Skin eruption shape-based pathological state inference for personalized treatment in chronic spontaneous urticaria

Summary: Abstract Body: Urticaria is a common skin disorder characterized by wheals that appear in various shapes. Chronic spontaneous urticaria (CSU), a major subtype which can persist for years or even decades, significantly impacts patients' QOL. Although it is well accepted that urticaria symptoms are induced by various mediators released from skin mast cells, including histamine, the mechanism of CSU remains elusive, largely due to the lack of animal models and specific clinical biomarkers. To address this, we have developed a novel systematic framework capable of inferring pathophysiological states from the shape of skin eruptions and linking them to patient-specific treatments. We first developed a mathematical model to capture the wheal dynamics in CSU, based on pathomechanisms inferred from in vitro experimental results. This mathematical model successfully reproduced the spatiotemporal dynamics of actual wheal patterns in CSU patients, and we validated it through a clinical study that compared live images of wheal dynamics with in silico data. To infer a patient’s pathological state from imaging data of wheal, we constructed a novel mathematical tool that quantitatively captures the geometrical features of wheals by integrating topological data analysis with the mathematical model. Using this tool, we successfully extracted not only the geometrical features of patient's wheals but also patient-specific parameter sets that reflect their in vivo pathological state. With the mathematical model incorporating these patient-specific parameters, we predicted the efficacy of H1-antihistamines. This approach provides a novel framework for dermatology, enhancing diagnostic accuracy and predicting treatment efficacy by leveraging features of skin eruption shape, integrating mathematical modeling, data analysis, and accessible clinical data. Sungrim Seirin-Lee³, Takahiro Hiraga³, Hiroshi Ishii¹, Daiki Matsubara², Ryo Saito², Shunsuke Takahagi², Michihiro Hide² 1. Hokkaido Daigaku, Sapporo, Hokkaido, Japan. 2. Hiroshima Daigaku, Higashihiroshima, Hiroshima, Japan. 3. Kyoto Daigaku, Kyoto, Kyoto, Japan. Bioinformatics, Computational Biology, and Imaging