Low-Power MRI Scanners Powered by Deep Learning Could Revolutionize Accessibility

A new study has demonstrated that machine learning can enable cheaper and safer low-power magnetic resonance imaging (MRI) without compromising accuracy. These advancements could lead to affordable, patient-centric, and deep learning-powered ultra-low-field (ULF) MRI scanners, addressing unmet clinical needs in various healthcare settings worldwide.

Addressing MRI Accessibility Challenges

Despite the revolutionary impact of MRI on healthcare, it remains largely inaccessible, particularly in low- and middle-income countries. This is mainly due to the high costs of standard superconducting MRI scanners and the specialized infrastructure required for their operation. To address these challenges, Yujiao Zhao and colleagues have developed a low-power and highly simplified ULF MRI scanner that operates on a standard wall power outlet without the need for radiofrequency (RF) or magnetic shielding.

Deep Learning Enhances Image Quality

The scanner uses a compact 0.05 Tesla (T) magnet, significantly lower than the 1.5 T to 7 T magnets used in most MRI devices. It also incorporates active sensing and deep learning to address electromagnetic interference signals and improve image quality. Moreover, the device consumes only 1800 watts (W) during scanning, compared to the 25000 W or more consumed by conventional MRIs.

Zhao et al. conducted imaging on healthy volunteers and demonstrated that the device could produce clear and detailed imaging comparable to that obtained by high-power MRI devices currently used in the clinic. However, Udunna Anazodo and Stefan du Plessis note in a related Perspective that there are still limitations and challenges to be addressed before low-field MRI can be widely applied for clinical use. They write, “Low-field MRI has yet to mature to enable cost-effective access to medical imaging. Its potential as an essential and environmentally sustainable health technology will be proven when many communities around the world can use low-field MRI without barriers.”

Keyword phrase: low-power deep learning-powered ultra-low-field MRI scanners