State-of-the-art review in AI and digital technologies for spine deformity

Artificial intelligence (AI) and digital technologies are increasingly being used across the continuum of spinal deformity care, from screening and diagnosis to surgical planning, intraoperative assistance, and postoperative follow-up. This State-of-the-Art Review provides a clinically oriented synthesis of current applications, summarizes the available evidence supporting their use, and discusses the requirements for safe and effective clinical translation.

Peer-reviewed literature published primarily within the past five years was reviewed, with earlier studies included when considered fundamental to the development or clinical understanding of the field. Applications were organized according to their role in screening and diagnosis, deformity assessment, prediction and decision support, preoperative planning, intraoperative technologies, postoperative follow-up, and clinical workflow support. Emphasis was placed on the maturity of available evidence, validation strategies, and potential clinical utility.

Published studies describe a rapidly expanding range of applications, including automated radiographic measurements, radiation-free screening approaches, automated deformity classification, prediction of curve progression and postoperative outcomes, preoperative planning, patient-specific physics-based biomechanical simulations, remote monitoring technologies, and natural language processing-based clinical tools. The strongest evidence currently supports automated image analysis, selected predictive models, and specific planning technologies. However, many systems remain at the proof-of-concept stage and have been developed using retrospective single-centre datasets. Although technical performance is often promising, independent multicentre validation, prospective evaluation, assessment of clinical impact, and demonstration of integration within routine workflows remain limited for many applications.

AI and digital technologies have the potential to improve the consistency, efficiency, and personalization of spinal deformity care. Nevertheless, most current applications should be considered decision-support tools rather than autonomous systems. Future progress will depend not only on algorithmic performance but also on rigorous validation, transparent reporting, prospective clinical evaluation, and risk-informed assessment frameworks that align the level of evidence with the intended clinical role and potential impact of each technology.