Predicted clinical impact of artificial intelligence in radiology : a rapid evidence assessment

Abstract

This Rapid Evidence Assessment (REA) reviews the literature sources related to the predicted clinical impact of Artificial Intelligence (AI) in radiology. It examines AI’s potential across five critical areas: diagnostic accuracy and early detection, treatment monitoring, workflow efficiency, patient safety, and the challenges faced by radiologists. By synthesizing evidence from peer-reviewed studies and incorporating insights from an expert panel, the REA provides a comprehensive evaluation of AI’s transformative capabilities and the practical considerations required for its implementation. The findings indicate that AI has the potential to significantly enhance radiology practice by improving imaging precision, enabling earlier disease detection, automating repetitive tasks, and personalizing treatment monitoring. These advancements address pressing challenges such as growing workloads, resource constraints, and radiologist burnout. However, the evidence base remains limited, as much of the research relies on retrospective studies or expert opinion, restricting generalizability. Challenges such as automation bias, ethical concerns, and interoperability issues are also identified, highlighting the need for cautious and deliberate integration of AI into clinical practice. Based on these findings, the REA recommends high-quality, multi-centre, prospective studies to validate AI’s applications and address current evidence gaps. It emphasizes the need for targeted training for radiologists to integrate AI effectively while maintaining their central role in patient care. Transparent and explainable AI systems are deemed essential to build trust and accountability, and a stepwise, scalable approach to AI implementation is suggested to align with clinical workflows and healthcare operations. The assessment concludes that while AI holds significant promise in transforming radiology, its clinical impact depends on evidence-driven implementation, ethical oversight, and strategic integration, ensuring that radiologists remain pivotal in leveraging AI to enhance accuracy, efficiency, and patient-centred care.