300 Section - Artificial Intelligence and personalized rehabilitation pathways in Adolescent Idiopathic Scoliosis

Adolescent idiopathic scoliosis (AIS) is a three-dimensional spinal deformity affecting a substantial proportion of the adolescent population. Its highly variable natural history and unpredictable progression complicate early prognostic stratification and therapeutic decision-making, highlighting the need for more accurate and reliable tools to support diagnosis, risk prediction, and treatment optimization.
In recent years, artificial intelligence (AI) and machine learning (ML) have emerged as innovative approaches with the potential to enhance multiple aspects of AIS management, from image analysis to clinical decision support. 
The aim of this review was to provide a timely synthesis of available research on AI applications in AIS and to assess the methodological quality of published systematic reviews.












Design: Systematic reviews and meta-analyses published in English up to April 2025 were identified through structured searches of Embase, PubMed/MEDLINE, and the Cochrane Database. Studies addressing scoliosis and any AI-based intervention were considered eligible. Two independent reviewers screened titles and abstracts according to PRISMA rapid review guidelines, evaluated full-text articles, and assessed methodological quality using the AMSTAR 2 tool.

Results: Five systematic reviews met the inclusion criteria. Among the studies included in these reviews, 55% employed AI algorithms based on convolutional neural networks, artificial neural networks, support vector machines, decision trees, or hybrid models. The most consistently validated applications in AIS were automated Cobb angle measurement, with reported accuracy within 3° in selected models, curve classification, and prediction of curve progression. Emerging applications also included patient education and clinical decision support systems based on language models.

Conclusion:

Overall, AI and ML represent a promising opportunity in AIS management, enabling earlier identification of patients at higher risk of progression and facilitating more personalized treatment strategies. However, broader external validation and stronger clinical integration are essential before these technologies can be effectively implemented in routine practice.