Nutrition deficiency has a significant impact on the results of antitumor treatment of cancer patients. The empirical approach to prescribing nutritional support is quite subjective, which is why errors are possible when deciding on the need for, type and composition of artificial nutrition. At the same time, artificial intelligence is increasingly being introduced into real clinical practice, but its use in the field of nutritional support is very limited. We conducted a review of the literature on this topic in order to highlight the current state of the problem. The article presents and analyzes literary data from medical databases PubMed, E-library, Sciencedirect for 40 years on the use of artificial intelligence, namely machine learning algorithms for the early detection of protein-energy deficiency and predicting its development in cancer patients. It is shown that predictive models based on artificial intelligence, as well as models for identifying protein-energy malnutrition, can be integrated into systems for supporting medical decision-making, ensuring its timely diagnosis and correction, which will avoid subjectivity and limitations inherent in the traditional, "empirical" approach to prescribing nutritional support. Errors that are often encountered in the implementation of nutritional support in oncological practice are considered, and opportunities for their leveling with the help of artificial intelligence are proposed. Thus, despite large-scale prospects, the use of artificial intelligence tools in the process of identifying nutritional deficiency and implementing nutritional support is still limited.
Kukosh et al. (Fri,) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: