Modified Lenke Classification Stratifies Brace Initiation Risk and Longitudinal Cobb Angles Trajectories in Female Patients With Adolescent Idiopathic Scoliosis

Study Design. Retrospective single-center cohort study. Objective. To determine whether baseline modified Lenke classification (MLC), assessed from standing whole-spine radiographs, stratifies brace initiation risk and longitudinal Cobb-angle trajectories in female patients with adolescent idiopathic scoliosis (AIS) before reaching the conventional bracing threshold. Summary of Background Data. AIS progression is heterogeneous. Baseline Cobb angle and skeletal maturity are established risk factors, but whether simple baseline curve-pattern classification adds risk-stratification information during nonoperative follow-up remains unclear. Methods. We included female AIS patients aged 10 to <18 years with a baseline maximum Cobb angle ≥10° and <25°. Patients with prior brace treatment, brace initiation at the first visit, insufficient pre-brace follow-up, or missing baseline variables were excluded. The primary outcome was brace initiation. Brace-free survival was assessed using Kaplan–Meier analysis and multivariable Cox models adjusted for baseline maximum Cobb angle, age, body mass index, menarche status, and Risser status. Model fit was compared between a conventional baseline model and a model additionally including MLC. Two-year risks were estimated at fixed baseline Cobb angles. Longitudinal trajectories were analyzed using a linear mixed-effects model with natural cubic splines, an age-by-MLC interaction, and patient-specific random intercepts and slopes for age. Results. A total of 1087 patients were included; 210 (19.3%) underwent brace initiation. Brace-free survival differed across MLC groups. In multivariable Cox analysis, MLC 2, 3, 4, and 6 showed higher hazards of brace initiation than MLC 1, with the highest risk in MLC 4. Adding MLC improved likelihood-based model fit (χ²=19.35, df=5, P =0.002) and reduced AICc, although BIC increased. At baseline Cobb 20°, adjusted 2-year risk ranged from 23.3% in MLC 1 to 55.9% in MLC 4. Longitudinal Cobb-angle trajectories differed according to MLC. Conclusion. Baseline MLC may provide complementary information for stratifying brace-initiation risk and longitudinal Cobb-angle trajectories before the conventional bracing threshold.