Abstract Background Recurrence after surgery for early-stage lung cancer is common, occurring between 30-50% of the time. Despite the popularization of prognostic gene signatures in early-stage lung cancer that allow us to better predict which patients may recur, why patients recur after surgery remains unclear. Methods Using a large cohort of lung adenocarcinoma patients with complete genetic, genomic, epigenetic and clinical profiling, a recurrence classifier was developed which identifies patients at highest risk of recurrence. The genetic, genomic, and epigenetic profiles of stage I patients with low-vs. high-risk of recurrence were compared. To characterize the tumor immune microenvironment of recurrent stage I tumors, single cell RNA-seq was performed on fresh tissue samples undergoing lung adenocarcinoma resection at UCSF to identify unique immune population markers and applied to the large stage I lung adenocarcinoma cohort using digital cytometry. Results Recurrence high-risk stage I lung adenocarcinomas demonstrated a higher mutation burden than low-risk tumors, however, none of the known canonical lung cancer driver mutations were more prevalent in high-risk tumors. Transcriptomic analysis revealed widespread activation of known cancer and cell cycle pathways with simultaneous downregulation of immune response pathways including antigen presentation and Th1/Th2 activation. Tumors at high-risk of recurrence displayed depleted adaptive immune populations, and depletion of adaptive immune populations was independently prognostic of recurrence in stage I lung adenocarcinomas. Conclusion Recurrent stage I lung adenocarcinomas display distinct features of genomic and genetic instability including increased tumor mutation burden, neoantigen load, activation of numerous mitotic and cell cycle genes, and decreased genome-wide methylation burden. Relative depletion of infiltrating adaptive immune populations may allow these tumors to escape immunosurveillance and recur after surgery.