Objective: 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.
Methods: 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.
Results: 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).
Interpretation: 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.
Keywords: HIV; aging; artificial intelligence; atrophy; low‐field MRI; white matter hyperintensities.
© 2025 The Author(s). Annals of Clinical and Translational Neurology published by Wiley Periodicals LLC on behalf of American Neurological Association.