Brain Volumetric Analysis Using Artificial Intelligence Software in Premanifest Huntington’s Disease Individuals from a Colombian Caribbean Population

Abstract

Background and objectives: The premanifest phase of Huntington’s disease (HD) is characterized by the absence of motor symptoms and exhibits structural changes in imaging that precede clinical manifestation. This study aimed to analyze volumetric changes identified through brain magnetic resonance imaging (MRI) processed using artificial intelligence (AI) software in premanifest HD individuals, focusing on the relationship between CAG triplet expansion and structural biomarkers. Methods: The study included 36 individuals descending from families affected by HD in the Department of Atlántico. Sociodemographic data were collected, followed by peripheral blood sampling to extract genomic DNA for quantifying CAG trinucleotide repeats in the Huntingtin gene. Brain volumes were evaluated using AI software (Entelai/IMEXHS, v4.3.4) based on MRI volumetric images. Correlations between brain volumes and variables such as age, sex, and disease status were determined. All analyses were conducted using SPSS (v. IBM SPSS Statistics 26), with significance set at p Results: The analysis of brain volumes according to CAG repeat expansion shows that individuals with ≥40 repeats evidence significant increases in cerebrospinal fluid (CSF) volume and subcortical structures such as the amygdalae and left caudate nucleus, along with marked reductions in cerebral white matter, the cerebellum, brainstem, and left pallidum. In contrast, those with Conclusions: These findings suggest that CAG expansion selectively impacts key brain regions, potentially influencing the progression of Huntington’s disease, and that AI in neuroimaging could identify structural biomarkers long before clinical symptoms appear.

Keywords

• artificial intelligence
• Huntington’s disease
• magnetic resonance imaging
• neuroimaging
• structural MRI