Identifying Lesions in Paediatric Epilepsy using Morphometric and Textural Analysis of Magnetic Resonance Images

Speaker: Sancgeetha Kulaseharan (MCSC)

Abstract:
Focal cortical dysplasia (FCD) is a brain malformation that is frequently responsible for epilepsy in children and accounts for approximately 26% of paediatric cases of intractable partial epilepsy. Surgical treatment through resection of the lesion responsible for epilepsy is an option in selected patients, however, FCD can be subtle and nearly undetectable using standard radiological assessment. Automated lesion detection through post-processing techniques using surface-based features and textural analysis in conjunction with classification have been shown to improve pre-surgical detection. An aim of this study is to develop an image processing pipeline using a 2-Step Naive Bayes classifier with a combination of established morphometric features and Gray-Level Co-occurrence Matrices (GLCM) textural analysis on MRI sequences. In addition, a goal is to identify lesions in patients with MRI-visible FCD, and to determine if the same pipeline can identify more subtle features in patients with MRI-negative FCD. Our analysis yields promising results on patient datasets.