Scientific Rationale: Acute kidney injury (AKI) affects 30-50% of hospitalized liver cirrhosis patients, with mortality exceeding 60% in severe cases. Despite its clinical significance, AKI remains underdiagnosed in early stages due to limitations of traditional biomarkers like serum creatinine, which is influenced by reduced hepatic production and muscle mass in cirrhotic patients. Current diagnostic criteria often identify AKI only after significant renal damage, delaying critical interventions. The UK Biobank’s multi-modal data offers opportunities to identify novel predictors and develop artificial intelligence (AI)-driven predictive tools for early risk stratification.
Objectives: Identify potential novel genetic, proteomic and clinical predictors of AKI in cirrhosis; Develop and validate an AI model integrating multi-omics data to predict AKI risk 48-72 hours before clinical onset; Elucidate biological pathways linking proteomic biomarkers to AKI pathogenesis.
Research Questions:
1. Do genetic variants, proteomic biomarkers, and clinical factors synergistically improve AKI risk prediction in cirrhosis?
2. Can a multi-omics AI model outperform current clinical scores in sensitivity and specificity?
3. What biological pathways link proteomic biomarkers to AKI pathogenesis in cirrhosis?
Public health effects: Understanding the novel predictors will inform the development of novel intervention strategies, including screening for early detection and therapeutics for timely intervention and improved prognosis.
We would envisage this project taking between 12 and 36 months to complete!but we hope to produce initial results in the next 12 months.