Efectividad de la inteligencia artificial en el diagnóstico médico por imágenes: Una revisión sistemática

This article aimed to analyze the effectiveness of artificial intelligence (AI) in medical imaging diagnosis through a systematic review of scientific literature published between 2021 and 2025. The integration of AI algorithms into clinical practice transformed diagnostic approaches across various medical specialties by improving the accuracy, efficiency, and speed of image analysis. The review was conducted following PRISMA guidelines and using databases such as PubMed, Scopus, EMBASE, Web of Science, and Science Direct. Ten relevant studies were selected, covering fields including ophthalmology, oncology, pulmonology, rare diseases, pediatric dentistry, and plastic surgery. The findings indicated that systems based on deep learning and machine learning achieved high levels of sensitivity, specificity, and positive predictive value, in some cases surpassing the performance of human specialists. However, significant limitations were identified, such as methodological heterogeneity, lack of standardized metrics, and the absence of robust ethical and regulatory frameworks. It was concluded that AI served as a valuable complementary tool in medical practice, with the potential to optimize imaging diagnosis and improve patient care. Nevertheless, its effective implementation depended on simultaneously addressing technical, ethical, educational, and structural challenges. This review provided a critical knowledge base to guide future research, clinical decisions, and public policy toward the responsible, safe, and humanized use of artificial intelligence in healthcare.