[Establishment and validation of a risk prediction model for scoliosis after Nuss procedure in children and young adults with pectus excavatum].

OBJECTIVE

To identify the risk factors associated with the development of post-Nuss procedure scoliosis in pectus excavatum patients and to establish a relevant predictive model.

METHODS

A retro-spective review was conducted on pectus excavatum patients who underwent Nuss procedure in Department of Thoracic Surgery at Beijing Jishuitan Hospital between January 2018 and February 2023. We gathered the patient demographic information (including age, sex, height, weight, and body mass index) and diagnostic imaging results (echocardiogram, chest CT scan, full-spine radiography, and PA/lateral chest X-rays), and measurements of Haller index, asymmetry index, sternal torsion angle (STA) index, and Cobb angle changes. The risk of post-Nuss procedure scoliosis was assessed. Cox regression analysis was performed to identify independent risk factors for scoliosis development in the pectus excavatum patients. Based on the results of the Cox regression analysis, a risk prediction model was established, and its specific predictive performance was assessed through internal cross-validation.

RESULTS

A total of 59 pectus excavatum patients who underwent Nuss procedure were included after applying inclusion and exclusion criteria. The median follow-up duration was 6.84 months, and the follow-up results showed significant improvements in Haller index, STA index, and asymmetry index postoperatively. Twelve patients developed scoliosis 3 months after Nuss procedure, while 47 patients did not, the incidence of scoliosis was 20.3%. Cox regression analysis identified preoperative pulmonary artery hypertension and preoperative asymmetry index as independent risk factors for post-Nuss procedure scoliosis. A predictive model was constructed based on single-factor Cox regression analysis results, incorporating age, height, weight, body mass index, preoperative pulmonary artery hypertension, preoperative Haller index, STA index, asymmetry index, and Cobb angle. The receiver operating characteristic (ROC) curve was plotted, and the area under the curve (AUC) for the overall predictive model was calculated to be 0.995. A calibration curve demonstrated good alignment between predicted values and actual values.

CONCLUSION

Nuss procedure achieved favorable clinical outcomes. However, postoperative scoliosis emerged as a significant complication with a high incidence rate. Pulmonary artery hypertension and asymmetry index were independent predictors of post-Nuss procedure scoliosis. The predictive model developed in this study demonstrated robust performance in estimating the risk of postoperative scoliosis.