Tracking diaphragm and chest wall movement on cine-MRI

Congenital muscular dystrophies (CMD) are amongst the most common early onset neuromuscular disorder that results in progressive weakness of skeletal, and respiratory muscles. Current methods based on breathing tests are not able to accurately characterize the respiratory muscle involvement. We examine the diaphragm and chest wall dynamics using cine magnetic resonance imaging (MRI). We propose a holistic neural network to handle the segmentation of lungs on poor quality cine MR images and develop a knowledge-based method to extract the diaphragm and chest wall from the image. The method was tested on a cohort of young patients with COL6- and LAMA2-related muscular dystrophy. Automated respiratory muscle movement measurements achieve comparable accuracy as manual measurements.