Jing Cheng, Jiaxing Yang, Xiaofeng Jia, Qilin Yang
This study investigates the characteristic changes of brain microstructures in end-stage renal disease patients with and without cognitive impairment (CI) and further predicts individual cognitive function in end-stage renal disease patients. In total, 32 healthy subjects, 22 end-stage renal disease patients with CI, and 22 end-stage renal disease patients without CI were included in our study. All subjects completed diffusion kurtosis imaging (DKI) examinations and the Montreal Cognitive Assessment (MoCA) test. Between-group differences in the DKI metrics were analyzed. In addition, correlation analysis between DKI data, clinical data, and MoCA scores was performed. Finally, prediction of cognitive function scores associated with end-stage renal disease patients with CI was conducted based on a support vector regression machine. We observed characteristic changes of brain microstructures in end-stage renal disease patients with and without CI, as indicated by significantly altered DKI parameters. In addition, a significant positive correlation was found between MoCA scores, axial kurtosis (AK) values, radial kurtosis values, and hemoglobin level. A significant positive correlation was found between hemoglobin level and AK values. And a significant negative correlation was found between creatinine level and AK values. Furthermore, this prediction model shows potential for estimating cognitive scores associated with end-stage renal disease, although further validation in larger cohorts is required. Our integrated analytical framework, combining voxel-wise group comparisons of DKI metrics, correlation analyses, and predictive modeling, identifies characteristic brain microstructural alterations and offers new insights into cognitive dysfunction among patients with end-stage renal disease. Notably, the significant DKI metrics extracted in this work may serve as potential imaging biomarkers for predicting cognitive performance related to end-stage renal disease.