This is a randomized, two-cohort, open-label, phase 3, clinical trial to compare the efficacy and safety of N-803 plus tislelizumab and docetaxel (cohort A) or prior failed Health Authority-approved antiprogrammed death-1 (PD-1) or anti-programmed death-ligand 1 (PD-L1) CPI and docetaxel (cohort B) versus docetaxel monotherapy (cohorts A and B).

For each cohort, enrolled participants will be randomized 2:1 to treatment in the experimental arm or the control arm. For cohort A, the randomization will be stratified by geographical region (North America vs Europe vs Asia vs Other), NSCLC histology (squamous vs nonsquamous), and actionable genomic alteration (AGA) (epidermal growth factor receptor \[EGFR\]/anaplastic lymphoma kinase \[ALK\]/ROS proto-oncogene 1, receptor tyrosine kinase \[ROS1\] vs Other AGA vs No AGA).

For cohort B, the randomization will be stratified by geographical region (Americas vs Asia Pacific \[PAC\] vs Other), NSCLC histology (squamous vs nonsquamous), and actionable genomic alteration (AGA) (Yes vs No).

This study is designed to assess the levels of drug exposure following treatment with tislelizumab administered as a subcutaneous (SC) injection compared to intravenous infusion (IV) as first-line therapy in adults with gastric or gastroesophageal junction (GEJ) that is locally advanced and cannot be surgically removed or has spread from the stomach to other areas of the body. Approximately 351 patients will be participating in this study. The study is composed of a screening period, a treatment period, and a follow-up period.