Artificial intelligence-based approaches for diagnosis of adverse reactions to food

Abstract

An increasing number of people worldwide suffer from adverse reactions to food (ARF). ARF can have both an immunological and a non-immunological background, which is relevant for both diagnosis and treatment. In everyday clinical practice, exact classification of ARF is sometimes challenging, as the symptoms can be relatively unspecific and overlap between ARF subgroups. In addition, some test systems frequently used in clinical routine have significant limitations. This concerns both their sensitivity and specificity as well as the relatively high resource demands. Use of artificial intelligence (AI) could represent a method to improve diagnosis of ARF in the future. Initial studies suggest that the use of AI can predict the individual risk of developing a food allergy as well as the allergic potential of new food proteins with a high degree of certainty. These and other examples of the successful use of AI applications in the diagnosis of ARF are encouraging and should provide an incentive for further studies.