Inference of Pathological States from Skin Eruption Morphology for Drug Efficacy Evaluation and Personalized Treatment in Skin Disease

Objectives: Urticaria is a common skin disorder characterized by wheals that appear in various shapes and typically resolve within hours or a day. Chronic spontaneous urticaria (CSU), a major subtype that can persist for years or even decades, significantly impacts patients’ quality of life. While it is widely recognized that urticaria symptoms are caused by the degranulation of skin mast cells and the release of various mediators, such as histamine, the underlying mechanism of CSU remains poorly understood. This is primarily due to the lack of animal models and specific clinical biomarkers. To address these challenges, we have developed a novel framework capable of inferring pathophysiological states from wheal morphology and connecting them to patient-specific treatments. Methods: We constructed a mathematical model to capture the dynamics of wheals in chronic spontaneous urticaria (CSU), based on pathomechanisms derived from in vitro experimental results. This model was validated by reproducing the spatiotemporal dynamics of actual wheals observed in live imaging data from CSU patients. In addition, we applied topological data analysis to extract key morphological features of the wheals. Results: Using clinical data from 105 patients with chronic spontaneous urticaria (CSU), our analysis revealed that wheal patterns could be categorized into five distinct types. The diversity of these patterns was found to be associated with the underlying pathophysiological balance between coagulation states and histamine release from mast cells [ 1,2, 3]. To infer a patient’s in vivo pathological state from wheal imaging data, we developed a novel mathematical tool that integrates topological data analysis and AI with our mathematical model, enabling the quantitative assessment of wheal geometrical features. By applying this tool, we successfully extracted the pathophysiological states of CSU patients from their wheal morphology. Furthermore, we evaluated the efficacy of antihistamines for each wheal type and proposed a new potential therapeutic approach based on the detailed analysis of wheal shapes. Conclusions: This study introduces a groundbreaking framework in dermatology that integrates mathematical modeling, AI, and data analysis to enhance diagnostic accuracy, evaluate drug efficacy, and enable personalized treatment strategies based on skin eruption morphology.