Predictive value of the ratio of fusion segments to main curve segments for postoperative curve progression in congenital scoliosis with solitary hemivertebra.

Purpose: To investigate the influence of the preoperative compensatory curve on the postoperative curve progression in congenital scoliosis (CS) patients following hemivertebra (HV) resection and short fusion and to analyze the risk factors for postoperative curve progress.

Methods: This study retrospectively reviewed a consecutive cohort of CS patients who underwent HV resection and short fusion with a minimum follow-up of 2 years. General demographic information, as well as preoperative, postoperative, and latest follow-up radiographic data, were collected. The fusion segment (FS) referred to the number of vertebrae within the anchoring range of the most cranial and caudal pedicle screws. The main curve segment (MCS) was defined as the number of vertebrae included in the main curve. Patients were divided into two groups based on the presence or absence of compensatory curves preoperatively: the compensatory curve group (Group C) and the non-compensatory curve group (Group NC). Furthermore, patients were categorized into progression (Group P) and non-progression (Group NP) groups based on the evolution of the coronal curve after surgery. Logistic regression analysis was employed to identify risk factors for curve progression, and receiver operating characteristic (ROC) curves were constructed to evaluate the predictive value of these risk factors.

Results: A total of 205 patients were included in this study, with 58 classified into Group C and 147 into Group NC. The incidence of postoperative progression of the coronal curve was notably higher in Group C compared to Group NC, with incidences of 29.3% and 12.9%, respectively (p = 0.005). Multivariate logistic regression analysis revealed that a lower FS/MCS ratio, a larger postoperative tilt of the UIV, and a larger postoperative tilt of the LIV were all independent risk factors contributing to the progression of the curve following surgery. Receiver operator characteristic (ROC) curves showed superior predictive value of FS/MCS for postoperative curve progression (AUC: 0.793; 95%CI: 0.675 ~ 0.912; p < 0.001). The optimal cut-off value for FS/MCS was 0.915 using the Youden index.

Conclusion: The presence of the compensatory curve appeared to increase the likelihood of postoperative curve progression in CS patients undergoing posterior HV resection and short fusion. Key risk factors for this progression include a lower FS/MCS ratio and more pronounced tilting of the postoperative UIV and LIV. We recommended proper alignment of FS with MCS, as well as horizontalization of both the UIV and LIV.