Background: Immune checkpoint inhibition has shown early promising results in patients with chemotherapy-refractory metastatic or advanced tumors of the esophagus, gastroesophageal junction, and stomach. We explore the cost-effectiveness of checkpoint inhibitors as second-line treatment agents for this group of patients using a decision analytic approach. Methods: A Markov model was developed to simulate the course of a virtual cohort of patients treated by (i) nivolumab 3 mg/kg, (ii) combination of ipilimumab 3 mg/kg and nivolumab 1 mg/kg, and (iii) best supportive care (BSC). Patients in the hypothetical cohort were 55-year-olds in an advanced/metastatic stage who had received at least one prior line of chemotherapy. Patients who remained stable in treatment were monitored for adverse events until death. Rates of cancer-specific mortality, disease progression, and drug-related adverse events were estimated using results from the CheckMate 032 clinical trial. The primary endpoints were survival, measured in life-years (LY), quality adjusted life years (QALY), and incremental cost-effectiveness ratios (ICER). Cost-effectiveness of each strategy was evaluated from a US-payer perspective considering costs of drugs, treatment, and management of immune-related adverse events. Cost-effectiveness was defined with a willingness to pay threshold of $100,000/QALY. Results: Combination therapy with nivolumab and ipilimumab yielded the highest effectiveness (QALYs = 0.47, LYs = 1.09) in our base case modeling results, compared with nivolumab (QALYs = 0.43, LYs = 1.03), and BSC (QALYs = 0.19, LYs = 0.42). Nivolumab had an incremental cost of $84,555/QALY compared with BSC, while nivolumab with ipilimumab resulted in an incremental cost of $1.1M/QALY compared with nivolumab alone. The cost gap between the two was associated with the higher price of ipilimumab, and costs of managing increased toxicity. Conclusions: Our modeling analysis finds that combination therapy of ipilimumab and nivolumab is the most effective, but from a cost-effectiveness perspective, it is expensive, making nivolumab monotherapy the cost-effective option. Additional clinical data are needed to confirm our modeling results.