Abstract: BACKGROUND: Current research on axial symptoms following posterior cervical surgery predominantly focuses on pathological mechanisms, while a clinically applicable risk prediction model remains lacking. 
OBJECTIVE: To investigate the factors influencing the occurrence of axial symptoms following laminectomy with lateral mass screw fixation for cervical compressive myelopathy and to develop and validate a predictive model for postoperative axial symptoms, and design a dynamic web calculator to achieve real-time evaluation and visualization of the risk of postoperative complications.
METHODS: A retrospective analysis was conducted on medical records of 127 patients who underwent laminectomy with lateral mass screw fixation for cervical compressive myelopathy at Affiliated Hospital of Xuzhou Medical University between February 2021 and April 2024. Patients were categorized into an axial symptom group and a non-axial symptom group based on the presence of axial symptoms within six months postoperatively. To handle missing data, the K-Nearest Neighbors Algorithm was employed for multiple imputation, and Synthetic Minority Over-sampling Technique was used to address class imbalance. Combining the results of univariate and multivariate binary Logistic regression analysis, the influencing factors that jointly affected the incidence of postoperative axial symptoms were screened out, that is, the predictive factors included in the prediction model. The dataset was stratified into a training set and a testing set in a 7:3 ratio. A nomogram prediction system was constructed based on the multivariate regression results, and its performance was evaluated using the receiver operating characteristic curve, calibration curve, and decision curve analysis. The reliability of the model was verified in three dimensions. A web-based calculator was developed to enable real-time individualized risk assessment and facilitate clinical decision-making through visualized results.
RESULTS AND CONCLUSION: (1) A total of 52 patients (40.9%) developed axial symptoms postoperatively. (2) Univariate analysis indicated significant differences between the two groups in age, operation time, intraoperative blood loss, early postoperative functional exercise, cervical collar wearing duration, preoperative Japanese Orthopaedic Association score, best postoperative Japanese Orthopaedic Association score, C2-C7 Cobb angle in neutral and flexion positions, and cervical lordosis loss (P < 0.1). (3) Multivariate logistic regression analysis using the forward stepwise selection method identified age, early postoperative functional exercise, cervical collar wearing duration, best postoperative Japanese Orthopaedic Association score, and cervical lordosis loss as independent predictors of axial symptoms. Among them, age, cervical collar wearing duration, and cervical lordosis loss were risk factors, whereas early postoperative functional exercise and best postoperative Japanese Orthopaedic Association score were protective factors. (4) A nomogram prediction model and a web-based calculator were established based on the analysis (available at http://121.40.90.248:8080/). The areas under the receiver operating characteristic curve values for the training and testing sets were 0.877 (95%CI: 0.798-0.955) and 0.834 (95%CI: 0.705-0.964), respectively, indicating excellent predictive performance. The calibration curve demonstrated strong agreement between predicted and observed outcomes. (5) It is indicated that age, cervical collar wearing duration, cervical lordosis loss, early postoperative functional exercise, and best postoperative Japanese Orthopaedic Association score are the key factors influencing the occurrence of axial symptoms following laminectomy with lateral mass screw fixation. Based on these findings, we developed a highly accurate nomogram prediction model and implemented a web-based calculator. 

Key words: cervical compressive myelopathy, laminectomy with lateral mass screw fixation, axial symptom, risk factor, nomogram, prediction model