Artificial intelligence and treatment algorithms in spine surgery

Yann Philippe Charles,Vincent Lamas,Yves Ntilikina

DOI: https://doi.org/10.1016/j.otsr.2022.103456

Abstract:Artificial intelligence (AI) is a set of theories and techniques in which machines are used to simulate human intelligence with complex computer programs. The various machine learning (ML) methods are a subtype of AI. They originate from computer science and use algorithms established from analyzing a database to accomplish certain tasks. Among these methods are decision trees or random forests, support vector machines along with artificial neural networks. Convolutive neural networks were inspired from the visual cortex; they process combinations of information used in image or voice recognition. Deep learning (DL) groups together a set of ML methods and is useful for modeling complex relationships with a high degree of abstraction by using multiple layers of artificial neurons. ML techniques have a growing role in spine surgery. The main applications are the segmentation of intraoperative images for surgical navigation or robotics used for pedicle screw placement, the interpretation of images of intervertebral discs or full spine radiographs, which can be automated using ML algorithms. ML techniques can also be used as aids for surgical decision-making in complex fields, such as preoperative evaluation of adult spinal deformity. ML algorithms "learn" from large clinical databases. They make it possible to establish the intraoperative risk level and make a prognosis on how the postoperative functional scores will change over time as a function of the patient profile. These applications open a new path relative to standard statistical analyses. They make it possible to explore more complex relationships with multiple indirect interactions. In the future, AI algorithms could have a greater role in clinical research, evaluating clinical and surgical practices, and conducting health economics analyses.