Automated Radiographic Bone Age Assessment Using Deep Learning: Clinical Applications for Bio-Banding and Maturity Profiling in Youth Sports — A Narrative Review

This narrative review aims to critically examine deep learning approaches to automated radiographic bone age assessment and their potential relevance to bio-banding and maturity profiling in youth sports. The review will discuss the methodological characteristics, reported accuracy, validation, and limitations of automated bone age assessment systems, including comparisons with conventional radiographic assessment methods. Particular attention will be given to the populations used for model development and validation, generalizability to young athletes, and the distinction between estimating skeletal age and informing sport-specific decisions. The synthesis will address potential clinical and practical applications, feasibility, radiation exposure considerations, and ethical issues associated with assessment in minors. Evidence obtained in general pediatric populations will be distinguished from evidence directly involving youth athletes. Applications to bio-banding will be considered critically, without assuming that accurate bone age estimation alone establishes the effectiveness or appropriateness of maturity-based grouping. The review will provide a qualitative, thematic synthesis of the literature. Expected outputs include an overview of available approaches, an appraisal of the evidence supporting their use in youth sports, and identification of research gaps and priorities for future validation. This study is a narrative review, rather than a systematic review or meta-analysis. The registration documents its objectives and methodological approach; the accompanying protocol will specify the literature search, selection approach, and stage of the review at registration.