Article type
Abstract
Backgrounds: Artificial intelligence is an emerging and powerful technology with growing use in orthopedics. The global morbidity of total knee arthroplasty is increasing.
Objectives: This systematic review investigated the application of artificial intelligence in periprosthetic of knee joint.
Methods: A systematic review was conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines. Deadline for all database searches was July 2023. All studies investigated the application on artificial.
intelligence based X-ray total knee arthroplasty were included. Studies with no full text available, reviews and meta-analyses were all excluded. For each included study, the characteristics, AI learning applications, algorithms, statistical performance, and results were summarized. Limitations and solutions to current AI applications and research were identified, including its visualization graphs to explain the "black box" effect, migration learning to explain the small amount of data, and data augmentation to improve model robustness and generalization performance.
Results: Nineteen studies were included in the final analysis. The applications of artificial intelligence in X-ray-based total knee replacement fall into four main categories: predicting the make and model of knee prosthesis, investigating postoperative satisfaction with the total knee, and diagnosing complications such as infection and loosening after total knee replacement.
Conclusions: Researches on artificial intelligence-based applications for knee arthroplasty are rapidly evolving. Further research will benefit from expanded datasets, external studies, and advancing technologies. The primary task for future AI is to create an open software program and commercial tools that will provide clinicians with a basis for preoperative optimisation, resource allocation and clinical decision-making.
Objectives: This systematic review investigated the application of artificial intelligence in periprosthetic of knee joint.
Methods: A systematic review was conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines. Deadline for all database searches was July 2023. All studies investigated the application on artificial.
intelligence based X-ray total knee arthroplasty were included. Studies with no full text available, reviews and meta-analyses were all excluded. For each included study, the characteristics, AI learning applications, algorithms, statistical performance, and results were summarized. Limitations and solutions to current AI applications and research were identified, including its visualization graphs to explain the "black box" effect, migration learning to explain the small amount of data, and data augmentation to improve model robustness and generalization performance.
Results: Nineteen studies were included in the final analysis. The applications of artificial intelligence in X-ray-based total knee replacement fall into four main categories: predicting the make and model of knee prosthesis, investigating postoperative satisfaction with the total knee, and diagnosing complications such as infection and loosening after total knee replacement.
Conclusions: Researches on artificial intelligence-based applications for knee arthroplasty are rapidly evolving. Further research will benefit from expanded datasets, external studies, and advancing technologies. The primary task for future AI is to create an open software program and commercial tools that will provide clinicians with a basis for preoperative optimisation, resource allocation and clinical decision-making.