Background:

The increasing prevalence of metabolic disorders, such as type 2 diabetes (T2D) and metabolic-associated fatty liver disease (MAFLD), require innovative approaches to dietary management. Traditional dietary interventions often overlook individual variability in metabolic responses, leading to non-optimal outcomes.

Methodology:

Recent studies have showed the integration of artificial intelligence (AI) in developing personalized nutrition plans. An important example is the application of Digital Twin (DT) technology, which uses machine learning models to predict individual postprandial glycemic responses (PPGRs) and adapt dietary recommendations accordingly. In a randomized controlled trial, 319 participants with T2D were assigned to either a DT-enabled personalized nutrition group or a standard care group. The DT group received AI-generated meal plans designed to minimize PPGRs, while the standard care group followed general dietary guidelines. The primary outcomes examined included changes in hemoglobin A1c (HbA1c) levels and medication usage. Secondary outcomes included liver function markers, body weight, and cardiovascular risk factors.

Results:

The DT group demonstrated a significant reduction in HbA1c levels (-2.9%) compared to the standard care group (-0.3%), with 72.7% of participants achieving T2D remission. Additionally, improvements were observed in liver fat scores, body weight, and blood pressure. Notably, 94% of the DT group discontinued T2D medications after one year.

Conclusions:

AI-driven personalized nutrition, as shown by DT technology, offers a promising strategy for managing metabolic disorders. By accounting for individual metabolic profiles, AI can optimize dietary interventions, leading to improved clinical outcomes and reduced reliance on pharmacological treatments. Further research is required to validate these findings across diverse populations and to explore the long-term benefits of AI-assisted dietary management.