Applications of Artificial Intelligence in osteoporosis and post-fragility fracture care: A Short Review

Background and aims: The increasing load of osteoporosis and fragility fractures highlightsthe need for better management of osteoporosis in the Healthcare system and is a proper contextfor digital health interventions including new artificial intelligence algorithms. Osteoporosis isone of the main causes of disability in old ages, reduced quality of life and loss of independence,therefore the use of artificial intelligence (AI) is an essential way to minimize the diagnostic errorsassociated with osteoporosis. The review study aims to address concerns and inform those interestedin the using artificial intelligence in osteoporosis management.Method: This review was conducted based on the review of articles available in pub med, googlescholar and med line databases from ۰۱/۰۱/۲۰۱۰ to ۰۱/۰۱/۲۰۲۳ with keywords AI (artificial intelligence),osteoporosis, fragility fracture and machine learning. It led to the inclusion of ۲۲ articlesin the review. ۵ articles were removed from the study due to similarity in ۳ bases, and ultimately۱۷ articles were read.Peer-reviewed articles cover ۵ areas of osteoporosis management. BMD predictive Variables(n=۱)Diagnosis, screening and classification of osteoporosis (n=۶) diagnosis and screening of fracture(n=۵)Forecast of fracture risk(n=۳) Auto-division of various images found (n= ۲)Results: Recent advances in machine learning (ML) have enabled the field of artificial intelligence(AI) to make dramatic advances in complex data environments in which human capacityfor high-dimensional relationships is limited. Different techniques to check bone health with AIwent beyond X-ray imaging such as: bone acoustics to regulate bone health, dental radiographyand BMD, use of MRI to assist with diagnose and distribution of energetic X-ray images dualenergy(DXA), image analysis and multi-row multi-detector (MDCT). Prediction algorithms usingdifferent input data sets based on known risk factors help physicians to calculate ۵ or ۱۰-yearfracture risk. The results of most of the studies showed better performance of CNN in fracture diagnosisthan doctors and orthopedists. This performance is improved by data enhancement techniquesof generative networks and digitally reconstructed radiographs, compared to those withoutbooster. AI-designed Channel Convolutional Neural Network are able to automatically detectcracks and trajectories at different levels of compression with high precision.Conclusion: Initial efforts to harness the power of machine learning algorithms like neural networksare still limited to the macroscale of bone, while there is a clear lack of their application atthe smaller scale, where the damage begins to erode. This approach is especially for developingof a powerful detection system to understand the initiation of bone microdamage propagation andpaves the way for the application of machine learning studies in bone micromechanics. Althoughthese recent advances have had successful initial application to osteoporosis research, their developmentis continuing to improve the assessing the effectiveness and affordability of this technologyrequires strong controlled studies.

کلیدواژه ها:

artificial intelligence ، osteoporosis ، fragility fracture ، bone density