Software to automatically measure muscle fiber types and sizes.
Despite generations of technological advancements in disease interpretation and testing technologies, accurate diagnosis of myopathy remains a challenge. Myopathy, a muscle fibers dysfunction disease, is typically diagnosed with a variety of tests including muscle biopsy which involves collecting a small amount of muscle tissue samples and analyzing the sample via histology. The physician then identifies the muscle myopathy manually, which is labor-intensive, time-consuming, and error-prone leading to up to 10 percent of all patient deaths and up to 17 percent of all hospital complications. An automated method to accurately detect the morphology of muscle samples is required to reduce/eliminate the manual diagnostic errors.
The Emory inventors have developed a machine learning based analytical tool called, Myosoft, to automate the analysis of muscle morphology and muscle fiber type in samples stained with fluorescence antibodies. Myosoft uses existing, open source machine learning algorithm designed for use in Image J or FIJI, an open access software platform used commonly by researchers to analyze images. The machine learning algorithm is pre-trained with datasets.