Background: Use of checkpoint inhibitors, alone or with chemotherapy, has emerged as the preferred standard treatment for patients with advanced, driver-negative non-small cell lung cancer (NSCLC). While outcomes are superior to chemotherapy alone, only a subset of patients achieve durable response and long term survival. One potential mechanism of primary resistance to checkpoint inhibitors is the lack of tumor-infiltrating lymphocytes. Inhibition of mitogen-activated protein kinase (MAPK) kinase (MEK) increases the number of CD8+ T-cell within a tumor and has shown synergy with anti-programmed death-ligand 1 (PD-L1) antibodies. The combination of the MEK inhibitor cobimetinib and the PD-L1 antibody atezolizumab has led to limited responses in colorectal cancer, a tumor typically non-responsive to checkpoint inhibition. This phase II trial explores the combination of cobimetinib and atezolizumab in patients with PD(L)1-refractory NSCLC. Methods: This phase II study is being conducted through the Experimental Therapeutics Clinical Trials Network (ETCTN #10166). Eligible patients have advanced NSCLC with primary resistance to anti-PD(L)1 therapy (defined as progression noted within 6 months of initiating therapy) and tumor amenable to serial core biopsy. Patients will receive atezolizumab 840mg intravenously every 2 weeks and cobimetinib 60mg orally for 21 days in 28-day cycles. Two cohorts will enroll in parallel, defined by presence or absence of a KRAS mutation. Each cohort will employ a Simon two-stage design to test a null rate of 5% vs. 25% (power = 0.90, □ = 0.10). If > 1 of 9 patients in stage 1 achieve a partial response, an additional 15 patients are enrolled and if > 3 patients achieve a durable response, the combination will be worthy of further investigation. The primary endpoint is durable (> 6 months) response rate. Secondary endpoints are overall response rate, progression free survival, overall survival, duration of response and adverse events. Biopsies performed at baseline and after 3 weeks of therapy will assess the change in the density of tumoral CD8+ T-cells. Whole exome sequencing and immune cell profiling will also be performed on serial samples. Enrollment was initially limited to KRAS-mutant NSCLC. Prespecified activity goal for the first stage of accrual has been met; second stage accrual began in September 2019. Enrollment to the KRAS wild-type cohort will commence. Clinical trial information: NCT03600701.