Background: LEAP-015 (NCT04662710) is a randomized, open-label, 2-part, phase 3 study of the safety and efficacy of lenvatinib + pembrolizumab + chemotherapy as a first-line treatment for advanced/metastatic gastroesophageal adenocarcinoma. We report findings from part 1, the safety run-in, of LEAP-015. Methods: Eligible patients had untreated, HER2-negative, locally advanced unresectable or metastatic gastroesophageal adenocarcinoma, measurable disease per RECIST v1.1, and ECOG performance status 0 or 1. In part 1, patients received induction with IV pembrolizumab 400 mg Q6W (×2) + oral lenvatinib 8 mg QD + investigator choice of chemotherapy (capecitabine + oxaliplatin [CAPOX] Q3W ×4 or 5-fluorouracil + leucovorin + oxaliplatin [mFOLFOX6] Q2W ×6) and consolidation with pembrolizumab 400 mg Q6W for ≤16 doses + lenvatinib 20 mg QD; dose-limiting toxicities (DLTs), defined as selected prespecified grade ≥3 adverse events (AEs) or any-grade thromboembolic events, were evaluated for 21 days after the first dose of study intervention. If ≥3 DLTs occurred in either oxaliplatin-containing regimen, then enrollment in part 2 may be delayed to allow for the examination of safety data and to consider design changes. In part 1, the primary end point was safety and tolerability in all patients. Preliminary efficacy was also assessed in part 1. Objective response and disease control rate were assessed per RECIST version 1.1 by blinded independent central review. Results: In part 1, 15 pts received ≥1 dose of lenvatinib + pembrolizumab + chemotherapy. 1 DLT of grade 3 asthenia occurred in the CAPOX cohort and 1 DLT of grade 4 neutropenia occurred in the FOLFOX cohort. Median time from first dose to data cutoff (Oct 13, 2021) was 7 mo (range, 7-9). Treatment-related AEs occurred in 14 patients (93%), with grade 3/4 events in 8 patients (53%). 4 patients (27%) discontinued any drug because of a treatment-related AE, and no patient discontinued all drugs because of a treatment-related AE. No patients died because of a treatment-related AE. No grade ≥3 immune-mediated AEs or infusion reactions occurred. Objective response was reported in 11 of 15 patients (73%; 95% CI, 45-92) who received ≥1 dose of treatment. Disease control rate was reported in 14 of 15 patients (93%; 95% CI, 68-100) who received ≥1 dose of treatment. Conclusions: In the safety run-in of LEAP-015, lenvatinib + pembrolizumab + chemotherapy was associated with a manageable safety profile in the first-line treatment of advanced/metastatic gastroesophageal adenocarcinoma. Preliminary antitumor activity was observed for lenvatinib + pembrolizumab + chemotherapy. Part 2 will evaluate the efficacy and safety of lenvatinib + pembrolizumab + chemotherapy versus chemotherapy in this same patient population and patient accrual is ongoing. Clinical trial information: NCT04662710.