8110 Background: The UC Davis - Jackson Labs Consortium has established a PDX resource of over 50 NSCLC models (12 EGFR mutant) for experimental therapeutics and personalized medicine, using the NSG mouse. Here we report on a unique PDX model, LG703, derived from a patient at time of acquired E resistance. This tumor harbors EGFR L858R with no T790M mutation, high MET and very high EGFR protein expression. The patient is in remission following treatment with AFAT-CET. In this model, we investigated the respective contribution of CET and AFAT relative to E on tumor growth and signal transduction. Methods: Individual mice implanted with LG703 tumor fragments were randomized to E (50 mg/kg qd po), AFAT (20 mg/kg qd po), CET (10 mg/kg twice weekly iv), AFAT-CET, or vehicle control (n per arm = 12) for 3 weeks followed by a 75-day monitoring period. Changes in signal transduction mediators and RTKs were assessed after 6 and 24h treatment exposures using kinase arrays (R&D systems) and immunoblotting. Results: AFAT, CET and AFAT-CET resulted in complete tumor response (CR) during the 21-day treatment period; whereas E resulted in temporary growth delay. After cessation of treatment, E-treated mice progressed rapidly and AFAT-treated mice progressed within 2 weeks. Mice treated with CET or AFAT-CET remained in complete remission. At 6h, E and AFAT significantly (>70%) reduced EGFR phosphorylation as well as that of AKT1, AKT2, ERK1, p38a, RSK1 and p70S6K. At 24h, E-treated tumors had returned to baseline-like states for all factors, AFAT showed an intermediate response, and CET, alone or with AFAT, achieved the greatest inhibition of pEGFR with sustained inhibition of all downstream effectors. Conclusions: In the LG703 PDX model, CET and AFAT+CET resulted in CRs, mirroring the patient's response to similar therapy, associated with sustained inhibition of pEGFR and multiple downstream signaling factors. In contrast, E exhibited only temporary growth delay associated with transient inhibition of EGFR pathway factors. These experiments demonstrate the potential of this PDX resource to assess new therapeutic strategies in models representing individual patients. Supported by BJALCF.