Predicting the Risk of Total Hip Replacement by Using A Deep Learning Algorithm on Plain Pelvic Radiographs:...

Area of Science:

• Artificial Intelligence in Medicine
• Medical Imaging Analysis
• Orthopedic Surgery Decision Support

Background:

• Total hip replacement (THR) is a standard treatment for degenerative hip disorders.
• Timely identification of patients needing THR is crucial to avoid complications from delayed treatment or unnecessary conservative measures.
• Deep learning (DL) shows promise in medical imaging but lacks application in predicting short-term THR needs.

Purpose of the Study:

• To develop a DL-based assistant system, SurgHipNet, for predicting the need for THR within 3 months using pelvic radiographs.
• To aid physicians in clinical decision-making for optimal THR timing.

Main Methods:

• A convolutional neural network-based DL algorithm was developed to analyze pelvic radiographs.
• The algorithm predicted hip region of interest (ROI) and determined the necessity of THR.
• A dataset of 4643 hip ROIs (3013 surgical, 1630 non-surgical) from 2008-2017 was used, split into training, validation, and testing sets.

Main Results:

• SurgHipNet achieved an area under the receiver operating characteristic curve of 0.994.
• The model demonstrated high performance with an accuracy of 0.977, sensitivity of 0.920, specificity of 0.932, and F1-score of 0.944.

Conclusions:

• The developed SurgHipNet system shows significant potential for clinical decision support in orthopedic surgery.
• This AI tool can assist physicians in promptly identifying patients who require total hip replacement, optimizing treatment timing.