Towards quantitative diagnosis and measurement of chronic spinal diseases: the role of image processing
Abstract: The technology of acquiring images of the spine (CT, MRI, static or dynamic X-ray imaging) is, in 2014, relatively mature. In contrast, the interpretation of spinal images remains qualitative (disease present/absent or, at best mild/moderate/severe), much like it was 30 years ago, rather than quantitative. Quantitative diagnosis makes it possible to study the natural history of disease and to better understand the effect of interventions. The proliferation of image processing algorithms in the last 20 years, more recent advances in machine learning, and, finally and most recently, establishment of anonymized image databases have significantly lowered the barriers to the development of clinically useful quantitative measures of spinal disease. In this talk, I will describe some of the common major kinds of chronic spinal pathology, emphasizing for each the specific ways in which image analysis tools and algorithms could be applied to quantitative analysis for disease measurement.
Prof. David R. Haynor
Professor of Neuroradiology
University of Washington