Abstract Melanoma has seen a remarkable improvement in the treatment of advanced disease. However, the majority of patients still relapse with the emergence of resistant disease. This is especially true for BRAF inhibitors, were multiple mechanisms of resistance have been uncovered. However, the lack of dependable models to develop approaches to overcome resistance, poses a real challenge for clinical decisions on second line therapies. In addition, combination therapies necessary to achieve long term responses increase the number of possibilities to be assessed. In this study, we developed patient derived xenograft (PDX) models from BRAF inhibitor resistant patients and, after characterization, tested rational second line combination therapies in co-clinical trials.Fresh tumor tissue from 12 patients after progression on vemurafenib or dabrafenib was implanted into immune deficient NSG mice. Once the tumor grafts were established, mice were continually dosed with PLX4720 to approximate clinical plasma levels to preserve the resistant phenotype. Xenografts were then characterized for genomic alterations and pathway activation status using a targeted sequencing array and a reverse phase protein array, respectively. We found at least 2 and up to 9 known deleterious alterations in all samples, many of which have been described as conferring resistance. The protein array clustered into two major groups: MAPK pathway activated and PI3K pathway activated, again consistent with previous reports. Integrating these data allowed us to design rational combination therapies for in vivo testing. We treated a PDX model with MET amplification and high phospho c-Met with the BRAF inhibitor LGX818 20mg/kg QD, the MEK inhibitor MEK162 3mg/kg QD, and the c-Met inhibitor INC280 25mg/kg QD (all p.o.). Although INC280 alone led to significant decrease in tumor growth, only the LGX818/ MEK162/ INC280 triple combination led to complete and sustained tumor regression in all mice. Further, we tested a second PDX model with a MET amplification (albeit broader and lower) but baseline pMET on the protein level. In this model, INC280 did not show any antitumor effect confirming the specificity for c-Met and the importance of validating genomic data.In summary, using our BRAF inhibitor resistant PDX model pipeline we were not only able to identify actionable alterations in all models, but also to test rational second line combination therapies in pre-clinical in vivo trials. Genomic data alone might not be sufficient to accurately predict therapy responses and protein arrays were a valuable tool to validate deleterious alteration calls. This study represents an important step towards improving personalized medicine in melanoma, and highlights the potential use of c-Met inhibitors in melanoma combination therapy in a defined subset of patients. Citation Format: Clemens Krepler, Min Xiao, Katrin Sproesser, Patricia Brafford, Batool Shannan, Marilda Beqiri, Wei Xu, Bradley Garman, Katherine L. Nathanson, Xiaowei Xu, Giorgos Karakousis, Gordon B. Mills, Yiling Lu, Giordano Caponigro, Markus Boehm, Malte Peters, Lynn Schuchter, Meenhard Herlyn. Personalized pre-clinical trials in BRAF inhibitor resistant patient derived xenograft models of melanoma identify c-Met as an effective second line combination therapy target. [abstract]. In: Proceedings of the 106th Annual Meeting of the American Association for Cancer Research; 2015 Apr 18-22; Philadelphia, PA. Philadelphia (PA): AACR; Cancer Res 2015;75(15 Suppl):Abstract nr 2842. doi:10.1158/1538-7445.AM2015-2842