A team of researchers from the Keio University Medical School, Kyoto Prefecture and Teikyo University, in collaboration with Atopiyo LLC, has developed a new model of artificial intelligence (AI) that can objectively evaluate the seriousness of eczema using images of smartphone. Their findings were recently published in AllergyThe official magazine of the European Academy of Allergy and Clinical Immunology.
Atopic dermatitis (AD) is a common time of skin that blinks repeatedly and often requires long -term surveillance and treatment arrangements. In recent years, smartphone applications and social media platforms have made it easier for patients to monitor their symptoms, learn about their condition, and change the changes of documents over time. However, the symptoms reported by the patient, such as itching or sleeping loss, are not always aligned with the visible severity of the disease. This gap emphasizes the need for more standardized, objective evaluation tools-and emphasizes the growing potential of digital biomarkers to cover this role.
To deal with this, the research team has used data from Atopiyo, Japan’s largest advertising platform, where more than 28,000 users have shared over 57,000 symptoms photos and personal comments since 2018.
Using a setting of 880 images with self -reported itching ratings, the AI model showed high diagnostic accuracy. In a validation study using 220 test images, the TIS (AI-TIS) based on AI showed a strong correlation with a dermatologist-based TIS ratings (R = 0.73, p <0.001) and a significant correlation with objective scores (R = 0.53, p = 0.04).
Many eczema patients are struggling to evaluate the severity of their disease on their own. The AI model allows us to objectively monitor real -time using only one smartphone, strengthening patients and possibly improving disease management. “
Dr. Takeya Adachi, the corresponding author of the study
Interestingly, the study also found that the ratings of severity derived from AI are only weakly associated with self -reported itching scores, emphasizing the divergence between perceived symptoms and observable inflammation. This supports the increasing need for digital biomarkers that can enhance accuracy in dermatological care.
Researchers aim to extend the utility of the model by incorporating wider skin types, age areas and incorporating additional clinical features from scoring systems such as Scorad and EASI. This project paves the way for viewing views based on AS that support both patients and clinicians in real arrangements.
Source:
Magazine report:
Okata-Karigane, U., et al. (2025) The evaluation of the objective severity of the A -of -atopic dermatitis using patients in actual arrangement: a digital biomarbon approach. Allergy. doi.org/10.1111/all.16586.
