Abstract
Introduction: The development of computer-aided diagnostic solutions based on Artificial Intelligence (AI) has enabled the diagnosis of skin cancer. Objective: To discuss advances in dermatoscopic and AI-based digital image solutions for the diagnosis of skin cancer, along with some future challenges and opportunities to improve these AI systems, in order to support dermatologists and increase their ability to diagnose skin cancer. Methods: The systematic review guidelines of the PRISMA Platform were followed. The search was conducted from September to October 2025 in the Web of Science, Scopus, PubMed, Science Direct, SciELO, and Google Scholar databases. The quality of the studies was based on the GRADE and AMSTAR-2 instruments, and the risk of bias was adequately analyzed according to the Cochrane instrument. Results and Conclusion: 87 articles were found. A total of 18 articles were qualitatively evaluated, and 13 were included and developed in this systematic review study. Considering the Cochrane tool for risk of bias, the overall assessment resulted in 5 studies with a high risk of bias and 50 studies that did not meet the GRADE and AMSTAR-2 criteria. It was concluded that accumulating evidence suggests that computer-aided diagnostic systems can offer their greatest benefit as assistive systems, as studies indicate that the combination of humans and machines achieves the best results. Artificial intelligence-based diagnostic systems are capable of detecting morphological characteristics quickly, quantitatively, objectively, and reproducibly, thus providing a more objective analytical basis to complement medical expertise.