Abstract / Summary
Abstract Purpose Hepatitis B virus-related acute liver failure (HBV-ALF) is a highly lethal syndrome lacking precise early biomarkers to guide the timing of liver transplantation (LT). Aims We aimed to identify core drivers of hepatic synthetic collapse and explore novel prognostic biomarkers and targeted therapeutic strategies for HBV-ALF. Methods An integrated machine learning framework (LASSO, SVM-RFE) was applied to multi-cohort transcriptomic datasets to identify crucial diagnostic hub genes. Findings were rigorously validated in a real-world clinical cohort of HBV-ALF patients using serum biochemistry and spatial histopathology (H&E and IHC staining). Results We identified a profound disruption of the complement-coagulation crosstalk as a hallmark of HBV-ALF. Three critical hub genes—C8B, CPB2, and MASP2—were significantly downregulated. In clinical validation, CPB2 levels plummeted to approximately 26% of the normal baseline, strongly correlating with declining prothrombin time activity (PTA) and surging MELD scores. Spatial histopathology revealed a paradoxical microenvironment: massive necrosis with a near-complete loss of these locally synthesized proteins, accompanied by intense overactivation of the terminal complement complex (C5b-9) strictly bordering the surviving parenchyma. Conclusions The precipitous depletion of C8B, CPB2, and MASP2 serves as a highly sensitive real-time biomarker for irreversible hepatic synthetic exhaustion, offering a novel metric to refine LT indications. Furthermore, targeting the localized C5b-9 overactivation presents a promising therapeutic strategy to halt the necrotic cascade and serve as a "bridge to transplant" in HBV-ALF.