INTRODUCTION: The phase 3 RATIONALE-303 trial (NCT03358875) investigated the efficacy and safety of tislelizumab versus docetaxel in pretreated patients with advanced NSCLC. Here, we report the efficacy and safety results and describe the exploratory biomarker analyses. METHODS: A total of 805 patients aged more than or equal to 18 years with locally advanced or metastatic squamous or nonsquamous NSCLC were randomized 2:1 to intravenous tislelizumab 200 mg or docetaxel 75 mg/m every 3 weeks. Co-primary end points were overall survival (OS) in the intent-to-treat (ITT) and programmed death-ligand 1 (PD-L1) tumor cell expression greater than or equal to 25% populations. The exploratory biomarker analyses included PD-L1 expression, tumor mutation burden, and gene expression profile. RESULTS: At the prespecified interim analysis (August 10, 2020), the co-primary end point of OS in the ITT population was met, with a statistically significant and clinically meaningful improvement in OS with tislelizumab versus docetaxel (median 17.2 versus 11.9 mo, respectively; hazard ratio [HR] = 0.64, p < 0.0001). At the final analysis (July 15, 2021), the other co-primary end point of OS in the PD-L1 tumor cell greater than or equal to 25% population was further met (median 19.3 versus 11.5 mo, respectively; HR = 0.53, p < 0.0001), and OS continued to improve in the ITT population (median 16.9 versus 11.9 mo, respectively, HR = 0.66). Exploratory biomarker analyses revealed the potential association of NOTCH1-4 mutations with improved tislelizumab efficacy for both OS and progression-free survival, whereas tissue tumor mutation burden correlated with progression-free survival benefit, but not OS benefit. No new safety signals were identified. CONCLUSIONS: Tislelizumab was found to have a significantly improved and long-term clinical benefit in OS versus docetaxel in pretreated patients with advanced NSCLC, regardless of PD-L1 expression.