Skin Cancer Detection through Mobile Applications using Artificial Intelligence: Technical Review

Artificial intelligence-powered mobile applications are transforming the early detection of skin cancer, offering an accessible and accurate alternative to traditional methods. This study aims to evaluate the effectiveness, impact, and limitations of these tools through a systematic review of scientific literature. Using a mixed qualitative-quantitative approach, 30 articles were analyzed from the Scopus database, adhering to stringent inclusion and exclusion criteria. The findings reveal that applications leveraging convolutional neural networks achieve up to $92 \%$ accuracy in classifying skin lesions, showcasing significant potential to enhance diagnostic outcomes. Additionally, these tools have demonstrated a positive impact on early detection, particularly in regions with limited access to dermatology specialists. However, key limitations were identified, including the lack of clinical validation across diverse settings, challenges in adoption among older and vulnerable populations, and the need for educational programs to build user trust in these technologies. Equity challenges also emerged, as the reliance on smart devices and internet connectivity restricts the reach of these applications in underserved communities. In conclusion, while these technologies represent a promising advancement in medical diagnostics, their effective implementation requires addressing barriers related to validation, adoption, and education to ensure their seamless integration into healthcare systems.