Computational Approach toward Pulmonary Functional Imaging

Quantitative functional imaging has improved clinical understanding of lung disease, enhanced researchers’ abilities to develop new treatments, and has the potential to improve care for millions of patients who suffer from COPD, asthma, cystic fibrosis, and other diseases. While all imaging modalities rely on physics and computation, functional lung imaging is especially amenable to and dependent on computational approaches. Functional imaging modalities, such as hyperpolarized noble gas magnetic resonance imaging (MRI), inhaled contrast computed tomography (CT), and nuclear imaging, require complex processing strategies for synthesis, quantification, and computer-aided diagnosis. Early investigation into the use of such modalities for the development of computational metrics specifically tailored for the quantitative assessment of lung function has led to advanced machine learning strategies for classifying patients and predicting disease progression. Such developments will have a lasting impact on future diagnostic capabilities and the potential treatment of lung disease. In this chapter, we review these functional lung imaging modalities, their corresponding computational methodologies, and potential clinical significance.