Automatic segmentation of Optic Pathway Gliomas in MRI

This paper presents an automatic method for the segmentation of Optic Pathway Gliomas (OPGs) from multi-spectral MRI datasets. The method starts with the automatic localization of the OPG and its core with an anatomical tumor atlas followed by a binary voxel classification with a probabilistic tissue model whose parameters are estimated from MR images. The method effectively incorporates prior location, shape, and intensity information to accurately identify the sharp OPG boundaries and to delineate in a consistent and repeatable manner the OPG contours that cannot be clearly distinguished on conventional MR images. Our experimental study on 15 datasets yield a mean surface distance error of 0.67 mm and mean volume overlap difference of 28.6% as compared to manual segmentation by an expert radiologist. To the best of our knowledge, this is the first method that addresses automatic OPG segmentation.