Using Artificial Intelligence to Predict the Development of Kyphosis Disease: A Systematic Review

Abstract

The use of artificial intelligence in the field of medicine - including spine surgery - is now widespread and prominent. Kyphosis is a prevalent disease in spine surgery with abundant morbidity. Predicting the development of kyphosis disease has been somewhat difficult, and the use of AI to aid in the prediction of kyphosis disease may yield new opportunities for spine surgeons. The aim of this review is to recognize the contributions of AI in predicting the development of kyphosis. Five databases/registers were searched to identify suitable records for this review. Nine studies were included in this review. The studies demonstrated that AI could be utilized to predict the development of kyphosis disease after corrective surgery for a variety of spinal pathologies, including thoracolumbar burst fracture, cervical deformity, previous kyphosis disease, and adult degenerative scoliosis. The studies utilized a variety of AI modalities, including support vector machines, decision trees, random forests, and artificial neural networks. Two of the included studies also compared the use of different AI modalities in predicting the development of kyphosis disease. The literature has demonstrated that AI can be utilized effectively to predict the development of kyphosis disease. However, the current research is limited and only sparsely covers this broad field. Therefore, we suggest that further research is needed to explore the uncharted opportunities in predicting the development of kyphosis disease. Also, further research would confirm and consolidate the benefits demonstrated by the literature included in this review.