Diagnostic Model for Cerebral Small Vessel Disease Based on Subregional Volumes and Quantitative Susceptibility Mapping

Motivation: By integrating MRI-based radiomics and clinical cognitive assessment, the study aims to establish a more quantitative and reliable approach to early diagnosis, focusing on specific brain regions associated with CSVD-associated cognitive decline. Goal(s): The primary objective is to evaluate the effectiveness of a hybrid diagnostic model combining radiomic features from QSM with cognitive and motor assessments. Approach: This study developed a diagnostic model for CSVD using QSM imaging and radiomics features, further enhanced by incorporating cognitive and motor test results. Results: The hybrid model demonstrated high diagnostic efficiency, with an AUC of 0.87 in the test set, outperforming the radiomics-only model. Impact: These findings suggest that the combination of radiomic features and cognitive tests can improve the quantitative diagnosis of CSVD at an early stage, and provide an important theoretical basis and practical method for accurate diagnosis and early intervention.