A predictive model for postoperative shoulder imbalance in Lenke 1 A/B and 2 A/B adolescent idiopathic scoliosis.

Abstract

This study aimed to investigate the risk factors associated with postoperative shoulder imbalance (PSI) and to construct predictive models for Lenke 1 A/B and 2 A/B adolescent idiopathic scoliosis (AIS) patients. Preoperative and first erect (FE) radiographic parameters were measured in 147 Lenke 1 A/B and 2 A/B AIS patients. Univariate analysis and multivariate logistic regression analysis were employed to compare clinical data between the PSI and non-PSI groups, leading to the establishment of a corresponding predictive model. At the last follow-up, 32 (21.8%) patients developed PSI. Six risk factors were identified: Risser grade, preoperative radiographic shoulder height (RSH), FE correction of the main thoracic curve (MTC), FE Cobb angle ratio of the proximal thoracic curve (PTC) to MTC (PTC/MTC), FE correction of the lumbar curve (LC), and FE RSH. A nomogram was used to construct a clinical prediction model for PSI. The Hosmer-Lemeshow test indicated a good fit between the nomogram-predicted probability and actual values. Additionally, the area under the receiver operating characteristic curve (AUC) was 0.833. Decision curve analysis (DCA) showed that the nomogram provided substantial net benefit within the threshold probability range. Risser grade, preoperative RSH, FE correction of MTC, FE PTC/MTC ratio, FE correction of LC, and FE RSH are risk factors for PSI in Lenke 1 A/B and 2 A/B AIS patients. The predictive model based on these six factors demonstrates excellent predictive accuracy. The nomogram derived from this model is practical and straightforward, effectively predicting the occurrence of PSI in these patients during follow-up and providing valuable guidance for surgical planning. However, these findings require further verification in larger-scale multicenter studies to confirm their universal applicability.