Wednesday, December 3 | 8:40 a.m.-8:50 a.m. | W1-SSNPM03-5 | Room S402
For radiology department leaders interested in sustainable imaging practices, AI-powered MRI offers a "green pathway" for routine clinical workflows, according to details of this Wednesday morning session.
The findings highlight the potential value of deep-learning MRI reconstruction techniques as a time- and energy-saving strategy.
"By optimizing acquisition efficiency while maintaining diagnostic accuracy, AI-enhanced MRI protocols represent a pivotal innovation for the future of eco-responsible radiology," presenter Tiziano Polidori, MD, and colleagues noted ahead of RSNA 2025.
The group conducted a retrospective study involving 148 patients divided into three groups: 45 undergoing upper abdomen MRI, 53 undergoing cardiac MRI, and 50 undergoing lumbar spine MRI.
Each patient underwent conventional and AI-assisted MRI acquisitions, with AI-enhanced protocols developed in collaboration with the vendor, according to the group. All were imaged using 1.5-tesla and 3-tesla scanners. The researchers evaluated energy consumption, greenhouse gas emissions, and scan time.
Importantly, Polidori's group reported "significant time savings per patient," as much as 58% in some upper abdomen MRI cases, for example, and about 50% within the other two imaging groups.
Beyond time saved per patient, energy reduction (kW/h) and carbon dioxide (CO2)-equivalent emissions were also recorded and statistically analyzed, with statistical significance established using paired t-tests with α = 0.05, according to the group.
Stop by to hear how the time savings reductions translated into energy savings.
