Abstract #4627

Towards a Better Understanding of Helium-3 MRI Manual Segmentation Error Using Fuzzy C-Mean Methods

Mohammadreza Heydarian¹, Andrew Wheatley¹, Grace Parraga^1,2

¹Robarts research Institute, London, Ontario, Canada; ²Medical Biophysics, University of Western Ontario, London, On, Canada

Hyperpolarized helium-3 MRI provides a way to visualize and quantify lung function based on segmentation of helium-3 ventilation images. Manual segmentation of 3He ventilation volumes is time consuming and prone to observer error. To address this limitation, we developed and applied a fully automated fuzzy c-mean (FCM) method for segmenting ventilated regions and observed significant associations between the automated and manual segmentation methods. FCM provides a fully automated, robust and efficient method for segmenting ventilated regions of hyperpolarized helium-3 images.