Artificial Intelligence and the Future of Environmentally Sustainable Radiology

Abstract

Background: Artificial Intelligence (AI) adoption in radiology has revolutionized diagnostic procedures by enhancing efficiency, accuracy, and clinical results. AI application, from image reception and reconstruction to report production, has optimized operations and reduced errors significantly linked to human interaction. Its environmental impacts, nonetheless, of releasing AI, especially the energy it demands for training and performing extensive models, are still insufficiently addressed. With the world's health systems inclining toward green policies and carbon neutrality, there is a need to assess if AI helps in reducing or adding to the environmental impact of radiology.

Aim: The aim of this research is to examine the role of AI in pro-environmental radiology in terms of its impact on energy consumption, carbon emissions, and resource usage in imaging procedures.

Methodology: A mixed-methods strategy was used that included a systematic review of peer-reviewed literature (2015–2024) and a comparative workflow analysis in a tertiary radiology department. AI-integrated workflows were contrasted with traditional practices through the use of key environmental metrics: energy use per exam, image processing time, dose of radiation, and medical waste (contrast media, use of film). Specific focus was placed on AI applications for automated protocol selection, low-dose reconstruction, and workflow automation.

Results: Integration of AI led to a 22% decrease in mean scan time and 15% lowering of contrast media usage through optimized protocols. AI-based reconstruction facilitated dose-reduction measures, reducing radiation exposure and energy usage by about 18% per scan. Yet the energy invested in training deep learning models was significant, emphasizing the need for green AI development methodologies. The results indicate a twofold effect: increased sustainability in clinical use but higher energy requirements in AI model development.

Conclusion: The AI is powerful in enabling environmentally friendly radiology through resource optimization. This can be achieved through embracing green development approaches and life cycle analysis.