Advancing ultrasonography with AI: recent innovations and future perspectives

Abstract Recent advancements in artificial intelligence (AI) and deep learning have significantly impacted medical imaging, particularly in ultrasonography. AI applications in this field span a wide range of tasks, including tumor detection, segmentation, recurrence prediction, and radiomics. These technologies enhance diagnostic accuracy, improve workflow efficiency, and enable personalized treatment strategies. The integration of AI with ultrasonography extends beyond prediction, offering insights into new diagnostic factors through explainable AI (XAI). XAI enhances transparency by visualizing regions of interest used in predictions, addressing the black-box issue and providing feedback to the medical field. This dual utility as both a predictive and discovery tool holds promise for uncovering hidden diagnostic factors and advancing medical knowledge. As AI and ultrasonography evolve, their convergence is expected to transform clinical practice by improving diagnostic accuracy, enabling personalized care, and reducing healthcare provider workloads. These advancements could establish a new standard in medicine, delivering high-quality, accessible healthcare to a broader population.