DOI: 10.20935/mhealthwellb8385 ISSN: 2997-9196

Visual MRI ratings and automated volumetry in Ugandan dementia care

Rita Nassanga, Noeline Nakasujja, Mark Kaddumukasa, Martha Sajatovic, Aloysius Gonzaga Mubuuke, Roy Clark Basiimwa, Michael Grace Kawooya, James K. Tumwine, Kamada Lwere, Joy Louise Gumikiriza-Onoria, Stephen E. Jones
Introduction: Dementia diagnosis in sub-Saharan Africa remains largely clinical, with limited neuroradiologic confirmation. This study assessed how visual magnetic resonance imaging (MRI) rating scales correspond with automated volumetry in older Ugandan adults with clinically suspected dementia and explored associations with global cognition.

Materials and methods: We conducted a cross-sectional hospital- and community-based study of adults aged ≥ 50 years with clinically suspected dementia who underwent the Mini-Mental State Examination (MMSE) and 1.5T brain MRI using a NeuroQuant®-compatible protocol. Automated outputs included age- and sex-adjusted normative percentiles; hippocampal occupancy (HOC) ≤ 5th percentile was classified as abnormal. Visual ratings included medial temporal atrophy, global cortical atrophy, Koedam posterior atrophy, and Fazekas white matter hyperintensity scores. Spearman rank correlations assessed predefined visual–quantitative pairs and exploratory associations with MMSE.

Results: Of 72 eligible participants, 63 (87.5%) completed MRI and automated volumetry successfully. Mean age was 75.6 years, and 77.8% were female. Abnormal HOC was present in 38.1%, white matter hyperintensities in 61.9%, cortical infarcts in 41.3%, and entorhinal cortex atrophy in 11.1%. Medial temporal atrophy (MTA) correlated inversely with HOC (ρ = − 0.40; p = 0.001), global cortical atrophy (GCA) with whole-brain volume percentiles (ρ = − 0.56; p < 0.001), and frontal atrophy with frontal cortical volume (ρ = − 0.32; p = 0.011). Higher hippocampal asymmetry was associated with lower MMSE scores (p = 0.009).

Conclusions: Standardized visual MRI rating scales showed moderate correspondence with automated volumetry, supporting a pragmatic dementia imaging pathway in low-resource settings.

More from our Archive