Artificial intelligence for predicting various conditions in spine surgery: a systematic review

Abstract

Introduction Artificial intelligence (AI) includes software systems combined with applied methods and algorithms which main feature is the ability to solve intellectual problems. One of the most popular trend of AI application is the prognosis of various situations, evaluation of any digital information with an attempt to give a conclusion and analysis of different data with a search for hidden patterns. Study design Systematic review. Purpose To assess current possibilities of artificial intelligence (AI) for predicting unfavorable conditions in spine surgery that require medical care, and future development prospects in this area. Materials and methods Until June 2020 we performed a search using the Protocol (PRISMA) "Preferred Reporting Parameters for Systematic Reviews and Meta-analyzes" and keywords for articles in Medline, Scopus and eLIBRARY intended to summarize the available data on algorithms for predicting any pathological conditions in spine surgery that require medical intervention using artificial intelligence technologies. Results 20 publications were selected for systematic review, which presented data on the application of artificial intelligence, machine learning and neural networks to predict any condition in spine surgery. According to the review, the data obtained indicate that AI can be successfully used to optimize prognosis in various diseases of the spine. Therefore, the application of AI in the clinical practice of the spine surgeons can improve treatment results. Conclusion The promotion of artificial intelligence application in medicine is inevitable. Currently AI shows good results in the area of making clinical decisions by the surgeons and the ability to predict treatment results depending on certain factors. It is imperative that spinal surgeons should realize the potential of these new technologies. Nevertheless, some factors that determine the clinical application of artificial intelligence, i.e. the ability to consider the context of a patient's history, are difficult to calculate mathematically and, so far, are difficult for an algorithmic approach. Eventually, the doctors will continue to play a vital role in patients’ treatment, and artificial intelligence will not depreciate their clinical skills, but make them even more important.

Keywords

• artificial intelligence
• systematic review
• machine learning
• spine surgeons
• neural networks in spine surgery
• prognosis in spine surgery