Diagnosing contact dermatitis using machine learning: A review.

Area of Science:

• Dermatology
• Medical Informatics
• Artificial Intelligence

Background:

• Machine learning (ML) presents a significant opportunity to advance contact dermatitis (CD) research.
• Integrating ML with comprehensive clinical data may enhance diagnostic capabilities and patch test precision.

Purpose of the Study:

• To conduct a comprehensive review of existing literature on the application of ML in contact dermatitis.
• To explore the potential of ML in supporting diagnosis and improving patch test accuracy in CD.

Main Methods:

• A systematic literature search was performed across Embase, Medline, IEEE Xplore, and ACM Digital Library up to February 7, 2024.
• Included studies focused on primary research reporting ML models applied to contact dermatitis.

Main Results:

• Six studies met the inclusion criteria after reviewing 7834 identified articles.
• ML models were utilized to identify biomarkers, differentiate between allergic contact dermatitis (ACD) and irritant contact dermatitis (ICD) using image and clinical data, and predict patch test positivity.
• Supervised learning was employed across all models, encompassing a total of 49,704 patients, though accuracy reporting was sparse.

Conclusions:

• Despite limited current research, ML demonstrates potential as a supportive tool for clinical diagnosis in contact dermatitis.
• Further investigation into the clinical implementation and validation of ML in CD management is warranted.