ABSTRACT Aim: MEDI4736 is a human IgG1 mAb, engineered to prevent antibody-dependent cell-mediated cytotoxicity (ADCC) activity, that blocks PD-L1 binding to PD-1 and CD-80. This expansion study (NCT01693562) assessed MEDI4736 in pts with NSCLC, melanoma (cutaneous and uveal), hepatocellular carcinoma (ca), squamous cell ca of the head and neck (SCCHN), gastroesophageal ca, pancreatic ca, and triple negative breast ca. Methods: 10–20 pts (performance status [PS] 0–1) were initially enrolled per tumor type/study arm, with expansion allowed upon observation of clinical activity. MEDI4736 was administered as 10 mg/kg IV every 2 wks (q2w) for up to 12 months, with retreatment permitted for progression during follow-up. Response was assessed by RECIST v1.1. Results: This multi-arm dose expansion study was initiated in Sep 2013. As of 14 April 2014, 288 pts median age 61y (20–96), 57% male, PS 0/1/unknown (31%/65%/4%), median 3 (1–11) prior treatments, had received a median of 3 (1–20) doses of MEDI4736 10 mg/kg q2w and were evaluable for safety. Safety profile appeared to be consistent across tumor types and with previous reports. Treatment-related adverse events (AE) in 34% of pts (any Grade [Gr]); Gr ≥3 in 5.6%; no Gr 4–5); AEs led to discontinuation of study drug in 5.8% (drug-related in 1 pt). Most frequent treatment-related AEs were fatigue (14%), nausea (8.7%), and rash (5%). Gr 2 pneumonitis (resolved with drug interruption/steroids) was seen in 1 pt (0.4%). No colitis reported. No immunogenicity detected at the 10 mg/kg dose; 1/89 with antidrug antibodies impacting PK. Tumor shrinkage has been seen across all histologies, with responses as early as 6 wks and after initial progression in some pts. Clinical activity to date appears durable (maintained ≥ 48 wks in 5 pts). Analysis of correlation between activity and clinical attributes or biomarkers including PD-L1 expression is ongoing. Conclusions: Results indicate that MEDI4736 has a manageable safety profile even in heavily pretreated pts. Evidence of clinical activity has been seen across all 9 tumor types; other tumor types will be evaluated. Current experience supports development of MEDI4736 as monotherapy and in combination with other anti-cancer therapies for multiple tumor types. Disclosure: N.H. Segal: Ad Board participation with MedImmune, Alkermes Scientific and Imugene. Research funding from BMS and Pfizer. MedImmune considers the research funding received for the conduct of a MedImmune-sponsored study as conflict of interest; O. Hamid: My only conflict here is the research funding that I received from Medimmune. MedImmune considers the research funding received for the conduct of a MedImmune-sponsored study as conflict of interest; W. Hwu: No financial interest in MedI4736,process involved in the research.I dont own stock,serve on ad board,board of director or other substantive relationship with research sponsor who consider research fund rec'd for conduct of Medi-sponsored study a COI; C. Massard: MedImmune considers the research funding received for the conduct of a MedImmune-sponsored study as conflict of interest; M. Butler: Received financial support from MedImmune Inc. to cover the costs of conducting MedImmune-sponsored clinical studies of Medi4736. MedImmune considers the research funding received for the conduct of a MedImmune-sponsored study as conflict of interest; S. Antonia: MedImmune considers the research funding received for the conduct of a MedImmune-sponsored study as conflict of interest; A. Blake-Haskins: Employee of MedImmune and owns stock/stock options in AstraZeneca; P.B. Robbins: Employee of MedImmune and owns stock/stock options in AstraZeneca; X. Li: Employee of MedImmune and owns stock/stock options in AstraZeneca; J. Vasselli: Employee of MedImmune and owns stock/stock options in AstraZeneca; S. Khleif: MedImmune considers the research funding received for the conduct of a MedImmune-sponsored study as conflict of interest. I also function as a clinical advisor.