Sustainable MRI: Energy and Environmental Benefits of Accelerated Imaging with Deep-Learning Reconstruction

Motivation: MRI significantly contributes to healthcare-related greenhouse gas emissions; therefore, reducing MRI energy use through accelerated imaging and deep-learning reconstruction (DLR) is essential. Goal(s): To quantify the energy savings and emissions reductions achievable using accelerated MRI combined with DLR, while ensuring minimal compromise in image quality. Approach: Energy consumption was measured on three 3T MRI scanners during standard and accelerated imaging with DLR, validated using ACR phantom tests and human subject imaging. Results: Accelerated MRI with DLR reduced scan times and energy consumption by 3-5 times without compromising diagnostic image quality, yielding potential institutional savings of 74.1 metric tons of CO₂eq emissions annually. Impact: Accelerated MRI with DLR enables significant clinical efficiency and sustainability improvements, facilitating rapid, high-quality imaging. Future studies can explore enhanced DLR algorithms, optimizing diagnostic performance while maximizing energy savings and emissions reductions across healthcare settings.