ISTA 2025 - Deep Learning-based Automatic Measurement of Liner Wear and Acetabular Cup Alignment

Description

In this presentation, Dr. Beck from the National University of South Korea addresses the significant issues caused by malposition of acetabular components in total hip arthroplasties, including osteoarthritis and dislocation, which can lead to revisions. The talk highlights traditional measurement methods that suffer from ethical concerns and reliability issues, prompting the development of a new deep learning tool designed to enhance measurement accuracy and reduce biases. Dr. Beck outlines the workflow of the deep learning model, detailing the steps of region localization, segmentation of components, and final alignment measurement. They discuss the demonstration of high detection accuracy, segmentation success, and close correspondence with ground truth data concerning inclination and anteversion metrics. The study concludes with an acknowledgment of limitations while emphasizing the tool's reliability and accuracy improvements over previous methods, advocating for its use in clinical settings to enhance orthopedic practices.

Specialties

Conferences