Portable Low-Field Magnetic Resonance Imaging in People With Human Immunodeficiency Virus.
Authors
Affiliations (9)
Affiliations (9)
- Department of Neurology, Massachusetts General Hospital and Harvard Medical School, Boston, Massachusetts, USA.
- Center for Genomic Medicine, Massachusetts General Hospital and Harvard Medical School, Boston, Massachusetts, USA.
- Division of Infectious Diseases, Massachusetts General Hospital and Harvard Medical School, Boston, Massachusetts, USA.
- Athinoula A. Martinos Center for Biomedical Imaging, Department of Radiology, Massachusetts General Hospital and Harvard Medical School, Boston, Massachusetts, USA.
- Department of Neurology, Center for Brain & Mind Health, Yale New Haven Hospital and Yale School of Medicine, New Haven, Connecticut, USA.
- Division of Neuroradiology, Department of Radiology, Massachusetts General Hospital and Harvard Medical School, Boston, Massachusetts, USA.
- Centre for Medical Image Computing, University College London, London, UK.
- Computer Science and Artificial Intelligence Laboratory, Massachusetts Institute of Technology, Cambridge, Massachusetts, USA.
- Vaccine and Immunotherapy Center, Massachusetts General Hospital, Boston, Massachusetts, USA.
Abstract
The aging population of people with HIV (PWH) raises heightened concerns regarding accelerated aging and dementia. Portable, low-field MRI (LF-MRI) is an innovative technology that could enhance access and facilitate routine monitoring of PWH. We sought to evaluate the feasibility of LF-MRI and apply a machine learning (ML) segmentation algorithm to examine atrophy and white matter hyperintensities (WMH) in PWH compared to people without HIV (PWoH) of similar age. Individuals with a confirmed diagnosis of HIV on antiretroviral therapy underwent LF-MRI (64 mT) acquisition in the outpatient neurology clinic. PWoH with > 1 vascular comorbidity (VC cohort, n = 25) or with mild cognitive impairment (MCI cohort, n = 24) due to Alzheimer's disease served as comparators. LF-MRI brain region segmentations were derived using the ML algorithm WMH-SynthSeg in FreeSurfer. Brain regions corrected for intracranial volume were compared between cohorts after adjusting for age and sex. Thirty virally suppressed PWH were included. LF-MRI derived brain volumes from PWH demonstrated a reduction in volume of the caudate relative to PWoH with VC (p < 0.05). Volume of the putamen and white matter was reduced in PWH compared to VC (p < 0.05). Hippocampal volume was comparable between PWH and PWoH (p ≥ 0.05), while volume of the amygdala was reduced in those with MCI alone (p < 0.05). No differences in WMH were seen between these cohorts (p > 0.05). LF-MRI is feasible in an outpatient setting, and ML algorithms enable detection of regional atrophy and WMH in PWH. LF-MRI may enable more frequent monitoring and earlier detection of atrophy in at-risk populations.