Segmenting the substantia nigra in ultrasound images for early diagnosis of Parkinson's disease

Early diagnosis of Parkinson's disease (PD) is of immense importance, since clinical symptoms do not occur until substantial parts of the substantia nigra (SN) in the brain stem have been irreparably damaged. Recent work suggests, that by means of transcranial sonography (TCS) it is possible to determine PD even in the preclinical state. In images of the mesencephalon, the SN shows a distinct hyperechogenic pattern on TCS, which is currently manually segmented. To remove this investigator dependence, we develop a semi-automatic algorithm to segment SN in TCS images. After some preprocessing steps, the actual segmentation works intensity-based with morphological operations, taking anatomical information into account. The resulting size of the SN serves as a risk factor for PD manifestation.