Abstract: BACKGROUND: To establish a model for predicting the progression of scoliosis at different time points after the first diagnosis of adolescent idiopathic scoliosis contributes to accurately identify and predict the risks of progression in the early stage of disease development.
OBJECTIVE: To explore the factors influencing the progression of scoliosis in patients with adolescent idiopathic scoliosis after first diagnosis, and to construct a nomogram risk prediction model.  
METHODS: The clinical data of patients with adolescent idiopathic scoliosis who were first diagnosed at the Second Hospital of Shanxi Medical University from January 2019 to June 2023 were retrospectively analyzed and followed up to the progress of lateral bending, that is, the progress of main bend Cobb angle ≥ 6°, the last visit or cutoff date (June 2023). The patients were randomly divided at the rate of 8:2 into training set and verification set, and the patients were divided into progressive group and non-progressive group according to whether the patients had lateral bending progress or not. Least Absolute Shrinkage and Selection Operator (LASSO)-Cox regression analyses were used to investigate the independent influences on the occurrence of scoliosis progression in adolescent idiopathic scoliosis and to construct a Cox regression algorithm-based risk prediction model for column-line diagrams based on the area under the curve (AUC) of the receiver operating characteristic curve of the subjects, calibration curve, and decision curve were used to validate and evaluate the differentiation, accuracy, and clinical application value of the model.  
RESULTS AND CONCLUSION: (1) A total of 294 patients with adolescent idiopathic scoliosis were included in this study, with a progression rate of 41.84%. (2) The results of LASSO-Cox regression analysis showed that age at first diagnosis ≥ 14.5 years, brace treatment, Risser’s sign ≥ 1 grade, parietal cone offset > 1.6 cm, and thoracic 1 cone tilt angle > 1.2° were protective factors for scoliosis progression, whereas age at first diagnosis Cobb’s angle > 16.5°, parietal cone rotation ≥ 2 degrees, and spinal growth rate > 4.5 cm/year were risk factors for scoliosis progression. (3) The nomogram model constructed according to the above factors had excellent predictive ability and clinical significance. The AUC values of 6, 12, 18 and 24 months in the training set and verification set were 0.731, 0.852, 0.855, 0.843 and 0.766, 0.850, 0.850 and 0.830, respectively, and the C-index of the model was 0.795 and 0.771 respectively, indicating that the model has good discrimination. The calibration curve showed that the actual observation results fitted well with the predicted results, and the accuracy of the model was high. In addition, the decision curve showed that the use of the model to make clinical decisions could bring net benefits to patients. Therefore, the adolescent idiopathic scoliosis progression risk prediction nomogram model constructed in this study can use simple variables to judge the risk probability of patients' progression at different time points in the future, and then guide clinicians to choose a more reasonable treatment.

Key words: adolescent idiopathic scoliosis, scoliosis progress, influencing factors, LASSO, Cox regression, prediction model, nomogram