Artificial Intelligence in medicine: Methods selection, applications and considerations (Part II)

Abstract

The spreading of artificial intelligence and machine learning (ML) methods in different healthcare areas is common. The second part of this review article describes the foresights when selecting different ML methods. It also presents an updated review of examples and the precautions or challenges that we will face in the future when using these technologies. We will describe how to know whether to use a descriptive or predictive approach, the characteristics of these methods and their potential applications. Later, we will discuss how the irruption of digital data, together with freely available algorithms and greater computational power, has made it possible to enhance the implementation of these models in medicine. We will review how ML has contributed to the development of diagnostic imaging, as well as the prediction of monitoring and clinical outcomes. Finally, we will analyze the challenges and ethical considerations associated with the implementation of ML in clinical practice.