-off), median follow-up was 8.5 months (m) with tisle and 5.8 m with ICC. The study met its primary endpoint: tisle clinically and significantly improved OS vs ICC in the ITT population (median OS: 8.6 vs 6.3 m; HR 0.70, 95% CI 0.57-0.85, p=0.0001). Tisle also demonstrated significant improvement in OS vs ICC in the PD-L1+ population (median OS: 10.3 vs 6.8 m; HR 0.54, 95% CI: 0.36-0.79, p=0.0006). Survival benefit was consistently observed across pre-defined subgroups, including baseline PD-L1 status and region. Treatment with tisle was also associated with a higher ORR (20.3% vs 9.8%) and more durable response (median DoR: 7.1 vs 4.0 m; HR 0.42, 95% CI 0.23-0.75) than ICC in the ITT population. Fewer pts had ≥Grade 3 (46% vs 68%) treatment-emergent adverse events with tisle vs ICC. Of these, fewer ≥Grade 3 AEs were treatment-related (TR) with tisle vs ICC (19% vs 56%). Fewer pts discontinued tisle vs ICC (7% vs 14%) due to a TRAE. Conclusion: Tisle demonstrated statistically significant and clinically meaningful improvement in OS vs ICC in pts with advanced or metastatic ESCC who had disease progression during or after first-line systemic therapy. Tisle showed a higher and longer response vs ICC. The safety profile of tisle was more favorable than ICC. Clinical trial information: NCT03430843