Abstract: Objective To investigate craniofacial and upper airway anatomical characteristics in children using cone beam computed tomography(CBCT)and their association with apnea hypopnea index(AHI), aiming to identify key predictive indicators for assessing the severity of obstructive sleep apnea(OSA). Methods A total of 80 pediatric patients who visited the Department of Orthodontics at Ningbo Stomatological Hospital between June 2022 and June 2024 with chief complaints of malocclusion or mouth breathing were consecutively enrolled in the study. The sample included 53 boys and 27 girls with a mean age of 9.36 years. Portable home polysomnography was performed to obtain AHI values, followed by CBCT scans to measure adenoid-to-nasopharynx volume ratio(AV/NPV), maxillary and mandibular width and position, hyoid bone position, and upper airway dimensions. Hierarchical regression analysis was conducted to evaluate the independent predictive value of craniofacial and upper airway variables for AHI after controlling for sex, body mass index(BMI), and adenoid size. Results Spearman correlation analysis showed a significant positive correlation between AV/NPV and AHI(P<0.05). In contrast, mandibular ramus length(Ar-Go), maxillary width(Mx-Mx),mandibular width(Go-Go), minimum axial airway area(MAA), retronasal airway volume(V-RNA), retropalatal airway volume(V-RPA)and retrolingual airway volume(V-RLA)were negatively correlated with AHI(P<0.05). In the hierarchical regression analysis, sex and BMI were entered as control variables in Step 1, accounting for only 0.2% of the variance in AHI(R²=0.002), with the model being non-significant(F_(2,77)=0.067, P=0.935). In Step 2, the addition of AV/NPV significantly improved the model, with an increase in explained variance(ΔR²=0.179). AV/NPV was a significant positive predictor of AHI(β=0.441, P<0.001). In Step 3, CBCT-derived variables significantly associated with AHI were added to the model. This step further increased the explained variance(ΔR2=0.273). Among these, Mx-Mx(β=-0.207, P=0.041),Go-Go(β=-0.263, P=0.018)and MAA(β=-0.346, P<0.001)were identified as independent predictors of AHI. Conclusion Upper airway constriction, maxillary and mandibular deficiency serve as independent risk factors for pediatric OSA. These three-dimensional anatomical variables may provide a theoretical basis for early clinical identification of high-risk children.

Key words: Cone beam computed tomography, Obstructive sleep apnea, Apnea hypopnea index, Adenoid, Hierarchical regression analysis