Background: Erythropoietin-producing hepatocellular receptor A2 (EPHA2) is overexpressed on the cell surface of many tumors and is associated with poor prognosis, suggesting EPHA2 as a target for cancer therapy. DS-8895a is an afucosylated, humanized anti-EPHA2 IgG1 monoclonal antibody with potent cytotoxicity. We report results from a phase I clinical trial to determine safety, tolerability, and pharmacokinetics (PK) of DS-8895a in Japanese patients with advanced solid tumors (NCT02004717). Methods: Step 1 (dose escalation cohort) had patients with advanced solid tumors and comprised of six dose levels (0.1–20 mg/mL, intravenous infusion, every 2 weeks [Q2W]) with a 28-day dose limiting toxicity (DLT) evaluation period. Step 2 (expansion cohort) patients had gastric or esophageal cancer confirmed to be EPHA2 positive by immunohistochemistry. Dose level in Step 2 was determined based on results obtained in Step 1. We evaluated safety, PK, potential biomarkers including circulating NK cells and cytokines, and tumor response. Results: Maximum tolerated dose was not reached in Step 1 (n = 22). DS-8895a was administered at 20 mg/kg Q2W in Step 2 (n = 15). Among 37 patients in the safety analysis set, adverse events (AEs) were reported in 97.3% (64.9% drug-related); 35.1% presented grade ≥ 3 AEs (8.1% drug-related). Dose delay and study discontinuation due to AEs (treatment related: grade 4 platelet decrease, hypoesthesia, hypotension, peripheral coldness, nausea, and vomiting) were observed in one and four patients (20 mg/kg), respectively. Infusion-related reactions occurred in 51.4% of patients resulting in 10 dose interruptions with one discontinuation. Serum inflammatory cytokines were transiently increased 4 h from the end of infusion drug administration. Serum DS-8895a maximum and trough concentrations increased dose-dependently. Biomarkers had no apparent relationship to best overall response. Seven patients in Step 1 achieved stable disease; in Step 2, six patients achieved stable disease and one patient achieved partial response. Conclusions: DS-8895a was safe and well tolerated up to 20 mg/kg. The PK of DS-8895a was dose-dependent as expected. Clinical trial information: NCT02004717