Abstract / Summary
Abstract Background Lenvatinib (LEN) is a standard first-line therapy for unresectable hepatocellular carcinoma (HCC), yet therapeutic outcomes are compromised by pharmacokinetic variability and resistance. The biological mechanisms linking sub-therapeutic exposure to metabolic and immune adaptations remain elusive. Methods We conducted a prospective multi-omics study involving 194 HCC patients. Plasma drug concentrations were correlated with clinical outcomes. Resistance mechanisms were interrogated using untargeted plasma metabolomics, in vitro cell validation, exosomal miRNA sequencing, multiplex immunohistochemistry (mIHC), and single-cell RNA sequencing. Results A C trough of 26.5 ng/mL was identified as the candidate threshold for predicting objective response rate. Furthermore, a combined model integrating this concentration threshold with tumor size demonstrated superior predictive performance across the training (AUC: 0.884), internal (AUC: 0.868), and external (AUC: 0.873) cohorts. Uric acid was the only clinical marker significantly correlated with LEN exposure ( r = 0.657, P < 0.001). Metabolomic profiling revealed that sub-therapeutic LEN exposure was characterized by distinct lipid metabolic reprogramming and the accumulation of acylcarnitines. Spatially, this sub-optimal exposure was associated with a “hot but suppressed” tumor microenvironment, characterized by high CD8 + T-cell and PD-L1 + densities, alongside the spatial colocalization and entrapment of T cells by CD163 + macrophages. The scRNA-seq further identified CD8 + T cells as the primary intercellular communication hub, linking metabolic shifts to spatial immune activation. Conclusion Optimal LEN efficacy relies on maintaining plasma concentrations > 26.5 ng/mL. Sub-therapeutic exposure is closely associated with resistance, potentially involving specific lipid metabolic reprogramming and macrophage-mediated spatial exclusion of T cells. These findings advocate for therapeutic drug monitoring and combinatorial strategies targeting metabolic-immune axes.