Background: Liver cancer hasahigh recurrence rate of 50%-70% for early-stage patients. Minimal residual disease (MRD) is closely correlated with cancer early recurrence. Here, we developed SeekInCure, a non-invasive assay to detect MRD and predict the prognosis of liver cancer patients after radical surgery. Methods: 32 liver cancer patients undergoing radical surgery were prospectively enrolled. 8 mL peripheral blood was collected from each patient before and after surgery, respectively. SeekInCure assay, which integrated the protein tumor marker (AFP) and cancer genomic hallmarks: copy number aberration, and fragment size from ctDNA to calculate a score (P_HCC), was utilized to detect the cancer signal in each blood sample. MRD status determined by P_HCC value was correlated with the clinical outcome (OS) and the benefit of treatment after radical surgery. Results: Out of the 32 liver cancer patients, 78.1% (25/32) were at early stages (Stages I/II: 65.6%/12.5%), and 59.4% (19/32) received systemic treatment after surgery. MTB- patients (12.5%, 4/32) showed better overall survival (OS) than MTB+ patients (87.4%, 28/32) with a high hazard ratio (HR) value of 1.19. After radical surgery, 23 HCC patients were MRD-, including 4 MTB- patients remaining negative; meanwhile, 9 out of 28 MTB+ patients remained MRD+. The OS of MRD- patients was significantly better than that of MRD+ patients (HR 5.67; 95% confidence interval (CI), 1.33 ~ 24.20; P = 0.008). Moreover, the OS of MRD- patients with and without systemic treatment showed no significant difference (P = 0.930). Conclusions: This prospective study demonstrated the potential clinical value of ctDNA in detecting MRD status to predict the prognosis in early-stage liver cancer patients and to identify patients who may benefit from avoiding overtreatment. This strategy requires validation in independent prospective studies.