Background: In May of 2017, the U.S. FDA granted accelerated approval for pembrolizumab + pemetrexed & carboplatin for previously untreated advanced non-squamous NSCLC—the only first-line indication for immunotherapy + chemotherapy (‘IO+Chemo’) in NSCLC. Preliminary value estimates are needed to inform decision making. Using data from the Phase II KEYNOTE-021 trial, we estimated the potential value of IO+Chemo from a U.S. payer perspective. Methods: We created a partitioned survival decision model to assess the cost-effectiveness of IO+Chemo vs. pemetrexed & carboplatin (‘Chemo’) in Stage IIIB/IV non-squamous NSCLC. One year overall (OS) and progression-free survival came from KEYNOTE-021 and were extrapolated with parametric curves. The base case used a Weibull curve for long-term OS (6% 5 year OS), and scenarios used Log-Logistic (21% 5 year OS) and Gompertz ( < 1% 5 year OS) curves to model more and less durable responses. First- and second-line therapy resource use and Grade 3/4 adverse event rates were derived from KEYNOTE-021. We applied 2017 Average Sales Price for drugs and 2017 CMS reimbursement for procedures. Utilities were derived from the literature. We estimated life years (LY), quality-adjusted life years (QALYs), and costs over a lifetime horizon. Outcomes were discounted at 3% per year. Results: In the base case, IO+Chemo and Chemo resulted in 2.23 and 1.43 LYs, 1.20 and 0.77 QALYs, and $328,640 and $147,418 cost, respectively. The IO+Chemo costs per LY and QALY gained were $227,149 and $422,313, respectively. The cost per QALY varied greatly in OS scenarios (See Table), showing the strong influence of durable responses on value. Conclusions: In the first cost-effectiveness analysis of frontline IO+chemo in advanced non-squamous NSCLC, we found that pembrolizumab-based therapy is unlikely to be cost-effective (ie < $150,000 per QALY) regardless of long-term OS assumptions. Future studies should reassess IO+chemo value with Phase III KEYNOTE-189 data.