Low field, high impact: democratizing MRI for clinical and research innovation

Derek K. Jones, Daniel C. Alexander, Karen Chetcuti, Mara Cercignani, Kirsten A. Donald, Mark A. Griswold, Emre Kopanoğlu, Ike Langunju, Johnes Obungoloch, Godwin Ogbole, Marco Palombo, Andrew WebbBJR|Open 2024 · DOI

TL;DR

This review argues for the widespread adoption of low-field MRI (LF-MRI) as a democratizing force in global healthcare and research, particularly in low- and middle-income countries (LMICs). By leveraging recent advancements in hardware, software, and AI, LF-MRI can provide clinically relevant images and rich research data at a fraction of the cost and complexity of high-field systems, fostering local innovation and equitable access to advanced diagnostics.

What Problem Does This Paper Address?

Magnetic Resonance Imaging (MRI) is a powerful diagnostic tool, but its high cost, complex infrastructure requirements, and need for specialized personnel make it largely inaccessible in low- and middle-income countries (LMICs). This creates a significant global divide, limiting access to essential diagnostics for millions and underrepresenting diverse populations in crucial neuroimaging research. The paper addresses this inequity by advocating for portable, low-field MRI (LF-MRI) systems as a viable and impactful solution.

Methods

The paper doesn't present new experimental methods but rather reviews and synthesizes existing advancements in LF-MRI technology. Key areas discussed include:

  • Hardware Engineering: Advances in high-remanence magnetic materials and efficient MR pulse sequences.
  • Acquisition Physics: Novel approaches to optimize imaging at lower field strengths.
  • Image Reconstruction: Techniques like AI-based approaches, compressed sensing, and sparse sampling to improve image quality and reduce scan times.
  • AI-based Image Enhancement: Methods like Image Quality Transfer and SynthSR to approximate high-field image contrasts from low-field data.
  • Open-Source Design: A strong emphasis on developing LF-MRI systems that are modular, upgradeable, and allow for local maintenance and adaptation, fostering a sustainable ecosystem.

Datasets

Not discussed in the paper. The paper focuses on the technology and ecosystem for LF-MRI rather than presenting specific experimental results from new datasets.

Key Insights and Ideas

  • Low Field ≠ Low Quality: The paper challenges the outdated assumption that low-field MRI inherently means low diagnostic utility. Recent technological advancements have significantly improved LF-MRI image quality, making it suitable for a wide range of clinical and research applications.
  • Democratization through Openness: The authors strongly advocate for open-source LF-MRI systems. This approach empowers local teams to maintain, adapt, and scale technology, fostering independence and innovation, rather than relying on proprietary, vendor-locked systems.
  • Beyond the Scanner: The Ecosystem Matters: Successful LF-MRI deployment requires more than just hardware. The paper highlights the critical need for robust data infrastructure, comprehensive training pathways, ethical data governance, and equitable collaboration models.
  • Africa as a Catalyst for Discovery: The unique genetic diversity and environmental exposures in Africa present unparalleled opportunities for breakthrough discoveries in brain health. LF-MRI can unlock these insights by enabling large-scale population studies in these historically underrepresented regions.
  • A Call to Action for a Global Paradigm Shift: The paper issues a strong call to researchers, vendors, and funders to reframe MRI as a globally accessible technology, moving beyond a "donate-and-forget" culture towards sustainable, co-developed partnerships.

Key Results

The paper doesn't present quantitative results from a specific study but rather synthesizes the impact of LF-MRI based on existing literature. Key qualitative findings include:

  • LF-MRI systems can produce clinically meaningful images for a range of neurological conditions, including stroke, epilepsy, and neurodegenerative disorders.
  • Advances in AI and image reconstruction can significantly enhance image quality, making LF-MRI competitive for research applications.
  • Portable LF-MRI systems can enable imaging in previously inaccessible settings, such as critical care units or remote locations.
  • Open-source LF-MRI platforms can foster local capacity building and reduce reliance on expensive international service contracts.

Limitations and Future Work

The authors acknowledge several limitations and areas for future work:

  • Certification and Standardization: Developing scalable approaches to certify open, modular LF-MRI systems to ensure safety and trust without stifling local innovation is a challenge.
  • Gradient Amplitude: While higher gradient amplitudes are desirable for certain advanced LF-MRI techniques (like diffusion imaging), their cost and power requirements need careful consideration to maintain affordability.
  • Cooling Systems: Integrated cooling systems for LF-MRI scanners are often lacking, which can lead to performance drift, especially in hot climates, requiring further research into stable solutions.
  • AI Model Efficiency: Current deep learning models often have high computational and energy demands, making them unsuitable for low-resource environments. Developing lightweight, energy-efficient AI models tailored for LMICs is crucial.
  • Data Infrastructure: Bridging digital health gaps requires tailored data solutions, including on-device processing, hybrid storage models, and robust offline capabilities.
  • Broader Applications: While much of the recent innovation has focused on neuroimaging, the paper notes that LF-MRI has potential for other anatomical targets (musculoskeletal, abdominal, etc.) with appropriate hardware and acquisition strategies, which warrants further exploration.

Summary updated

Abstract

MRI is a cornerstone of modern clinical medicine and neuroscience, yet it remains largely inaccessible in low- and middle-income countries (LMICs) due to high costs, complex infrastructure requirements, the need for specialized personnel, and dependence on proprietary systems. Portable low-field MRI (LF-MRI), operating below 100 mT, offers a compelling alternative: low-cost, more accessible, and increasingly powerful, thanks to advances in hardware engineering, acquisition physics, image reconstruction, and open-source software. Reviewing and building upon recent progress, we, a multidisciplinary team of clinicians, physicists, engineers, and global health researchers based both in LMIC and HIC settings, present a formal argument for the adoption of LF-MRI as a catalyst for discovery science and healthcare innovation in LMICs. LF-MRI can produce clinically meaningful images and rich research data, enabling population-scale studies in neurodevelopment, ageing, and neurogenetics. But we argue that systems must be open, upgradeable, and co-developed, allowing potential for local teams to maintain, adapt, and scale technology according to their needs. Beyond the scanner, we outline the ecosystem required for success: data infrastructure, training pathways, ethical data governance, and equitable collaboration. We issue a call to researchers, vendors, and funders to reframe MRI as a globally accessible technology, capable of supporting diverse research agendas and delivering transformative health impact, particularly where it is needed most.
BibTeX
@inproceedings{jones2024low,
  author = {Derek K. Jones and Daniel C. Alexander and Karen Chetcuti and Mara Cercignani and Kirsten A. Donald and Mark A. Griswold and Emre Kopanoğlu and Ike Langunju and Johnes Obungoloch and Godwin Ogbole and Marco Palombo and Andrew Webb},
  title = {{Low field, high impact: democratizing MRI for clinical and research innovation}},
  booktitle = {BJR|Open},
  year = {2024},
  doi = {10.1093/bjro/tzaf022},
}
Figure