Machine learning for spine imaging
Dr. Liang Liang
University of Miami
About this study
Degenerative disk disease, osteoarthritis of the spine, is a condition in which the discs (cushioning in your spine) break down. An estimated 266 million individuals (3.63%) worldwide have lumbar degenerative spine disease each year, a condition which can cause debilitating lower back pain. Machine Learning is a method of data analysis that automates analytical model building. Machine Learning is a branch of artificial intelligence (AI), a concept that systems can learn from data, identify patterns, and make decisions with little human intervention. In recent years, Machine Learning has revolutionized fields and has been embraced for analyzing medical images for diagnosis. Currently, intervertebral disc degeneration (IDD) is diagnosed using advanced imaging techniques, especially magnetic resonance imaging (MRI), and must be read manually by a Radiologist. By utilizing modern Machine Learning technology, we propose to develop a machine learning system (i.e. a computer program with ML algorithms) for fully automatic spine image analysis, which will lead to significant enhancement in the diagnosis of IDD, improving both quality and affordability of care.
An estimated 266 million individuals (3.63%) worldwide have lumbar degenerative spine disease each year, which can cause debilitating lower back pain and immobility. Information obtained from advanced imaging, particularly magnetic resonance imaging (MRI), is necessary for proper diagnosis and treatment planning but is currently labor-intensive and costly. The proposed ML system will lead to significant enhancement in radiology for the diagnosis of intervertebral disc degeneration, to provide better and more affordable services to the patients. Furthermore, the ML framework can be generalized for the geometry reconstruction of other organs and other imaging modalities. Thus, this project will have a profound clinical impact, as MRI imaging has become a key component of diagnosis for many health conditions.
Subscribe to updates on this research
Favorite this research