AI14 Artificial intelligence-driven digital applications and tools for atopic dermatitis: understanding the current landscape

Abstract Atopic dermatitis (AD; synonymous with eczema) is a chronic, relapsing inflammatory skin condition. It affects approximately 20% of paediatric patients and up to 10% of adult patients in high-income countries. It is characterized by severe pruritus and recurrent skin lesions, significantly impairing quality of life and contributing to sleep disturbances, mental health issues and school absenteeism. The challenges of managing AD underscore the need for innovative technological solutions. Artificial intelligence (AI) and computational tools are emerging as transformative technologies in the diagnosis, monitoring and management of AD. In this scoping review, we systematically assessed studies on AI-driven digital applications for AD. In total 3122 papers were identified through searches in Medline, Embase, Web of Science and Scopus. After our initial search screening, 136 studies met our criteria, categorized as follows: diagnostic tools (102), symptom tracking tools (23), predictive tools (17), teledermatology platforms (9), personalized treatment tools (8) and natural language processing or large language models (3). It is worth noting that some papers developed multiple AI tools. Currently, many AI-driven imaging tools utilize convolutional neural networks to automate severity scoring from digital images of AD lesions. They provide an automated version of traditional scoring methods such as the Eczema Area and Severity Index or SCORing Atopic Dermatitis. ChatGPT and similar AI-based chatbots are being utilized to answer patient queries and support clinical decision making, by providing evidence-based information on managing AD and addressing treatment-related concerns. While AI advancements in AD management show promise, challenges remain, including data privacy, algorithmic bias and clinical integration barriers. Additionally, real-world testing is crucial to ensure reliability, accuracy and usability across diverse populations. To maximize impact, AI tools must prioritize patient care, clinical decision making, and both clinical and AI best practices, throughout development. These innovations should focus on practical utility, seamless integration and improvements in patient outcomes. To achieve these goals, collaborative efforts are needed between academic, clinical and industry experts to accelerate the development, validation and optimization of AI-driven tools for AD management.