Source-linked AI summary
Towards Sustainable Magnetic Resonance Imaging: Insights from long-term, high-resolution energy recordings across an entire scanner fleet
Florian Leonhard Raab, Fiona Mankertz, Nour Maalouf, Josephine Berger, Andreas Lingg, Reza Dehdab, Sebastian Werner, Judith Herrmann, Andreas Brendlin, Sebastian Gassenmaier, Suhas Siddaramu, Fabian Wagner, Julian Wohlers, Shreeja Varadarajan, Gurlal Singh, Jens Gühring, Rainer Schneider, Konstantin Nikolaou, Saif Afat, Thomas Küstner
TL;DR
MRI energy demand and its operational drivers remain insufficiently characterized, despite MRI’s energy intensity in healthcare. This study integrates energy recordings and system logs across seven scanners to examine use, utilization, and acquisition-related factors. Scan duration had a stronger relationship with energy than average power, and idle and eco-power modes together consumed more energy than active scanning.
Problem
Method
The study integrated scanner energy recordings with MRI system logs to characterize energy use and operational patterns across seven scanners over 1.5 years.
Source: §1 · p. 2, §2.1 · p. 3, §4 · p. 7Results
R2 was 0.769 for energy versus scan duration and 0.426 for energy versus average power; idle and eco-power modes together consumed more cumulative energy than active scanning.
Source: §4 · pp. 7–8Takeaways & Limitations
The integrated baseline provides a foundation for future data-driven research on energy-saving strategies and for further investigation of acquisition-related factors.
Source: §1 · p. 2, §5 · p. 9The retrospective study covered seven scanners from one manufacturer at a single hospital, and network interruptions caused incomplete measurements and data loss.
Source: §4 · pp. 7–8
Abstract
from arXiv · showhide
Magnetic resonance imaging (MRI) is among the most energy-intensive diagnostic modalities in healthcare, yet its energy consumption and the factors influencing it remain insufficiently understood. This study aims to establish a comprehensive baseline of MRI energy consumption by characterizing energy demand across a scanner fleet, examining scanner utilization and operating patterns in clinical practice. Concurrently, it investigates the relationships between energy consumption and relevant operational and acquisition features. On average, a single MRI measurement consumed 0.43 kWh, while a complete examination consumed 13.50 kWh. In general, substantial differences in energy consumption were observed between MRI scanners and their corresponding operating modes (scan, idle, and eco-power mode). These variations may be related to differences in scanner operating patterns, employed examination protocols, and their resulting acquisition parameters. Idle and eco-power modes accounted for more cumulative energy consumption than active scanning. However, these energy shares should always be interpreted in relation to scanner occupancy, as utilization patterns strongly influence the distribution of energy across those operating modes. Lastly, linear regression analysis showed that energy consumption was more strongly associated with scan duration than with average power, suggesting that scan duration may be an important factor influencing MRI energy consumption. ...
1 Introduction
Healthcare’s substantial energy and emissions footprint makes sustainability an operational concern, and MRI is a major energy user within radiology. Prior MRI studies have examined particular operating factors or interventions, while this study links energy demand with scanner utilization, examinations, protocols, and acquisition parameters across a seven-scanner fleet.
1 Introduction
- Healthcare accounts for about 4.4% of global net emissions, while radiology imaging systems can use up to 7.3% of a hospital’s electricity.MRI scanners represent the largest energy share among imaging modalities, and sequence-design or software-based changes may avoid major structural modifications.Source: §1 · p. 2
- Hospitals face rising operating costs alongside pressure to reduce their environmental footprint, while conventional building-efficiency projects can be costly and constrained.Radiology offers an energy-optimization opportunity within hospital operations.Source: §1 · p. 2
- MRI energy studies have mainly examined individual operating factors or interventions, leaving a need to connect scanner demand with utilization, examinations, protocols, and acquisition parameters.The study analyzes seven scanners using one-second data over 1.5 years to establish a hospital baseline.Source: §1 · p. 2
2 Methods
The study combines high-resolution scanner energy recordings with system logs to characterize energy use and utilization across seven clinical MRI scanners. It analyzes activity, operating modes, examinations, and sequence-level energy relationships, with filtering and visualization choices shaped by the dataset and intended analyses.
2 Methods
- Seven scanners at one university hospital were studied retrospectively, with scanner energy and MRI system logs collected and integrated for analysis.Power meters sampled the complete scanner system, including its reconstruction workstation, at 1 Hz; logs supplied operating-state, activity, event, and scan-parameter information.Source: §2 · p. 3, §2.1 · p. 3
2.1 Data Collection and Preprocessing
- Energy recordings and logs were time-aligned to calculate energy per examination and measurement and daily energy in scan, idle, and eco-power modes.Eco-power is the vendor-defined energy-saving mode in which the helium pump switches off and reactivates at predefined temperature or pressure thresholds; idle and eco-power were separated using power thresholds. Examinations include multiple sequences, such as localizers and adjustment scans.Source: §2.1 · pp. 3–4
- The dataset combined scanner, examination, measurement, and acquisition records; observations exceeding three-hour examination or 30-minute measurement limits were excluded.The initial dataset contained 600,293 measurements, and the resulting collection comprised 26,471 files spanning January 1, 2024, to November 30, 2025.Source: §2.1 · p. 4
2.2 Data Analysis
- Analyses summarized daily activity, scanner occupancy, weekday and operating-mode energy, and sequence-level energy relationships with scan duration and average power.Energy was averaged per scanner before fleet aggregation to account for differences in data availability; sequence relationships were evaluated with scatter plots and linear regression.Source: §2.2 · p. 4
- The interactive dashboard supports filtering and inspection of selected data, while energy plots omit error bars because variability measures could misrepresent recording uncertainty.The dashboard also includes a live data connection for continuous monitoring; energy and power values came from log data rather than repeated recordings subject to measurement noise.Source: §2.2 · p. 4
3 Results
Energy use varied by scanner, examination, measurement, operating mode, and clinical application. Scan duration was more strongly associated with sequence energy than average power, while fleet-level idle energy exceeded scan-mode energy and scanner occupancy differed across machines.
3 Results
- The fleet averaged 42.8 examinations and 1329.3 measurements per day, with notable differences between scanners.Measurement counts include adjustment and localizer scans, and data-transfer interruptions produced periods of missing data.Source: §3 · p. 5
- 0.43 kWh per measurement and 13.50 kWh per examination were the respective fleet averages, with scanner rankings differing between examination and measurement levels.Examination energy ranged from 9.70 kWh for Aera (UFK) to 17.51 kWh for VidaFit (CRONA); measurement averages ranged from about 0.36 kWh for five scanners to 1.10 kWh for Aera (UFK).Source: §3 · p. 5
- Mean scanner occupancy was 19.38%, with scanner averages ranging from 12.74% for Aera (UFK) to 20.04–23.15% for four scanners.Occupancy decreased on weekends; Vida (CRONA) operated on Saturdays but not Sundays, whereas Aera (UFK) was not used on weekends.Source: §3 · p. 5
- Across the fleet, average daily idle energy was 1181.60 kWh, exceeding scan-mode energy at 1041.27 kWh; eco-power contributed 132.00 kWh.Average daily fleet consumption was 2354.89 kWh, while weekend totals were lower than weekday totals.Source: §3 · pp. 5–6
- Active scan-mode energy averaged 262.40 kWh for VidaFit (CRONA), compared with 97.07–165.24 kWh for other scanners.Idle-mode consumption varied more narrowly, from 135.73 to 188.99 kWh, and only four scanners used eco-power mode, with consumption ranging from 5.26 to 58.03 kWh among them.Source: §3 · p. 6
- Breast sequences averaged 1.20 kWh, the highest body-region value, while cardiovascular sequences averaged 0.18 kWh, the lowest.Aera (UFK) breast examinations averaged 1.20 kWh compared with 0.66 kWh per breast measurement on AvantoFit (CRONA); the difference may partly reflect dedicated versus screening protocols and their acquisition times.Source: §3 · p. 6
- R2 was 0.769 for sequence energy versus scan duration, compared with 0.426 for energy versus average power.Sequences with similar durations could still vary in energy, consistent with scanner power demand and acquisition parameters also contributing to energy variability.Source: §3 · pp. 6–7
4 Discussion
The discussion links MRI energy demand to scan duration, scanner utilization, examination composition, and operating modes. It also identifies limits of the retrospective single-hospital analysis and calls for controlled recordings and more advanced modeling.
4 Discussion
- The integrated analysis and interactive dashboard support exploration of energy patterns, while providing a foundation for future energy-optimization approaches.The dashboard incorporates live hospital data for continuous visualization and real-time monitoring of scanner power consumption over the preceding three hours.Source: §4 · pp. 7–8
- Energy varied across scanners, examinations, and measurements, while temporal patterns were associated with scheduled examination types and scanner utilization.Daily demand was lower on weekends because of fewer routine scans; elevated Saturday energy use on the Vida scanner was likely related to overnight research scans.Source: §4 · p. 7
- Idle and eco-power modes consumed more cumulative energy than active scanning, but comparisons between scanners should account for occupancy.Idle and eco-power consumption remained relatively constant through the week, consistent with continuous cooling requirements.Source: §4 · p. 7
- Grouping examinations by body region was too broad to explain energy variation because examinations can differ in sequences and acquisition parameters.The authors suggest sequence-based grouping that combines sequence information with acquisition parameters, since sequence names alone are inconsistent and incomplete.Source: §4 · pp. 7–8
- R2 = 0.769 for scan duration’s relationship with energy consumption, compared with R2 = 0.426 for average power.The authors conclude that scan duration is more strongly associated with energy consumption, while power differences may contribute additional variability.Source: §4 · p. 8
- Very low energy occurred across a wide range of average power values, and energy also varied substantially at similar power levels.The near-zero band may reflect very short scan durations; vertical clusters at similar durations suggest power demand also contributes to variability.Source: §4 · p. 8
- The retrospective clinical setting introduces variability, and the seven scanners came from one manufacturer at one hospital, limiting generalizability.Network interruptions also caused incomplete temporal measurements and data loss.Source: §4 · pp. 7–8
- The analysis primarily used linear methods, which may miss nonlinear or more complex dependencies among acquisition parameters and energy demand.Future work should use controlled recordings and advanced models that account for nonlinear relationships and interactions.Source: §4 · pp. 8–9
5 Conclusion
The study integrates MRI sequence and acquisition data with energy recordings to characterize demand beyond scanner and examination levels. Its findings identify scan duration as an important factor associated with consumption, while power variation contributes less; further work is needed to identify underlying acquisition parameters.
5 Conclusion
- Merging MRI sequence and acquisition parameters with energy recordings enables characterization of energy consumption beyond examination and scanner levels.Source: §5 · p. 9