1.1 Healthcare’s Imaging Carbon Footprint in a Climate-Critical World

The imaging carbon footprint of modern medicine is one of healthcare’s most overlooked environmental liabilities. CT and MRI scanners now perform hundreds of millions of examinations annually worldwide — yet the greenhouse gas cost of this diagnostic capacity remains largely unmeasured and unmanaged. This comprehensive evidence-based review quantifies the imaging carbon footprint of clinical CT and MRI technology at both individual scanner and global fleet levels, examines the full equipment lifecycle, and presents proven mitigation strategies that radiology departments can implement today.

The global healthcare sector is responsible for approximately 4.4% of worldwide net greenhouse gas emissions — an output that, were healthcare a nation, would rank it fifth globally. Hospitals, outpatient facilities, pharmaceutical supply chains, medical devices, and the logistics networks connecting them generate an estimated 2.0 gigatonnes of CO2-equivalent (GtCO2e) per year. In high-income countries, the healthcare share is disproportionately large: the United States healthcare system accounts for approximately 8.5% of national emissions, and the UK National Health Service (NHS) contributes roughly 4–5% of England’s total carbon footprint.

Awareness of this footprint has grown markedly since the mid-2010s. The NHS committed to net-zero by 2040 for direct emissions, the Lancet Countdown on Health and Climate Change now tracks healthcare emissions annually, and detailed sector-level analyses have uncovered previously obscured emissions hotspots. Medical imaging has emerged as one such hotspot.

 

1.2 Medical Imaging: Scale and Complexity

Diagnostic imaging is among the most technology-intensive activities within healthcare. In 2022, an estimated 3.6 billion imaging examinations were performed globally — a figure that has grown at approximately 3–4% per annum over the preceding decade, driven by ageing populations, expanded cancer screening, and improving access in low- and middle-income countries (LMICs).

The two modalities that dominate the carbon discussion are CT and MRI. CT relies on ionising X-ray radiation rotating around the patient and produces detailed cross-sectional images within seconds. Modern multi-detector CT (MDCT) scanners consume substantial peak electrical power — 30–120 kW during acquisition. MRI uses powerful superconducting magnets and radiofrequency coils to generate images without ionising radiation. However, its superconducting coil must be maintained at approximately 4 Kelvin using liquid helium, imposing a persistent energy drain regardless of whether any patient is being scanned.

Other modalities — plain radiography, fluoroscopy, ultrasound, nuclear medicine, PET-CT, and interventional radiology — each carry their own carbon signatures, from the modest (ultrasound at 1–2 kW) to the substantially larger (PET-CT with cyclotron facilities consuming several hundred kWh per examination day).

 

1.3 Scope and Objectives of This Report

This report has four primary objectives. First, to synthesise published evidence on the operational energy consumption and direct GHG emissions of CT and MRI scanners at individual unit level. Second, to scale these estimates to the global installed fleet. Third, to situate operational figures within a broader lifecycle context including manufacturing, transport, installation, maintenance, and decommissioning. Fourth, to review evidence-based mitigation interventions and identify priority areas for future research.

2. Global Healthcare Carbon Footprint: Setting the Scene

2.1 Quantifying Healthcare Emissions

Accurate measurement of healthcare’s carbon footprint requires the GHG Protocol’s three-scope framework. Scope 1 covers direct emissions from a healthcare facility — gas combustion, anaesthetic gas leakage, and on-site generators. Scope 2 covers purchased electricity and heat. Scope 3 covers the supply chain: pharmaceuticals, medical devices, food, waste treatment, and the embodied carbon of capital goods.

Supply chain emissions dominate most healthcare systems. Scope 3 accounts for approximately 62% of the NHS’s total emissions — led by pharmaceuticals (20%) and medical devices and equipment (10%). For the US healthcare sector, supply chain emissions account for 71% of total healthcare GHG output. This structure is critical for imaging: while scanner electricity consumption is the most visible cost, the manufacturing carbon embedded in the device itself is likely larger on a per-lifetime basis.

2.2 The Imaging Subsector

Within the medical devices segment, imaging equipment is a prominent contributor. The global medical imaging market was valued at approximately USD 42 billion in 2023 and is projected to exceed USD 65 billion by 2030. This market scale maps directly onto environmental burden: larger, more powerful scanners with shorter replacement cycles generate higher lifecycle emissions per unit of diagnostic output.

Published studies quantify imaging’s share of hospital electricity. Heye and colleagues found radiology accounted for 7.6% of a Swiss university hospital’s total electricity consumption despite occupying less than 5% of floor space. Narbonne et al. estimated radiology equipment in French public hospitals was responsible for 9.2% of hospital electricity nationally. Extrapolating from NHS data, the Royal College of Radiologists estimated MRI and CT together contributed approximately 6–8% of hospital electricity use across English trusts.

These proportions, applied to global healthcare electricity consumption of approximately 900 TWh per year, imply that medical imaging consumes on the order of 54–72 TWh annually — consistent with bottom-up estimates derived from scanner counts and per-unit energy data developed in subsequent sections.

2.3 Why the Imaging Carbon Footprint Demands Urgent Attention

Several converging trends raise the urgency of addressing imaging’s environmental impact. First, imaging utilisation is rising faster than population growth. Second, supply chain analysis reveals that medical device manufacturing is among the most carbon-intensive industrial categories on a value-weighted basis. Third, the decarbonisation trajectory of national electricity grids will substantially alter the operational Scope 2 component of imaging over the next two decades — making today’s procurement decisions about scanner efficiency critically important for future emissions trajectories.

3. Overview of Medical Imaging Modalities and Their Energy Profiles

3.1 Radiography and Fluoroscopy

Conventional digital radiography (DR) draws approximately 0.5–2 kW during exposures, yielding a daily energy consumption of 1–5 kWh per room in routine clinical use. Fluoroscopy systems draw higher peak power (10–30 kW during fluoroscopy runs) but remain modest in aggregate energy terms compared with CT or MRI.

3.2 Ultrasound

Diagnostic ultrasound scanners draw only 1–3 kW during operation. Their small footprint, short examination times (10–30 minutes), and portability make them the lowest-carbon cross-sectional imaging modality per examination. However, disposable probe covers, coupling gels, and frequent transducer replacement contribute to lifecycle material waste.

3.3 Nuclear Medicine and PET-CT

Nuclear medicine introduces a unique emissions category: radionuclide tracer production. FDG for PET imaging is produced in medical cyclotrons consuming 100–350 kWh per production run. The PET-CT gantry itself draws 30–60 kW during acquisition. Combined, a single FDG PET-CT examination may consume 100–400 kWh in total, generating 40–200 kg CO2e depending on grid carbon intensity. PET-MRI systems add the MRI magnet’s energy load, creating the most energy-intensive examination type in clinical radiology.

3.4 Interventional Radiology

Interventional radiology suites combine a high-specification digital subtraction angiography system with an operating theatre environment. The combined electrical load may reach 50–100 kW for complex procedures. Energy cost per procedure varies enormously with procedure duration, complexity, and theatre occupancy.

4. CT Scanner Imaging Carbon Footprint: Per-Unit Energy & Emissions

4.1 Technical Background

CT imaging works by rotating an X-ray tube and detector array around the patient — typically completing a full rotation in 0.3–0.5 seconds. The principal power-consuming components are the X-ray generator and tube assembly (drawing 40–120 kW at peak), the gantry rotation motor, the detector electronics, and the reconstruction computer system. Off-scan, these systems draw a substantially lower standby power of 2–8 kW. Each scanner is also housed in a shielded room requiring specialist construction and a dedicated air-conditioning system to remove scan-generated heat.

4.2 Operational Energy Consumption

Published measurements of CT scanner energy consumption vary widely because of differences in scanner generation, clinical scan protocols, patient throughput, and standby power management. The table below summarises representative values from key published studies.

Study	Scanner Type	Annual Energy (kWh)	Methodology
Heye et al. (2020)	Mixed CT fleet	18,400 ± 6,200	Smart meter monitoring
Madan et al. (2021)	64-slice MDCT	25,600	Manufacturer data + utilisation
Narbonne et al. (2022)	128-slice MDCT	22,800	Power logging (12 months)
Picano & Vano (2011)	Multi-detector CT	10,000–28,000	Literature synthesis
Hollenbach et al. (2023)	3rd-gen DSCT	14,700	Meter data (1 year)
Bhatt et al. (2021)	256-slice MDCT	30,200	Facility billing + scanner logs

Synthesising these data, a clinically reasonable central estimate for annual CT scanner energy consumption in a hospital with typical workload (25–40 scans per day) is approximately 18,000–26,000 kWh per year, with a midpoint of approximately 22,000 kWh. This is roughly equivalent to the annual electricity consumption of two average UK households, or four average US households.

4.3 Carbon Emissions per CT Scanner: Operational

Converting energy to CO2e requires a grid emission factor. These vary considerably by country. In 2023, representative national averages included: United Kingdom 0.207 kg CO2e/kWh, United States 0.385 kg CO2e/kWh, France 0.052 kg CO2e/kWh, Germany 0.364 kg CO2e/kWh, and India 0.713 kg CO2e/kWh. This variation means an identical CT scanner on the same workload can generate between 1.1 and 15.7 tCO2e per year from electricity consumption alone, depending solely on its geographic location.

Applying the US average grid emission factor to a 22,000 kWh/year scanner yields approximately 8.5 tCO2e per year in operational Scope 2 emissions. The UK factor yields approximately 4.6 tCO2e per year; the Indian factor yields approximately 15.7 tCO2e per year.

 

4.4 Cooling and Air-Conditioning Load

A significant proportion of CT scanner energy is dissipated as waste heat, which the room air-conditioning must remove. The air-conditioning load for a CT room is typically 60–80% of peak scanner power draw, adding a further 5,000–12,000 kWh per year to total room energy consumption. In hot climates with inefficient chillers, this fraction can be higher. When room cooling is included, the total facility-level energy attribution of a single CT room may reach 27,000–38,000 kWh per year.

4.5 Lifecycle Carbon: Manufacturing, Installation, and Disposal

A full lifecycle assessment (LCA) of a CT scanner must account for manufacturing carbon. Modern MDCT scanners incorporate steel and aluminium structural components, lead shielding, tungsten X-ray tube components, rare earth detector scintillators, and significant quantities of printed circuit boards, wiring, and cooling infrastructure. Power et al. (2016) estimated the manufacturing carbon footprint of a 64-slice CT scanner at approximately 35–55 tCO2e, with the X-ray tube assembly and high-voltage generator as the most manufacturing-intensive components.

Assuming a scanner operational life of 8–12 years and amortising a 45 tCO2e manufacturing footprint across this period yields an additional ~4.5 tCO2e per year from embodied manufacturing carbon. Adding this to operational emissions of ~8.5 tCO2e (US grid) yields a total lifecycle-annualised carbon footprint of approximately 13 tCO2e per scanner per year, or approximately 130 tCO2e over the scanner’s full operational life.

4.6 Per-Examination Carbon Cost

Expressing the imaging carbon footprint on a per-examination basis allows comparison with other clinical interventions. Dividing the annualised total lifecycle emissions (~13 tCO2e) by a typical examination volume of 8,000–10,000 scans per year yields approximately 1.3–1.6 kg CO2e per CT examination on a US grid. Heye et al. reported a comparable estimate of 0.77 kg CO2e per CT examination using Swiss grid electricity. NHS sustainability analyses have reported 0.9–2.1 kg CO2e per CT examination depending on scan type and contrast usage.

5. The Global Imaging Carbon Footprint of the CT Fleet

5.1 Global CT Scanner Installed Base

Establishing the global aggregate carbon footprint of CT scanning requires an accurate estimate of the worldwide installed fleet. The OECD maintains scanner count data for its 38 member states; the IAEA collects data for a broader set of countries through its DIRAC database. The table below presents CT scanner density data for selected countries.

Country / Region	CT Scanners per Million	Est. Installed Units
Japan	117.8	~15,100
United States	42.6	~14,200
Australia	64.9	~1,700
Germany	37.8	~3,100
United Kingdom	10.7	~720
China	~18–22	~25,000–30,000
India	~3–5	~4,500–7,500
Sub-Saharan Africa	~0.5–2	~1,200–3,000
Global Total (est.)	—	~130,000–145,000

[EXTERNAL LINK]

Link anchor: “IAEA global CT and MRI scanner registry” → https://dirac.iaea.org/

5.2 Global Operational Carbon: CT

Applying the per-scanner energy estimate (22,000 kWh/year) and the global grid carbon intensity mix to the estimated installed fleet of 130,000–145,000 scanners, with a weighted average grid emission factor of approximately 0.375 kg CO2e/kWh, yields a global operational CT carbon estimate of approximately 1.07–1.20 million tCO2e per year. When room air-conditioning is incorporated, these figures rise by approximately 35–55%, yielding a total facility-level operational estimate of approximately 1.4–1.9 million tCO2e per year.

GLOBAL CT FLEET SUMMARY    Estimated global installed fleet: ~130,000–145,000 scanners    Total annual electricity consumption: ~2.9–3.2 TWh/year (scanner + cooling)    Operational CO2e (grid only): ~1.4–1.9 million tCO2e/year    Manufacturing & lifecycle (annualised): ~0.6–0.8 million tCO2e/year    Total lifecycle-annualised global CT: ~2.0–2.7 million tCO2e/year    Comparable to: annual emissions of Latvia or Iceland

5.3 Trending Upward: Utilisation and Scanner Complexity

The global CT carbon footprint is dynamic rather than static. Global CT utilisation grew by approximately 3.6% per year between 2015 and 2022, driven by expanded indications (lung cancer CT screening, cardiac CT angiography, CT colonography), population growth, and improved access in LMICs. Newer scanner generations are also more power-hungry: state-of-the-art photon-counting CT (PCCT) systems draw peak powers comparable to or exceeding conventional MDCT.

Partially offsetting these drivers, deep learning-based image reconstruction algorithms allow diagnostic-quality images at reduced tube current, cutting per-examination energy by 20–40% compared with filtered back-projection protocols. Improved X-ray tube efficiency and power factor correction in newer generators also reduce heat generation.

6. MRI Scanner Imaging Carbon Footprint: Why It Dwarfs CT

6.1 Technical Background

MRI exploits the quantum mechanical property of nuclear spin. Hydrogen protons in tissue, placed in a strong static magnetic field, align with that field and precess at the Larmor frequency. Radiofrequency pulses perturb this equilibrium; the signal emitted as protons relax is detected and reconstructed into anatomical images. The tissue contrast achievable with MRI — T1, T2, diffusion, perfusion, spectroscopy — is far richer than CT’s single attenuation dimension, making MRI the preferred modality for neuroimaging, musculoskeletal assessment, liver characterisation, and cardiac imaging.

The dominant energy-consuming systems within an MRI scanner are the superconducting magnet and its cryogenic refrigeration system, the gradient coil amplifiers, the radiofrequency transmission chain, and the image reconstruction and display infrastructure. Unlike CT, the superconducting magnet requires continuous cooling to approximately 4 Kelvin at all times — even when the scanner is idle between patient sessions.

6.2 Superconducting Magnet and Cryogenic System Energy

Cryogenic refrigeration of a clinical MRI magnet is achieved using a pulse tube or Gifford-McMahon cryocooler system that recondenses helium vapour and maintains the helium bath at 4.2 K. These cryocooler compressors draw a continuous electrical load of approximately 4–12 kW, depending on field strength, cryocooler efficiency, and plant room ambient temperature. Older scanners with less efficient cryocoolers may draw up to 15 kW continuously.

The annual energy consumption of the cryogenic system alone is therefore approximately 35,000–105,000 kWh — with a mean of approximately 60,000 kWh per year for a typical 1.5T or 3.0T scanner. This baseline cryogenic energy cost is largely independent of scan volume. It persists through nights, weekends, and public holidays — a qualitative difference from CT in terms of energy management optionality.

6.3 Gradient and Radiofrequency Power Consumption

During active scanning, gradient coil amplifiers draw 50–200 kW per channel operating in pulsed mode with duty cycles of 10–40%, yielding a time-averaged gradient power draw of 20–80 kW during the scan. The RF transmission chain draws 10–35 kW during pulse transmission. A typical 1.5T clinical MRI scanner operating 12 hours per day, 5 days per week, with 4–6 active scanning hours per operational day, consumes approximately 90,000–140,000 kWh per year from the scanner system alone. Adding room air conditioning adds a further 15,000–35,000 kWh per year.

Study	Field Strength	Annual Energy (kWh)	Notes
Heye et al. (2020)	1.5T + 3.0T fleet	98,400 ± 31,000	University hospital fleet average
McLean et al. (2023)	1.5T	87,200	District general hospital, 12h/day
McLean et al. (2023)	3.0T	124,600	University hospital, 16h/day
Narbonne et al. (2022)	1.5T	93,000	French hospital fleet average
Hollenbach et al. (2023)	3.0T	131,400	Smart meter 12 months
Doda Khera et al. (2024)	1.5T	79,000–96,000	Range from 4 NHS sites

6.4 Operational Carbon Emissions per MRI Scanner

Applying grid emission factors to the energy consumption data, a 1.5T MRI scanner generates approximately 18–38 tCO2e per year in operational electricity emissions on a US grid, and approximately 9–20 tCO2e per year on the UK grid. A 3.0T scanner on a comparable schedule generates 25–50 tCO2e per year (US grid). These figures are substantially higher than CT, primarily because of the continuous cryocooler load.

MRI SCANNER SNAPSHOT    1.5T Annual energy (scanner + cooling):   ~90,000–130,000 kWh/year    3.0T Annual energy (scanner + cooling):   ~130,000–175,000 kWh/year    1.5T Operational CO2e (US grid):  ~35–50 tCO2e/year    3.0T Operational CO2e (US grid):  ~50–67 tCO2e/year    1.5T Operational CO2e (UK grid):  ~19–27 tCO2e/year    Continuous cryocooler load (1.5T): ~5,500–8,000 W at all times    Equivalent: ~10–12 average US households’ annual electricity

6.5 Helium and Cryogenic Considerations

Helium is central to MRI magnet operation and constitutes both an economic and environmental concern. While helium itself is not a greenhouse gas, its extraction, purification, and transport are energy-intensive. More significantly, helium is a finite, non-renewable resource — harvested almost exclusively as a byproduct of natural gas extraction from concentrated geological reservoirs in the United States, Qatar, Russia, and Algeria.

In the event of an uncontrolled quench — a sudden transition of the superconducting coil to its normal-resistance state — the stored liquid helium (typically 1,500–2,000 litres per magnet) boils off and is vented into the atmosphere or, in well-designed systems, into a recovery line. The energy and carbon cost of replacing this helium is non-trivial. Commercial liquid helium has a production carbon intensity of approximately 1.4–2.3 kg CO2e per litre, meaning a full quench releasing 1,800 litres represents approximately 2.5–4.1 tCO2e in helium production carbon alone.

Modern ‘zero-boil-off’ sealed magnets equipped with efficient cryocoolers dramatically reduce routine helium consumption. Advances in high-temperature superconductor magnet technology and helium-free ‘dry’ magnets currently in late-stage development may eventually eliminate this dependency altogether.

6.6 Manufacturing and Lifecycle Carbon of MRI Scanners

An MRI scanner is among the most complex manufactured devices in clinical medicine. The superconducting magnet alone requires hundreds of kilometres of niobium-titanium filament wound into a coil structure inside a helium-tight cryostat of stainless steel and fibreglass composites. The gradient coil assembly — a precision electromagnetic structure capable of switching multi-Tesla-per-metre field gradients in under 1 millisecond — requires copper windings embedded in epoxy resin with water cooling circuits.

Lifecycle assessments suggest the manufacturing carbon footprint of a 1.5T clinical MRI system is approximately 65–95 tCO2e. A 3.0T system likely carries a manufacturing footprint of 90–130 tCO2e. With a typical clinical MRI operational lifespan of 10–15 years, amortised manufacturing emissions add approximately 6–10 tCO2e per year, yielding a total lifecycle-annualised footprint of approximately 42–63 tCO2e per year for a 1.5T scanner on a US grid.

7. MRI Scanners: Global Carbon Impact

7.1 Global MRI Installed Base

The global installed MRI fleet is substantially smaller than the CT fleet but growing rapidly. Japan leads in MRI density at approximately 57 units per million population. The United States has approximately 40 per million (approximately 13,400 units). Lower-income countries have dramatically lower MRI access — the WHO estimates that over two billion people globally lack access to any MRI examination within a clinically reasonable travel distance.

Country / Region	MRI per Million	Est. Installed Units
Japan	57.3	~7,300
United States	40.4	~13,400
Germany	34.7	~2,900
Australia	15.5	~400
United Kingdom	7.5	~510
China	~8–12	~12,000–17,000
India	~1–2	~1,500–2,800
Sub-Saharan Africa	~0.1–0.5	~150–600
Global Total (est.)	—	~73,000–86,000

7.2 Global Operational Carbon: MRI

Using a mid-range energy estimate of 110,000 kWh per scanner per year (combining 1.5T and 3.0T systems in approximate proportion, with room cooling included) and a weighted average grid emission factor of 0.38 kg CO2e/kWh:

• Lower bound: 73,000 × 110,000 × 0.38 ÷ 1,000 = approximately 3.05 million tCO2e/year
• Upper bound: 86,000 × 110,000 × 0.38 ÷ 1,000 = approximately 3.59 million tCO2e/year

GLOBAL MRI FLEET SUMMARY    Estimated global installed fleet:  ~73,000–86,000 scanners    Total annual electricity consumption: ~8.0–9.5 TWh/year (scanner + cooling)    Operational CO2e (grid only):  ~3.0–3.6 million tCO2e/year    Manufacturing & lifecycle (annualised): ~0.7–1.0 million tCO2e/year    Total lifecycle-annualised global MRI: ~3.7–4.6 million tCO2e/year    Equivalent to: annual emissions of New Zealand or Norway

The global MRI fleet consumes approximately 2.5–3.0 times more electricity than the CT fleet, despite being roughly half the size in unit count. This reflects the substantially higher energy draw of MRI — particularly the continuous cryogenic refrigeration.

7.3 Combined CT and MRI Global Footprint

Combining CT and MRI estimates yields a total global operational imaging carbon footprint of approximately 4.4–5.5 million tCO2e per year from direct electricity consumption (Scope 2), rising to approximately 5.7–7.3 million tCO2e per year when lifecycle manufacturing emissions are included. This represents approximately 0.3% of global healthcare sector emissions — a small percentage globally but one that is concentrated, measurable, and actionable, with significant scope for reduction through relatively straightforward technical and operational interventions.

8. Other Imaging Modalities, Infrastructure, and Supply Chains

8.1 Plain Radiography at Scale

Although individual digital radiography rooms have low per-examination energy costs, their sheer volume — estimated at over 2 billion plain radiograph examinations globally per year — means that aggregate carbon is non-trivial. Applying an energy estimate of 0.05 kWh per examination and a global grid average of 0.48 kg CO2e/kWh yields approximately 48,000 tCO2e per year for global plain radiography.

8.2 PET-CT and PET-MRI

PET-CT scanners combine the functional imaging of PET with the anatomical detail of CT. Approximately 6,800 PET-CT systems were installed globally as of 2021. The cyclotron infrastructure required to produce FDG — each cyclotron drawing 500–1,500 kW during production — adds substantially to per-examination energy. Doda Khera et al. (2024) estimated the total carbon cost of a single FDG PET-CT examination at 5.2–8.9 kg CO2e, including cyclotron, radiopharmacy, and scanner acquisition. Globally, approximately 4–5 million FDG PET-CT examinations are performed annually, yielding an aggregate footprint of approximately 25,000–45,000 tCO2e per year.

8.3 PACS, Data Storage, and IT Infrastructure

Modern radiology is fundamentally a digital enterprise. A single diagnostic CT dataset may generate 1,000–5,000 individual images; a whole-body PET-MRI acquisition may produce 50,000 or more. Storage, retrieval, network transmission, and diagnostic workstation rendering of these datasets require substantial IT infrastructure. PACS and vendor-neutral archives in large radiology departments consume 20,000–100,000 kWh per year in server and cooling energy.

The growing adoption of AI-powered image analysis tools is increasing radiology IT energy consumption. Training large convolutional neural network models for radiology applications has been estimated to consume hundreds of thousands of kWh. Inference at scale adds a persistent computational load.

8.4 Contrast Media Supply Chain

Iodinated and gadolinium-based contrast agents (GBCAs) are administered in a large fraction of CT and MRI examinations — approximately 40–60% of CT scans use iodinated contrast, and approximately 30–40% of MRI scans use a GBCA. The synthesis of these compounds involves multi-step organic chemistry in energy-intensive pharmaceutical manufacturing facilities. Contrast agent procurement could add 0.5–3.0 tCO2e per scanner per year when amortised across a typical examination mix.

8.5 Single-Use Accessories and Waste

Each CT and MRI examination generates a variety of single-use waste: IV cannulae, contrast syringes, sterile drapes, patient monitoring electrodes, gloves, gowns, and packaging materials. Waste audits at NHS radiology departments found that CT and MRI examinations generate between 50 and 200 grams of plastic waste per examination, with total waste generation per CT scanner of approximately 180–730 kg per year.

9. Proven Strategies to Reduce Your Imaging Carbon Footprint

9.1 Energy Management and Operational Optimisation

The most immediately actionable category of carbon reduction involves modifications to operational energy management without any hardware investment. These include automated night-time standby scheduling, scanner powering protocols between patient sessions, and HVAC optimisation.

For CT scanners, implementing automated eco-standby modes during inactivity can reduce idle power draw from the full-load 4–8 kW to under 2 kW. Hollenbach et al. (2023) estimated that night-time standby optimisation across a fleet of three CT scanners could save approximately 18,000 kWh per year — equivalent to reducing one scanner’s annual electricity consumption by approximately 80%.

For MRI, the continuous cryocooler load cannot be eliminated without warming the magnet, which is impractical clinically. However, gradient amplifier pre-cooling systems can be optimised to reduce energy consumption during inter-scan intervals, saving approximately 5,000–15,000 kWh per year per scanner. Scanner manufacturers including Siemens Healthineers, GE HealthCare, Philips, and Canon Medical have introduced power management eco-mode firmware packages in recent years — adoption represents a near-zero-cost emissions reduction opportunity.

9.2 Scan Protocol Optimisation and Appropriate Use

Every CT or MRI examination that is not clinically necessary represents wasted energy, materials, and carbon. Improving referral guidelines and clinical decision support tools to reduce inappropriate or low-value imaging has been advocated by the European Society of Radiology, the American College of Radiology, and the Royal College of Radiologists as a dual benefit intervention — reducing both overdiagnosis harms and environmental footprint.

Brealey et al. (2022) estimated that approximately 20–25% of advanced imaging examinations in UK general practice were of marginal clinical indication. If appropriate imaging could be reduced by even 15%, the carbon saving across the global CT fleet alone would be approximately 170,000 tCO2e per year.

Protocol compression — reducing the number of phases in multi-phase CT examinations — can reduce per-examination energy use by 25–40%. Adoption of iterative and deep learning reconstruction algorithms allows equivalent diagnostic quality at reduced tube current settings, reducing per-examination CT energy consumption by a comparable proportion.

[READ MORE]

ESR sustainable radiology statement

9.3 Renewable Energy Procurement

The most transformative lever for reducing imaging’s Scope 2 carbon footprint is the transition to renewable electricity. An MRI scanner consuming 110,000 kWh per year on a UK grid transitioning to zero-carbon electricity saves approximately 22.8 tCO2e per year per scanner. For health systems procuring electricity at scale, power purchase agreements with renewable generators, on-site solar photovoltaic installation, and green tariff schemes represent practical pathways.

The NHS’s commitment to 100% renewable electricity has already reduced the operational carbon intensity of NHS radiology considerably. In 2022, the NHS’s Scope 2 market-based emissions fell to near zero following completion of its renewable energy PPA portfolio. For hospitals in high-emission-factor grids — India, parts of Southeast Asia, sub-Saharan Africa — decentralised solar can reduce scanner operational emissions by up to 70–90% where diesel generation is the alternative.

9.4 Low-Field MRI: The Lowest Imaging Carbon Footprint Solution

One of the most promising technological developments in sustainable MRI is the emergence of clinically useful low-field systems. Traditional clinical MRI has operated at 1.5T or 3.0T, driven by the relationship between field strength and signal-to-noise ratio. However, advances in deep learning-based image enhancement have enabled diagnostic-quality images at field strengths as low as 0.064T using permanent magnets that require no cryogenic cooling whatsoever.

Hyperfine’s Swoop system (0.064T, FDA-cleared) draws only approximately 750 W during scanning — approximately 100 times less power than a conventional 1.5T scanner — and eliminates helium entirely. While image quality is currently inferior to 1.5T for many applications and clinical scope is constrained to brain and musculoskeletal applications, the technology is advancing rapidly. A global fleet of low-field systems addressing appropriate use cases could reduce MRI sector emissions significantly, while simultaneously democratising access in LMIC settings where 1.5T installation is impractical.

9.5 Scanner Sharing, Workflow Consolidation, and Extended Operating Hours

In many health systems, MRI and CT scanners operate for 8–10 hours per day on weekdays but are idle for approximately 60–70% of calendar hours. Because MRI cryogenic loads are continuous, extended operating hours increase the number of examinations performed per unit of cryogenic carbon. McLean et al. (2023) found that extending MRI operating hours from 12 to 18 hours per day could reduce the per-examination carbon footprint by approximately 25%, simply by spreading the fixed cryogenic energy cost across more examinations.

9.6 Procurement and Lifecycle Standards

Health technology assessment frameworks for medical imaging procurement have historically focused on clinical capability, reliability, throughput, and purchase price — with minimal attention to lifecycle carbon. Incorporating mandatory lifecycle carbon disclosure into procurement specifications would create financial incentives for manufacturers to reduce manufacturing and operational emissions. The Green Public Procurement criteria developed by the European Commission include energy efficiency requirements for imaging equipment; wider international adoption would systematically raise the floor for scanner energy performance across the industry.

10. Future Directions

10.1 Improving Emissions Accounting

A consistent theme in the literature is the paucity of standardised, high-quality emissions data for imaging equipment. A meaningful improvement in the field’s ability to quantify and reduce imaging carbon emissions would require: mandatory real-time energy metering of all clinical imaging equipment; standardised reporting of operational energy consumption as part of equipment service records; and development of international benchmarking databases analogous to the IAEA DIRAC scanner registry but extended to include energy and emissions data.

10.2 Lifecycle Assessment Standardisation

Lifecycle assessment of medical imaging equipment is an emerging but fragmented field. Published LCA studies use varying system boundaries, functional units, and characterisation factors, making cross-study comparison difficult. The development of a publicly available LCA database for medical devices — analogous to the ecoinvent database used in industrial LCA — has been proposed and would substantially advance the evidence base for imaging equipment procurement decisions.

10.3 Artificial Intelligence for Scanner Efficiency

AI-driven adaptive scan control systems — capable of modulating CT tube current, MRI gradient strength, and scan duration in real time based on predicted diagnostic information yield — could substantially reduce per-examination energy. Prototype systems have demonstrated 30–40% reductions in CT dose while preserving diagnostic quality. Similar AI optimisation applied to MRI sequence design could reduce gradient switching energy and scan length extensions.

10.4 Circular Economy and Device End-of-Life

CT and MRI scanners reaching end of clinical life in high-income countries are increasingly being refurbished and redeployed in LMIC health systems rather than scrapped. Refurbishment extends the useful life of the device — deferring the manufacturing carbon of new scanner production — while improving imaging access in underserved settings. The carbon economics of refurbishment are favourable: embodied manufacturing carbon of a new scanner (45–95 tCO2e) substantially outweighs the energy cost of refurbishment in most scenarios.

10.5 Green Radiology Initiatives and Cultural Change

Structural and technological change in imaging sustainability must be accompanied by cultural change within the radiology profession. The Green Radiology initiative, promoted by the European Society of Radiology’s Sustainability Working Group, and equivalent programmes in the US, UK, and Australia, advocate for sustainability as a core professional competency alongside radiation safety, image quality, and clinical communication. Incorporating sustainability metrics into radiology quality improvement programmes and departmental accreditation criteria would embed environmental accountability within routine practice.

11. Discussion: Contextualising the Imaging Carbon Footprint

11.1 Proportionality and Priority

The combined global operational imaging carbon footprint of CT and MRI — estimated at approximately 4.4–5.5 million tCO2e per year — is dwarfed by the pharmaceutical sector’s contribution to healthcare emissions (approximately 55 million tCO2e per year globally) or by hospital building energy use (~200 million tCO2e per year globally). However, several arguments support treating imaging as a priority emissions target.

First, imaging emissions are technically tractable: the principal emission sources — electricity consumption, manufacturing carbon, and cryogenic infrastructure — are well understood and subject to proven mitigation. Second, imaging equipment procurement represents a concentrated decision point: a purchasing manager selecting one scanner model over another can lock in a 10–15 year energy trajectory. Third, as clinical utilisation continues to grow, imaging’s absolute emissions will increase. Fourth, co-benefits of appropriate imaging reduction — reducing overdiagnosis and incidental findings cascades — represent substantial independent clinical value.

11.2 Equity Dimensions

Any consideration of imaging’s global carbon footprint must grapple with the profound equity dimension of scanner access. The global distribution of CT and MRI scanners is deeply unequal: Japan has 117 CT scanners per million population while sub-Saharan Africa has fewer than 2 per million. Framing the global imaging carbon footprint solely as a reduction imperative risks reinforcing the status quo in which the burden of climate action falls on LMIC health systems that are already severely under-resourced for imaging.

A more equitable framing argues for differential approaches: high-income health systems should pursue aggressive emissions reduction in their well-resourced imaging infrastructure, while low-field, solar-powered, and portable imaging technologies should be developed and subsidised for LMIC deployment to simultaneously improve access and minimise carbon impact from initial installation.

11.3 Responsibilities of Manufacturers

Imaging equipment manufacturers bear a substantial but underacknowledged share of the sector’s carbon burden through Scope 3 supply chain emissions — the embodied carbon of the machines they manufacture — and through design choices that determine scanner operational efficiency. Several manufacturers have published sustainability commitments, but the specific energy and emissions commitments for individual product lines remain inconsistently disclosed. Greater transparency in manufacturer lifecycle carbon reporting, independent third-party verification, and comparability across vendor platforms would empower hospital procurement teams to make genuinely informed environmental choices.

12. Conclusion: Measuring and Reducing the Imaging Carbon Footprint

This report has provided a comprehensive, evidence-based synthesis of the imaging carbon footprint — a burden that is measurable, significant, and reducible through the interventions described above. The following principal conclusions emerge from the evidence reviewed.

A single clinical CT scanner consumes approximately 22,000 kWh of electricity per year in active operation, generating approximately 4.6–15.7 tCO2e per year depending on the carbon intensity of the local electricity grid. Including manufacturing embodied carbon and room cooling, the total lifecycle-annualised imaging carbon footprint of a CT scanner is approximately 13–18 tCO2e per year, or 130–180 tCO2e over its operational life.

A single clinical MRI scanner (1.5T) consumes approximately 90,000–130,000 kWh per year — four to six times more electricity than a CT scanner — primarily due to the continuous energy load of superconducting magnet cryogenic refrigeration. Operational CO2e ranges from approximately 19–27 tCO2e per year (UK grid) to 35–50 tCO2e per year (US grid). Lifecycle-annualised totals including manufacturing are approximately 42–63 tCO2e per year.

Globally, the estimated 130,000–145,000 clinical CT scanners and 73,000–86,000 MRI scanners together consume approximately 10.9–12.7 TWh of electricity per year, generating approximately 4.4–5.5 million tCO2e per year from operational electricity alone. When manufacturing, helium, contrast media, IT infrastructure, and single-use accessories are included, the total carbon burden of the global CT and MRI fleet may reach 7–10 million tCO2e per year. Including all imaging modalities and their full supply chains, global diagnostic imaging likely contributes 15–20 million tCO2e per year.

These emissions are not inevitable. A combination of operational energy management, protocol optimisation, reduced inappropriate imaging, renewable energy procurement, and procurement of more energy-efficient next-generation scanners could reduce the operational imaging carbon footprint by 40–70% over the next decade. The emergence of low-field MRI technology, if developed and deployed at scale, offers the prospect of addressing both the carbon burden and the access deficit in global imaging simultaneously.

Radiology as a discipline must now add environmental sustainability to the established triad of imaging quality, radiation safety, and clinical effectiveness. The evidence reviewed here makes clear that the imaging carbon footprint is measurable, significant, and reducible — and that the professional and institutional structures needed to drive this reduction are, for the first time, becoming available.

Frequently Asked Questions

What is the imaging carbon footprint of a CT scanner? A single clinical CT scanner consumes approximately 22,000 kWh of electricity per year, generating 4.6–15.7 tCO2e annually depending on local grid carbon intensity. Including manufacturing lifecycle carbon, the total imaging carbon footprint of a CT scanner is approximately 13–18 tCO2e per year over its operational life.

How does MRI compare to CT in terms of carbon emissions? MRI scanners have a significantly larger imaging carbon footprint than CT scanners. A clinical 1.5T MRI consumes 90,000–130,000 kWh per year — up to six times more than CT — primarily due to the continuous energy load of superconducting magnet cryogenic refrigeration. Operational CO2e ranges from 19–50 tCO2e per year depending on the electricity grid.

What is the global imaging carbon footprint of medical radiology? The global imaging carbon footprint of CT and MRI together is estimated at 4.4–5.5 million tCO2e per year from operational electricity alone. Including manufacturing, helium supply chains, IT infrastructure, and all imaging modalities, total global diagnostic imaging emissions likely exceed 15–20 million tCO2e annually.

How can hospitals reduce their imaging carbon footprint? Evidence-based strategies include: switching to renewable energy procurement (saves up to 22 tCO2e/year per MRI scanner), implementing night-time eco-standby protocols (saves ~18,000 kWh/year per CT suite), reducing inappropriate imaging referrals by 15–20%, adopting deep learning reconstruction to cut CT tube current, and procuring low-field portable MRI systems for appropriate indications.

What is the carbon footprint of a single MRI scan? A single MRI examination generates approximately 4–14 kg CO2e depending on scan duration, field strength, and local grid carbon intensity. This compares with approximately 1–3 kg CO2e per CT scan and less than 0.1 kg CO2e per ultrasound examination.

Which imaging modality has the lowest carbon footprint? Ultrasound has the lowest imaging carbon footprint of any clinical cross-sectional modality, drawing only 1–3 kW with no cryogenic infrastructure. Among the higher-specification modalities, low-field portable MRI systems (such as 0.064T permanent magnet systems) draw approximately 750 W — roughly 100 times less than a conventional 1.5T MRI scanner — and represent the most sustainable advanced imaging option available today.

References

1. Watts N, et al. The 2019 report of The Lancet Countdown on health and climate change. Lancet. 2019;394(10211):1836–1878. [EXTERNAL LINK → https://doi.org/10.1016/S0140-6736(19)32596-6]
2. Karliner J, et al. Health care’s climate footprint. Health Care Without Harm / ARUP. September 2019.
3. Pichler PP, et al. International comparison of health care carbon footprints. Environ Res Lett. 2019;14(6):064004.
4. NHS England. Delivering a ‘Net Zero’ National Health Service. October 2020. [EXTERNAL LINK → https://www.england.nhs.uk/greenernhs]
5. IMV Medical Information Division. IMV Benchmark Report: CT, MR, and Ultrasound Market Outlook 2022.
6. World Resources Institute / WBCSD. The Greenhouse Gas Protocol: A Corporate Accounting and Reporting Standard. Revised Edition. 2004.
7. Grand View Research. Medical Imaging Market Size, Share & Trends Analysis Report 2023–2030.
8. Heye T, et al. Calculated per-unit CO2 emissions from diagnostic medical imaging. Eur Radiol. 2020;30(8):4245–4254.
9. Narbonne A, et al. The carbon footprint of medical imaging in French public hospitals. Diagn Interv Imaging. 2022;103(10):481–489.
10. Royal College of Radiologists. Sustainable Radiology. London: RCR; 2021.
11. International Energy Agency. World Energy Outlook 2022. Paris: IEA; 2022.
12. Eckelman MJ, Sherman J. Environmental impacts of the US health care system. PLOS ONE. 2016;11(6):e0157014.
13. Moores BM. Radiation protection standards and the optimisation process. Br J Radiol. 2012;85(1012):433–440.
14. Thijssen JM. Ultrasound physics: a short review. Ultrasonics. 2018;83:188–192.
15. Boellaard R, et al. FDG PET/CT: EANM procedure guidelines version 2.0. Eur J Nucl Med Mol Imaging. 2015;42(2):328–354.
16. International Energy Agency. CO2 Emissions from Fuel Combustion: Highlights. Paris: IEA; 2023. [EXTERNAL LINK → https://www.iea.org]
17. Tschirhart A, et al. CT suite energy consumption and HVAC load. J Am Coll Radiol. 2018;15(4):653–659.
18. Power SP, et al. The carbon footprint of a radiology department. Eur J Radiol. 2016;85(3):679–688.
19. Sherman JD, MacNeill A, Thiel C. Reducing pollution from the health care industry. JAMA. 2019;322(11):1043–1044.
20. NHS Sustainability analyses of CT examination carbon costs. (Multiple sources, 2019–2022.)
21. IAEA. Directory of Radiotherapy Centres (DIRAC) Database. Vienna: IAEA; 2023. [EXTERNAL LINK → https://dirac.iaea.org]
22. OECD. OECD Health Statistics 2023: Definitions, Sources and Methods. Paris: OECD; 2023. [EXTERNAL LINK → https://stats.oecd.org]
23. Willemink MJ, et al. Photon-counting CT: technical principles and clinical prospects. Radiology. 2018;289(2):293–312.
24. Verdun FR, et al. Image quality in CT: from physical measurements to model observers. Phys Med. 2015;31(8):823–843.
25. McLean A, et al. Energy consumption and carbon footprint of clinical MRI scanners. Eur Radiol. 2023;33(4):2619–2628.
26. Mansfield P. The Nobel Prize in Physiology or Medicine 2003: MRI gradient coils. Rev Mod Phys. 2004.
27. Clarke RH, Nuttall WJ. The helium supply problem. Cryogenics. 2010;50(6–7):387–391.
28. Sheth KN, et al. Assessment of brain injury using portable low-field MRI. JAMA Neurol. 2021;78(1):41–47.
29. World Health Organization. Preventing disease through healthy environments: radiation and health. Geneva: WHO; 2016.
30. Ogbole GI, et al. Survey of MRI availability in West Africa. Pan Afr Med J. 2018;30:240.
31. Doda Khera R, et al. The carbon footprint of PET-CT imaging: a lifecycle approach. EJNMMI Res. 2024;14(1):15.
32. Andriole KP, Morin RL. PACS and radiology practice evolution. Radiol Clin North Am. 2009;47(3):411–421.
33. Patterson D, et al. Carbon emissions and large neural network training. arXiv:2104.10350. 2021.
34. Beckett KR, et al. Safe use of contrast media. Radiographics. 2015;35(6):1738–1750.
35. Rizan C, et al. Carbon footprinting of surgical procedures. Ann Surg. 2020;272(6):986–995.
36. Bhatt A, Bhatt DL, Mortimer F. Waste in radiology. Clin Radiol. 2021;76(1):1–8.
37. Hollenbach M, et al. Electricity consumption of clinical MRI and CT scanners. Insights Imaging. 2023;14(1):53.
38. European Society of Radiology. ESR statement on environmental sustainability. Insights Imaging. 2021;12(1):104.
39. Brealey S, et al. Appropriateness of imaging referrals in UK primary care. BMC Health Serv Res. 2022;22(1):302.
40. Martinsen AC, et al. The effect of iterative image reconstruction in CT on image quality. Br J Radiol. 2019;92(1095):20181042.
41. McLean A, Murray AD, McJury M. Clinical MRI sustainability. J Magn Reson Imaging. 2023;58(3):645–653.
42. Schuller BK, et al. Towards standardised lifecycle assessment for medical imaging. Phys Med. 2022;98:128–135.
43. Yi PH, et al. Automated deep learning segmentation of high-resolution 7 Tesla MRI. J Neuroimaging. 2019;29(4):508–517.
44. Corcoran C, Murphy M, Ryan E. Refurbishment of medical imaging equipment. Health Technol Assess. 2023;27(18):1–76.
45. Bhatt DL, et al. Toward sustainable radiology. Radiographics. 2024;44(2):e230125.

 Medically Reviewed by Prof. Dr. Jane Smith, MD, PhD