Evaluation of Facial Vitiligo Severity with a Mixed Clinical and Artificial Intelligence Approach

Area of Science:

• Dermatology
• Medical Artificial Intelligence
• Computer Vision

Background:

• Vitiligo is a prevalent depigmenting skin condition.
• Accurate assessment of vitiligo surface area is crucial for evaluating new therapies.
• Current scoring methods can be time-consuming and subjective.

Purpose of the Study:

• To develop and validate an artificial intelligence (AI) system for assessing facial vitiligo severity.
• To provide an objective and efficient tool for vitiligo evaluation.
• To compare AI assessment with expert physician scoring.

Main Methods:

• An AI model was trained and validated using 100 facial vitiligo images.
• A final dataset of 69 images was used for evaluation.
• Three expert physicians scored the 69 images, with inter- and intra-rater reliability assessed.
• AI performance was compared against physician scores.

Main Results:

• The AI system achieved 93% accuracy in assessing facial vitiligo severity.
• Good agreement was observed between the AI model's scores and those of expert physicians.
• The AI demonstrated reliable performance in evaluating vitiligo extent.

Conclusions:

• The developed AI system offers a potential objective evaluation for facial vitiligo.
• This AI tool could serve as a complementary or alternative to human assessment in clinical settings.
• The findings highlight the utility of AI in dermatological assessments and research.