Artificial intelligence in spinal deformity

Abstract

Objective: To facilitate discussion surrounding the use of AI in medicine so that it may lead to improved surgeon and patient outcomes in the future. Background: Artificial intelligence (AI) refers to algorithms that utilize data to mimic human cognition. Machine learning (ML) is a subset of AI that enables the algorithm to improve without explicit human direction. Methods: Narrative overview of the literature synthesized from searches of computerized databases and authoritative texts. Discussion: There are forms of ML in use both in research and clinical settings today, which can personalize medical care in the future. In spine surgery, AI can affect care in the following domains: pre-operative workup, surgical planning, and outcome prediction. The use of AI in adult spinal deformity (ASD) poses unique opportunities for growth, as outcomes after ASD are often difficult to predict due to disease complexity. Recently, the use of AI modeling has gained traction with large multi-institutional organizations, leading to robust publications aided by an abundance of prospectively collected data. However, current AI usage still has concerns that should not be taken lightly. Conclusion: In the following review, we outline the basis of AI, its current clinical uses and potential benefits, and its various challenges.

Keywords

• Spinal deformity
• Artificial intelligence
• Machine learning
• Deep learning
• Neural network
• Patient reported outcome measures