Integrating Histological, Biochemical and Molecular Endpoints to Identify Predictors of Diet-Induced Hepatic Injury and the Protective Effects of Berry-Derived Bioactive Compounds

Diet-induced hepatic injury is a multifactorial pathological condition involving hepatic lipid accumulation, oxidative stress, inflammatory activation, mitochondrial dysfunction, and progressive hepatocellular damage. Conventional assessments frequently rely on isolated biochemical biomarkers or histopathological observations, potentially overlooking complex interactions among tissue morphology, metabolic disturbances, and molecular signalling pathways. This study proposes an integrated experimental and computational framework for identifying predictors of diet-induced hepatic injury and evaluating the hepatoprotective effects of berry-derived bioactive compounds. A novel Berry-Hepatic Multimodal Predictive Algorithm (BHMPA) is introduced to integrate quantitative histological, biochemical, and molecular endpoints through adaptive feature weighting, attention-based multimodal fusion, pathway-informed regularization, and explainable machine learning. The proposed experimental design incorporates normal-diet controls, high-fat/high-fructose diet-induced hepatic injury groups, and dietary intervention groups receiving standardized anthocyanin-rich blueberry, blackberry, or bilberry extracts at predetermined concentrations.