Investigation of stratum corneum cell morphology and content using novel machine‐learning image analysis

• DOI: 10.1111/srt.13565
• Publication Date: 2024-01-27T07:14:50Z
• Authors: Takeshi Tohgasaki; Saki Aihara; Mariko Ikeda; Minako Takahashi; Masaya Eto; Riki Kudo; Hiroshi Taira; Ai Kido; Shinya Kondo; Shioji Ishiwatari
• Keywords: Machine Learning; 03 medical and health sciences; 0302 clinical medicine; Artificial Intelligence; Humans; Original Articles; Epidermis; Biomarkers; Skin

Abstract Background The morphology and content of stratum corneum (SC) cells provide information on the physiological condition skin. Although morphological biochemical properties SC are known, no method is available to fully access interpret this information. This study aimed develop a comprehensively decode skin, based SC. Therefore, we established novel image analysis technique artificial intelligence (AI) multivariate predict skin conditions. Materials Methods samples were collected from participants, imaged, annotated. Nine biomarkers measured in using enzyme‐linked immunosorbent assay. data then used teach machine‐learning models recognize individual cell regions estimate levels nine images. Skin indicators (e.g., barrier function, facial analysis, questionnaires) or obtained participants. Multivariate including biomarker ​​and structural parameters as variables, was these indicators. Results We two models. accuracy recognition assessed according average intersection over union (0.613), precision (0.953), recall (0.640), F‐value (0.766). predicted significantly correlated with levels. questionnaire answers strong correlations correct answer rates. Conclusion Various conditions can be images AI analysis. Our expected useful for dermatological treatment optimization.