Computer Vision News - July 2021

2 Summary AI Rese ch 4 Axial Spondyloarthritis Axial Spondyloarthritis (axSpA) is defined as a chronic, inflammatory rheumatic disease that affects primarly the axial skeleton, causing severe pain, stiffness and fatigue. The disease typically starts in early adulthood. The current clinical standard of detection is with a short-tau inversion recovery (STIR) magnetic resonance imaging (MRI). In order to evaluate and treat the disease, the need of identification and quantification of inflammation is crucial. As counter-intuitive as it may seem, at the moment the only way to describe the inflammation in clinical practice is a visual assessment of the MRI with a verbal description that includes no numerical metric to allow quantification. There is clinical research on using semi-quantitative scoring to describe bone marrow edema (BME, the inflammation in the subchondral bone marrow), but those are limited. They take a very long time to be calculated as well and they only take information from very few slices. The research of the month is: Towards Deep Learning-assisted Quantification of Inflammation in Spondyloarthritis: Intensity-based Lesion Segmentation. It is currently pre-published on ArXiv by authors Carolyna Hepburn, Hui Zhang, Juan Eugenio Iglesias, Alexis Jones, Alan Bainbridge, Timothy JP Bray and Margaret A Hall- Craggs. Towards Deep Learning-assisted Quantification of Inflammation in Spondyloarthritis: Intensity-based Lesion Segmentation A new article for a new month! Happy fourth of July to our US readers This month’s article diverts from the previous ones, concentrating more on the quantification side of medical imaging and specifically the quantification of inflammation in spondyloarthritis. It’s not a surprise anymore that some techniques of deep learning will be used! Read on to explore more of this very interesting topic.