Abstract
The clinical efficacy of 3D ultrasound for spinal curvature assessment was evaluated in adolescent idiopathic scoliosis (AIS) patients through comparative analysis with conventional radiography. A prospective cohort of 230 preoperative and 22 postoperative AIS patients was evaluated over a 24-month period. Spinal curvature assessments were performed using a standardized 3D ultrasound system (SCN801), with reliability quantified via intraclass correlation coefficients (ICC) and validity assessed through Spearman correlation analysis. Ultrasound measurements demonstrated good reliability (ICC > 0.70), particularly for transverse process (TP)-based angle measurements, which showed superior consistency (ICC > 0.80). Strong correlations were observed between radiographic Cobb angles and TP-derived ultrasound curve angles (UCA) in mild-to-moderate curves (R2 > 0.70), though correlations diminished in severe deformities (R2 > 0.6). Mean scanning time measured 37.3 ± 6.8 s, with diagnostic-quality images obtained in 81 % of initial acquisitions. Multivariate regression analysis identified thoracic kyphosis magnitude (B = 3.334, p = 0.012) and vertebral rotation severity (B = 1.862-2.209, p < 0.05) as significant predictors of measurement variability in thoracic and (thoraco) lumbar curves. The findings demonstrate that TP-based 3D ultrasound enables accurate radiation-free spinal assessments in mild-to-moderate AIS, with measurement accuracy significantly influenced by vertebral rotation and thoracic kyphosis severity. Despite reduced performance in complex deformities, the method demonstrates operational efficiency, diagnostic precision, and an enhanced safety profile, supporting its clinical utility as a screening alternative to conventional radiography.