Machine Learning Approaches To Segment And Cluster Cells Of The Cartilage And Capsule In Rat Elbow Histology Sections

Purpose: Cartilage and capsule are tissues of interest for trauma-induced joint disorders, such as post-traumatic contracture and osteoarthritis of the elbow. Our group has developed a rat preclinical rat model to study post-trauma elbow disorders and commonly uses histological analysis to evaluate the morphological and biological properties of soft tissues on the cellular level. However, histological analysis: i) is time-consuming, ii) relies on a pathologist’s inherently subjective evaluation, and iii) can be overly laborious when extracting quantitative data from tissue sections via histomorphometry. Furthermore, unbiased identification of various cell types might identify novel cellular targets, particularly in relatively understudied elbow disorders. New techniques leveraging machine learning (ML) may accelerate and objectify histology analysis, thereby dramatically advancing soft tissue analysis in healthy and diseased states. Thus, this study developed ML models to segment and cluster the tissues and cells of the cartilage and capsule in rat elbow histology sections.