Artificial Intelligence in the Diagnosis of Atopic Dermatitis

Atopic dermatitis (AD) is the most common chronic inflammatory skin disease in high-income countries and is characterized by marked clinical heterogeneity and a relapsing course. Despite its high prevalence, the diagnosis of AD remains largely clinical and is often challenging due to overlapping features with other inflammatory dermatoses, the absence of definitive biomarkers, and variability in disease presentation across age groups and skin types. These challenges may lead to diagnostic delay or misclassification, with potential implications for disease management and patient outcomes. In recent years, artificial intelligence (AI), particularly machine learning and deep learning approaches, has emerged as a promising tool in dermatology. AI-based systems have demonstrated encouraging performance in diagnosing various skin diseases using clinical photographs, advanced imaging, molecular and laboratory approaches, and have increasingly been applied to AD, achieving high sensitivity and specificity in several studies. Such tools may support clinicians by enhancing diagnostic accuracy, assisting in differential diagnosis, and serving as effective screening instruments, especially in primary care or resource-limited settings. Importantly, AI is intended to complement rather than replace clinical expertise. This review summarizes the current literature on AI applications in the diagnosis of AD, critically evaluates existing models, and discusses their readiness for clinical translation, while highlighting key gaps that must be addressed to bridge research and routine clinical practice.