Prediction of Fracture Union in Long Bones Using AI Models

This chapter sums up the recent findings on artificial intelligence (AI) application to the healing of long bone fractures, such as important algorithms, data sources and clinical variables to be included in the model construction. It additionally provides discussion of how the imaging modalities can be integrated with AI to increase the accuracy of the diagnosis. The chapter also mentions such challenges as the quality of data, the explainability of models and their implementation in clinical practice. It shows that future AI-based decision support systems can personalize treatment, enhance clinical decision-making, and eventually improve the healing process. There are four broad categories of AI models, which comprise supervised and deep, unsupervised, and hybrid/ensemble models. AI models predictive of fracture union have been assessed in a number of studies and have proven to be significantly accurate and applicable in the clinical setting.