The authors of a new study published in the American Association for the Advancement of Science (AAAS) have found that machine learning enables cheaper and safer low-power magnetic resonance imaging (MRI) without sacrificing accuracy. These advances pave the way for affordable, patient-centric, and deep learning-powered ultra-low-field (ULF) MRI scanners, addressing unmet clinical needs in diverse healthcare settings around the world.
MRI scans offer noninvasive and radiation-free imaging, but despite its five decades of development it can remains inaccessible, particularly in low- and middle-income countries due to the high costs associated with standard superconducting MRI scanners and the specialised infrastructure required for their operation.
These scanners are typically housed in specialised radiology departments or large imaging centres, limiting their availability in smaller medical facilities.
Additionally, the necessity for radiofrequency (RF)-shielded rooms and substantial power consumption further limits access to MRI technology. To address MRI accessibility challenges, Yujiao Zhao and colleagues present a low-power and highly simplified ULF MRI scanner that operates on a standard wall power outlet and without the need for RF or magnetic shielding.
The scanner uses a compact 0.05 Tesla (T) magnet (most MRI devices use a 1.5 T magnet, but some can go as high as 7 T) and incorporates active sensing and deep learning to address electromagnetic interference signals and improve image quality.
In addition, the device used only 1800 watts (W) during scanning, while conventional MRIs can consume 25000 W or more. Zhao et al. conducted imaging on healthy volunteers and show that the device was able to produce clear and detailed imaging on par with that obtained by high-power MRI devices currently used in the clinic.