Urinary Proteomics Enables Noninvasive and Longitudinal Monitoring of Tumor Burden Dynamics and Early Recurrence in Hepatocellular Carcinoma

Abstract

To develop a noninvasive, urine-based approach for dynamic monitoring of tumor burden and early detection of recurrence in hepatocellular carcinoma (HCC), addressing the limited sensitivity of conventional serum biomarkers such as AFP and DCP, particularly for minimal residual disease (MRD) assessment. We established a prospective, multi-cohort urinary proteomics framework encompassing four longitudinal clinical cohorts (378 patients, 972 urine samples). In the discovery cohort, 26 patients contributed 130 longitudinal urine samples from those undergoing primary and secondary resections, which were analyzed by mass spectrometry at five standardized follow-up time points to identify proteins associated with tumor burden dynamics. The validation cohort (n = 46) used parallel reaction monitoring (PRM) to confirm candidate biomarkers and construct a composite urine-based tumor burden monitoring model integrating HPGD, AFP, DCP, and GGT. The model was then applied to an early recurrence cohort (306 patients, 612 urine samples) to detect MRD and predict recurrence prior to radiological confirmation. Among 8563 quantified urinary proteins, 217 significantly correlated with tumor burden, with HPGD closely mirroring dynamic changes. The integrated model achieved a pre-recurrence AUC of 0.86, sensitivity of 73%, and specificity of 87%, outperforming AFP (0.73, 39%, 96%) and DCP (0.64, 59%, 88%). It predicted recurrence a median 4.1 months earlier than imaging and served as an independent prognostic factor for recurrence-free (RFS) and overall survival (OS, p < 0.001). This urine-based model enables dynamic assessment of tumor burden and early recurrence detection, surpassing conventional serum biomarkers and providing a clinically actionable tool for personalized surveillance and therapeutic decision-making in HCC.