Exactech Unveils Large-Scale CT Radiomic Analysis of Shoulder Bones and Muscles

Exactech shared results of a landmark research study comprising the first large-scale, fully automated CT radiomic analysis of the bones and muscles in the shoulder.

Using an AI-based framework, researchers analyzed pre-operative CT images and clinical outcomes from more than 4,000 patients treated with Exactech’s Equinoxe shoulder system and automatically extracted radiomic features—quantitative numerical representations of segmented image shape, pixel density and textures—from the deltoid muscle and scapular bone.

Numerous radiomic measurements were identified that were predictive of pain, motion, and function following shoulder arthroplasty. Identification of clinically relevant radiomic measurements, that can be automatically extracted from pre-operative CT images, has potential to expand the capabilities of pre-operative planning software and provide more-personalized predictions, with additional clinical decision support.

In addition to identifying the most clinically relevant radiomic features associated with shoulder arthroplasty clinical outcomes, this study also proposed a novel unsupervised machine learning based clustering analysis as a method to create new classification systems for bone and muscle morphology. This clustering analysis identified several unique morphologies of the deltoid muscle and scapula bone that were associated with differences in clinical outcomes before and after shoulder arthroplasty.

“Our team has created a novel automated image analysis tool that will change how orthopedic surgeons utilize CT image data to diagnosis their patients and plan treatment with shoulder arthroplasty,” said Chris Roche, Exactech’s Senior Vice President of Extremities. “Automated radiomic analysis of pre-operative CT image data allows us to fully characterize the bones and muscles of an individual patient’s shoulder joint, evolving diagnosis from a subjective classification to quantification on a continuum. This evolution will represent a paradigm shift for the future of pre-operative planning and clinical decision support software. Capabilities like these are only possible with the help of artificial intelligence.”

Source: Exactech