Background and purpose: To assess cost-effectiveness of late time-window endovascular treatment (EVT) in a clinical trial setting and a real-world setting. Methods: Data are from the randomized ESCAPE trial and a prospective cohort study (ESCAPE-LATE). Anterior circulation large vessel occlusion patients presenting > 6 hours from last-known-well were included, whereby collateral status was an inclusion criterion for ESCAPE but not ESCAPE-LATE. A Markov state transition model was built to estimate lifetime costs and quality-adjusted life-years (QALYs) for EVT in addition to best medical care vs. best medical care only in a clinical trial setting (comparing ESCAPE-EVT to ESCAPE control arm patients) and a real-world setting (comparing ESCAPE-LATE to ESCAPE control arm patients). We performed an unadjusted analysis, using 90-day modified Rankin Scale(mRS) scores as model input, and analysis adjusted for baseline factors. Acceptability of EVT was calculated using upper/lower willingness-to-pay thresholds of 100,000 USD/50,000 USD/QALY. Results: Two-hundred-forty-nine patients were included (ESCAPE-LATE:n=200, ESCAPE EVT-arm:n=29, ESCAPE control-arm:n=20). Late EVT in addition to best medical care was cost-effective in the unadjusted analysis both in the clinical trial and real-world setting, with acceptability 96.6%-99.0%. After adjusting for differences in baseline variables between the groups, late EVT was marginally cost-effective in the clinical trial setting(acceptability:49.9%-61.6%), but not the real-world setting(acceptability:32.9%- 42.6%). Conclusion: EVT for LVO-patients presenting beyond 6hours was cost-effective in the clinical trial setting and real-world setting, although this was largely related to baseline patient differences favoring the real-world EVT group. After adjusting for these, EVT benefit was reduced in the trial setting, absent in the real-world setting.