Artificial intelligence in trichology: A systematic review of current applications in diagnosis, severity assessment, and treatment monitoring.

Background: Artificial intelligence (AI) methods are increasingly used to assess hair and scalp disorders. However, the range of clinical tasks addressed, model performance, and methodological robustness across indications remains unclear. Methods: We conducted a systematic review following PRISMA guidance. PubMed/MEDLINE, Web of Science, Scopus, and Cochrane CENTRAL were searched from inception to December 2025, with additional reference screening. We included interventional or observational studies that developed, validated, or clinically evaluated AI and machine learning (ML) models using scalp photography, dermoscopy/trichoscopy, or related imaging modalities and reported extractable performance or agreement outcomes. Risk of bias was assessed using QUADAS-2 for diagnostic studies, PROBAST+AI for prediction models, and the NIH/NHLBI tool for the single before-after trial. Due to methodological heterogeneity, findings were synthesized narratively. Results: From 2,256 records, 15 studies published between 2020 and 2026 met the inclusion criteria, with seven published in 2025 or 2026. Thirteen were retrospective image analysis or model development studies, and two were prospective investigations. Seven studies addressed diagnostic classification, and six focused on severity scoring or quantitative assessment; prognosis and treatment response were each evaluated in one study. Deep learning approaches, predominantly transfer-learned convolutional neural networks, were most common, alongside segmentation frameworks and occasional traditional ML algorithms. Reported performance was generally high in controlled datasets and structured tasks, including strong agreement for automated Severity of Alopecia Tool (SALT) estimation and robust discrimination between psoriasis and seborrheic dermatitis. Conclusions: AI demonstrates promising potential for scalp disease diagnosis and objective severity quantification, particularly for automated SALT estimation and patterned hair-loss assessment. However, the current evidence base is dominated by retrospective single-source datasets with limited external validation, restricting confidence in generalizability and clinical implementation.