Vertebral rotation as a predictor of residual deformity following scoliosis correction in spinal muscular atrophy: a retrospective analysis

This study investigated the impact of preoperative vertebral rotation (VR) on the surgical outcomes in spinal muscular atrophy (SMA) scoliosis. A retrospective analysis of 27 SMA patients (mean age 13.4 ± 4.3 years) who underwent scoliosis surgery between 2015 and 2019 was conducted. Preoperative VR was measured using Aaro-Dahlborn's and Ho's methods on computed tomography images. Surgical outcome prediction was evaluated using the Pearson correlation coefficient, linear stepwise regression, receiver operating characteristic (ROC) curve, and logistic regression analyses. Ho's method yielded significantly higher VR measurements than Aaro-Dahlborn's ( P < 0.001). Postoperative Cobb angles correlated positively with preoperative Cobb angles ( r ² = 0.425, P = 0.0002), Ho's method VR ( r ² = 0.449, P = 0.0001), and Aaro-Dahlborn's method VR ( r ² = 0.4352, P = 0.0002). Stepwise regression identified preoperative Ho's method VR and Cobb angles as independent predictors of postoperative Cobb angles. Postoperative Cobb angles >30 ° indicated increased risk of deformity progression in SMA. ROC curve analysis showed preoperative Ho's method VR significantly predicted postoperative Cobb angles >30 ° (area under the curve: 0.813, P = 0.006), with an optimal cutoff of 35 °. Logistic regression analysis revealed patients with preoperative Ho's method VR > 35 ° had a higher risk of postoperative Cobb angles >30 ° (odds ratio: 10.36, 95% confidence interval: 1.050-102.261, P = 0.045). This study demonstrated that Ho's method better predicted surgical outcomes, with preoperative Ho's method VR > 35 ° at the apex associated with higher initial residual scoliosis curves after surgery. These findings could enhance surgical planning and improve outcome predictions in SMA scoliosis correction.