Artificial Intelligence in Spine and Paraspinal Muscle Analysis

Abstract

Disorders affecting the neurological and musculoskeletal systems represent international health burdens. A significant impediment to progress with interventional trials is the absence of responsive, objective, and valid outcome measures sensitive to early disease or disorder change. A key finding in individuals with spinal disorders is compositional changes to the paraspinal muscle and soft tissue (e.g., intervertebral disc, facet joint capsule, and ligamentous) structure. Quantification of paraspinal muscle composition by MRI has emerged as a sensitive marker for the severity of these conditions; however, little is known about the composition of muscles across the lifespan. Knowledge of what is "typical" age-related muscle composition is essential in order to accurately identify and evaluate "atypical," with a potential impact being improvements in pre- and postsurgical plan and measurement of surgical implants, exoskeletons, and care on a patient-by-patient basis.

Keywords: Biological aging; Convolutional neural networks; MRI; Skeletal muscle; Spine.