Texture features of T2w magnetic resonance imaging mirror tissue changes in Parkinson’s disease models

After successful applications in the oncology field to provide new in vivo diagnosis and prognosis imaging features, texture analysis and more generally radiomics were also reported as having the potential to provide markers of different neurodegenerative processes. Indeed, in neurodegenerative diseases such as Parkinson's disease (PD), there is a need for neuroprotective therapies, the development of which will be fundamentally aided by imaging biomarkers capable of inferring tissue changes such as loss of neurons in the nigro-striatal pathway or alpha synuclein aggregates that characterize PD. In this study, we therefore sought to decipher the relationship between signal changes measured using brain MRI texture features and histological changes in preclinical models of this disease. Three rodent models were used: two toxin-based models, one involving 6-hydroxydopamine injection and the other using methyl-phenyl-tetrahydropyridine, and a third model based on alpha-synuclein overexpression. Animals had MR imaging with a T2w sequence evaluation and were sacrificed for histological analyses of the brains. Texture features were measured in different brain structures. The association analyses revealed significant correlations between the imaging features measured in the substantia nigra and the striatum with dopaminergic degeneration, as well as significant correlations between texture features in key structures (substantia nigra, striatum, thalamus, hippocampus and associative and cingulate cortices), and alpha-synuclein quantified in these regions. These preliminary results suggest that MR signal changes captured using texture features reflect the underlying tissue changes occurring in the brain such as neuronal death and proteins accumulation.