Background: Artificial intelligence (AI) is rapidly influencing dermatology worldwide. India’s unique mix of dense urban tertiary centres and vast rural populations presents both opportunities and challenges for AI deployment
particularly mobile/low-cost tools that could extend diagnostic reach. The objective of this review was to assess the scope, quality, and clinical readiness of AI applications in dermatology in India, with emphasis on rural applicability.
Methods: We followed PRISMA guidance for systematic reviews. Databases searched included PubMed/PMC, Embase, Scopus, IEEE Xplore, and selected Indian journals and conference proceedings up to June 2026 (search terms:
“artificial intelligence”, “dermatology”, “India”, “mobile app”, “deep learning”, “dermoscopy”). Inclusion: original studies of AI applied to dermatologic diagnosis, triage, or monitoring with Indian datasets or Indian clinical deployment; exclusion:
non-dermatologic AI, editorial/opinion pieces without primary data. We extracted study design, dataset size, setting (urban vs rural), algorithm type, target conditions, performance metrics (sensitivity/specificity/accuracy), validation method, and
implementation features. Risk of bias was assessed using QUADAS-2 domains. Results (summary): The Indian literature includes a handful of robust, largesample validation studies of mHealth apps and CNN models with mixed results;
a landmark multi-site smartphone app study validated on ~5,000 patients demonstrated promising diagnostic breadth in clinical settings but variable performance across Fitzpatrick types and disease categories. Several pilot implementations and acceptability studies highlight strong clinician interest but raise concerns regarding skin-tone bias, dataset representativeness, and regulatory
Original Article
English
P. 7-13