Image quality assessment of advanced reconstruction algorithm for point-of-care MRI scanner

Elizabeth A Krupinski; DeAngelo Harris; Lori R Arlinghaus; Jo Schlemper; Michal Sofka

doi:10.1117/1.JMI.10.S1.S11913

Image quality assessment of advanced reconstruction algorithm for point-of-care MRI scanner

J Med Imaging (Bellingham). 2023 Feb;10(Suppl 1):S11913. doi: 10.1117/1.JMI.10.S1.S11913. Epub 2023 May 22.

Authors

Elizabeth A Krupinski¹, DeAngelo Harris¹, Lori R Arlinghaus², Jo Schlemper², Michal Sofka²

Affiliations

¹ Emory University, Department of Radiology & Imaging Sciences, Atlanta, Georgia, United States.
² Hyperfine, Inc., Guilford, Connecticut, United States.

Abstract

Purpose: Portable magnetic resonance imaging (pMRI) has potential to rapidly acquire images at the patients' bedside to improve access in locations lacking MRI devices. The scanner under consideration has a magnetic field strength of 0.064 T, thus image-processing algorithms to improve image quality are required. Our study evaluated pMRI images produced using a deep learning (DL)-based advanced reconstruction scheme to improve image quality by reducing image blurring and noise to determine if diagnostic performance was similar to images acquired at 1.5 T.

Approach: Six radiologists viewed 90 brain MRI cases (30 acute ischemic stroke (AIS), 30 hemorrhage, 30 no lesion) with $T 1$ , $T 2$ , and fluid attenuated inversion recovery sequences, once using standard of care (SOC) images (1.5 T) and once using pMRI DL-based advanced reconstruction images. Observers provided a diagnosis and decision confidence. Time to review each image was recorded.

Results: Receiver operating characteristic area under the curve revealed overall no significant difference ( $p = 0.0636$ ) between pMRI and SOC images. Examining each abnormality, for acute ischemic stroke, there was a significant difference ( $p = 0.0042$ ) with SOC better than pMRI; but for hemorrhage, there was no significant difference ( $p = 0.1950$ ). There was no significant difference in viewing time for pMRI versus SOC ( $p = 0.0766$ ) or abnormality ( $p = 0.3601$ ).

Conclusions: The deep learning (DL)-based reconstruction scheme to improve pMRI was successful for hemorrhage, but for acute ischemic stroke the scheme could still be improved. For neurocritical care especially in remote and/or resource poor locations, pMRI has significant clinical utility, although radiologists should be aware of limitations of low-field MRI devices in overall quality and take that into account when diagnosing. As an initial triage to aid in the decision of whether to transport or keep patients on site, pMRI images likely provide enough information.

Keywords: observer performance; portable magnetic resonance imaging; viewing time.