L’impiego dell’intelligenza artificiale (IA) nella diagnosi del melanoma cutaneo

Artificial intelligence (AI) is reshaping melanoma diagnosis through convolutional neural networks (CNNs), achieving accuracy comparable to, and sometimes exceeding, that of expert dermatologists. Prospective studies and systematic reviews highlight the benefits of Augmented Intelligence, combining algorithmic output with clinical judgment to enhance sensitivity and specificity. Nonetheless, important limitations persist, including dataset bias, underrepresentation of darker skin types, and insufficient external validation. Ethical, regulatory, and trust-related issues remain critical. Overall, AI represents a promising support tool in dermatology, but requires transparency, robust clinical validation, and specialist supervision to ensure safe and responsible integration into clinical practice.