Quantification of Spatial Ventilation Defect Distribution in Hyperpolarized Gas MRI of Lungs Using A 3D Clustering Algorithm

Hyperpolarized gas (HPG) MR imaging allows for quantification of a patient’s lung function. The spatial distribution of ventilation defect patterns is often overlooked in quantitative analyses but may be important for further understanding the nature of lung disease. Here we present a method for quantifying the extent to which defect voxels tend to be sparsely distributed or clustered. This technique involves spherical region-growing for each defect voxel to assess the fraction of neighbors which are also part of a defect. A ‘clustering index’ is reported which quantifies the extent to which defect voxels are spatially congregated or scattered.