PD282 - RESULTS FROM A SYSTEMATIC REVIEW ON ARTIFICIAL INTELLIGENCE IN CLINICAL NUTRITION: A NEW PARADIGM IN THE ERA OF PERSONALIZED NUTRITION

View poster

Linked sessions

PD282

RESULTS FROM A SYSTEMATIC REVIEW ON ARTIFICIAL INTELLIGENCE IN CLINICAL NUTRITION: A NEW PARADIGM IN THE ERA OF PERSONALIZED NUTRITION

 

S. Ebrahimpour-Koujan1,*, M. Aghasi1, A. Haghighi2

1Department of Clinical Nutrition, Tehran University of Medical Sciences, Tehran, Iran, Islamic Republic Of, 2Department of Medicine, Brigham and Women’s Hospital and Harvard Medical School, Boston, United States

 

Rationale: Artificial intelligence (AI) has reached all-time-high in critical aspect of patient care and it has the potential to revolutionize clinical nutrition. This review focused on the possibilities of AI application in clinical nutrition toward precision nutrition. Moreover, utilization of healthcare database to develop deep learning and machine learning (ML) algorithms on clinical nutrition screening, assessment, prediction of clinical events and outcomes were discussed.

Methods: We searched published articles using relevant keywords in PubMed, Scopus, google scholar by April 1th, 2026.  

Results: Classically, AI and ML algorithms can be applied to the all aspects of clinical nutrition regarding patients' nutritional requirements and management as well as risk prediction. Therefore, it enables clinicians to make evidence-based nutritional recommendations. In oncology setting, combining ML in practice improves screening tools and identifies malnourished cancer patients or obesity using large databases. Interestingly, in ICU patients, it has been able to predict enteral feeding intolerance, diarrhea, or refeeding hypophosphatemia as well as response to treatment and overall mortality. Accordingly, the outcome of patients can also be improved. As microbiota and metabolomics profiles are better integrated with the clinical condition, it can effectively manage using ML. However, implementing AL and ML in clinical nutrition cause several ethical and regulatory limitations regarding data policy and bias. 

Conclusion: AI strongly supports and targets personalized clinical care. It can potentially transform to clinical nutrition. However, further researches are needed to assess the effectiveness, adaptation, patients’ safety and benefits of all AI-powered clinical nutrition efforts.

Disclosure of Interest: None declared