Neuroradiology – Segmentation of Corpus Callosum

The corpus callosum (CC) has become very important in pathophysiological models of schizophrenia. Volume and structural integrity are usually examined by volumetric and diffusion tensor imaging (DTI) studies, but the results are often inconsistent across the methods and studies applied. Problems arise from the varying methodologies and the accuracy of measurements based on a single slice or small regions of interest. Usually volume and diffusion values are not examined in the same group of patients, and thus the relationship between these anatomical measures is not clear. In addition, not all regions are affected equally by anatomical changes. This emphasizes the importance of using different methods in the evaluation of white matter (WM) in schizophrenia to avoid false negative findings.

Based on Definiens Cognition Network Technology®, in a research study an application was developed which fully automatically segments and quantifies the shape and sub-segments of corpus callosum from sagittal T1-weighted MRT images of the human brain. Besides automation, it allows visual inspection and ergonomic correction of segmentation errors if needed. The application can also be applied to mask the corpus callosum  from diffusion tensor imaging (DTI) data. The insights gained can be generalized and transferred to the analysis of other anatomical structures and data from different modalities, like computer tomography.

For further details, please see the poster Fully automated segmentation of corpus callosum from sagittal MRT images.