MRI and sustainability

Reducing energy consumption without compromising 

Philipp Grätzel von Grätz
Published on January 30, 2023

MRI machines are power-hungry, no secret about that. But this doesn‘t mean that their appetite cannot be tamed. Average energy consumption of MRI units has decreased since the early years of the technology. And there are some promising avenues currently being explored that could enhance energy efficiency further in the years to come.

Among modern imaging modalities, MRI is probably the most power-hungry technology that is regularly used in clinical routine. “The main reason for this is that an MRI machine needs energy 24 hours a day, 7 days a week,” according to Sören Grübel, Team Lead Hardware Development for Magnetic Resonance Imaging (MRI) at Siemens Healthineers in Erlangen, Germany.

While a CT machine can be shut down and won’t require relevant amounts of energy in sleep mode, the fact that the MRI magnet needs to be cooled constantly means that energy is required more or less constantly: “We are talking about 6 to 7 kilowatts (kW) that we need as background energy all the time, plus additional energy once the machine turns into productive mode.”

When it comes to discussions about MRI sustainability, the relevance of this background energy consumption should not be underestimated. The exact amount obviously depends on the type and strength of the magnet used.

Energy consumption of MRI machines should be reduced, this is something all MRI producers can easily agree on. One problem is that it is not easy to standardize the measurement of energy consumption in a way that makes the scanners of various providers or even different generations of devices in different medical institutions comparable. According to Grübel, there are, for example, MRI units in a country such as Turkey that diagnose more than 100 patients per day, meaning that the machine is in productive mode not only during the day, but more or less around the clock. The type of examination an MRI is used for also has to be factored in. There are relevant differences, says Grübel:

“Standard examinations might require about 25 kW during scanning. More demanding examinations that use energy intensive sequences can go up to 70 kW or even 80 kW.”
Constant cooling of an MRI magnet means constant energy is required.
A typical example for a very energy-demanding examination are head examinations with echo-planar imaging (EPI). Certain examinations of the knee that don’t use standard orientations but need special angles are also quite taxing in terms of power consumption. Given all this: How do you measure MRI energy efficiency in a standardized fashion? An organization that has addressed this exact question is the European Coordination Committee of the Radiological, Electromedical and Healthcare IT Industry (COCIR), which brings together major medical technology manufacturers. COCIR initiated a self-regulatory initiative (SRI) for the eco-design of medical imaging equipment in 2011. As part of it, a standardized methodology was developed that can be used to measure energy consumption and make it comparable across manufacturers and device generations.

The SRI’s most recent report from 2018 shows that there has been a certain amount of success in making MRI more sustainable. According to the report, the daily average energy consumption per MRI scanner decreased from 226 kWh in 2011 to 165 kWh in 2017, a drop of a nearly 30 percent. The total daily energy consumption of MRI as a whole nevertheless increased by around 15 percent because the installed base grew considerably in these six years. So what exactly made the reduction in energy consumption per unit possible, and what other possibilities could be conceived to improve energy efficiency even further? A big chunk of the improvement is due to what manufacturers refer to as economic power mode or EPM. This acts on the level of the cooling system. It is about changing MRI refrigeration from continuous cooling to a mode in which the helium refrigeration compressor turns on and off in certain intervals mainly overnight, according to Simon Calvert.

There are some other levers to reduce energy consumption, says Grübel: “For example, when the table is in home position outside the bore, switching off the gradient power amplifiers will reduce energy consumption, and this is also possible during the day.” There are also other components that can be switched off when the table is in home position, specifically parts of the electronics. “We have introduced these features 10 years ago already, and it really has zero impact on the workflow for our customers. We are currently trying to shut down some more components, but the challenge is really to do that without affecting workflows. It all comes down to whether it is possible to switch components on and off fast enough. 

Customers don’t want clumsy workflows, and they certainly don’t want to compromise on quality.” Software can also help. Across the MRI industry, there is a trend towards shorter scanning times, optimized sequences and better post-processing, made possible by sophisticated digital tools that only became available in recent years. Artificial intelligence algorithms should help with further reducing scan times and thus energy consumption in productive mode in the years to come. Unfortunately, there is little reward in terms of publishable numbers for this kind of innovation, says Grübel: “COCIR doesn’t really look at these topics. It’s difficult to quantify the impact.”
Grübel sees some bigger trends that go beyond the individual MRI machine and its magnet: “I think we will have to rethink the integration of the MRI scanner into the building and its overarching energy infrastructure.” The catchword here is ‘smart buildings’, meaning that hospitals, especially big ones, do not only consume electricity and water, but also become, at least partly, producers. In such an ecosystem, digital regulation tools could model and forecast critical parameters like energy demand and necessary water supply and regulate producing capabilities against consuming capabilities in order to make the overall building more sustainable. Such a smart building would “know” when an MRI needs a lot of power and thus water; it could “check” whether this could be accounted for by boosting internal production capacities, or by rebalancing production and consumption. Doing so would increase, possibly substantially, the overall building efficiency, without having to compromise in an area in which compromise is not an option.

By Philipp Grätzel von Grätz

Philipp Grätzel von Grätz lives and works as a freelance medical journalist in Berlin. His specialties are digitalization, technology, and cardiovascular therapy.