Develop an AI Food Allergy Diagnostics Platform for Patient Care and Treatments
Funded by Food Allergy Fund
Despite significant advances in Artificial Intelligence (AI) and Machine Learning (ML) for medical diagnostics, current models for food allergy diagnosis have not yet fully integrated multimodal data, limiting their effectiveness and precision. Multimodal data encompasses diverse sources of information, including genetic data, immunological markers, electronic health records, patient-reported symptoms, dietary logs, and environmental factors. Most AI and ML models have traditionally focused on isolated datasets, such as specific IgE levels or skin prick test results, without considering the rich, interconnected nature of these data types. The lack of integration leads to a fragmented understanding of food allergies, potentially overlooking complex interactions and contextual factors critical to accurate diagnosis. Incorporating multimodal data could significantly enhance AI models' diagnostic capabilities, enabling a more holistic view of a patient's condition and more accurate predictions. This integration would require sophisticated data fusion techniques and robust computational frameworks, presenting a substantial challenge but also a promising frontier for future research and development in food allergy diagnostics.
We aim to develop a multimodal AI platform for food allergy diagnostics that integrates diverse data sources, including skin tests, blood tests, molecular data, histological data, and medical history. This platform will leverage advanced machine learning algorithms to analyse and correlate these varied datasets, providing a comprehensive and nuanced understanding of food allergies. Initial findings suggest that our pipeline, which integrates physiological markers across multiple data sources, can predict oral food challenge outcomes months in advance of the challenge. In addition, our causal learning model improves decision-making by informing recommendations on whether to continue or discontinue oral immunotherapy. By incorporating the different data modalities, the AI system will identify complex patterns and interactions that single-modal approaches might miss. This integration will enable more accurate and personalized diagnostics and allow for earlier and more precise identification of food allergies.
