Diagnostic factors influencing orthodontic extraction decision-making: a systematic review

Introduction Orthodontic extraction decision-making requires integration of multiple diagnostic characteristics, yet the factors underlying extraction vs. non-extraction treatment selection have not been comprehensively synthesized across conventional clinical research and contemporary prediction models. This systematic review aimed to identify and critically evaluate the diagnostic factors influencing orthodontic extraction decisions and determine their consistency across the available evidence. Methods PubMed, MEDLINE, Embase, Scopus, Web of Science, and the Cochrane Library were searched from inception through July 2026 for studies evaluating diagnostic factors associated with orthodontic extraction decisions or statistical or artificial intelligence-based prediction models. Two reviewers independently performed study selection, data extraction, and risk-of-bias assessment. Substantial clinical and methodological heterogeneity precluded meta-analysis, and findings were synthesized narratively. Results Nineteen studies met the inclusion criteria. Dental crowding and tooth size -arch length discrepancy were the most consistently evaluated diagnostic factors but were typically considered alongside facial profile, incisor position, skeletal relationships, and occlusal characteristics. No consistent numerical threshold for recommending extraction was identified. Artificial intelligence-based prediction models demonstrated promising performance, with interpretable predictors substantially overlapping with diagnostic variables identified in conventional clinical studies. Methodological quality ranged from moderate to critical risk of bias, with most prediction-model studies judged to be at high risk of bias. Discussion These findings indicate that orthodontic extraction decision-making is fundamentally multifactorial rather than governed by a single diagnostic measurement or threshold. Current artificial intelligence models should be considered decision-support tools pending higher-quality prospective research and external validation; whether image-derived or latent features provide genuinely novel determinants of extraction need remains uncertain. Systematic Review Registration https://osf.io/x86zy