LASSO Logistic Regression was Used to Analyze the Risk Factors for Cauda Equina Injury Secondary to Lumbar Spinal Stenosis and to Build a Risk Model.

Abstract

Objective: To analyze the risk factors for secondary cauda equina injury in lumbar spinal stenosis using LASSO logistic regression and to construct a risk prediction model in the form of a nomogram.

Methods: Patients with lumbar spinal stenosis were divided into a secondary injury group (90 cases) and a non-secondary injury group (110 cases). LASSO logistic regression was applied, and a risk nomogram was generated. The predictive efficacy of the model was evaluated using receiver operating characteristic (ROC) curves and calibration curves.

Results: The ROC curve analysis showed that the area under the curve (AUC) of the risk nomogram model was 0.865 (95% CI: 0.755-0.948), with a sensitivity of 91.11% (82/90), specificity of 93.64% (103/110), and accuracy of 92.50% (185/200). The risk nomogram model demonstrated good fit (χ² = 3.347, df = 7, P = 0.341), and the C-index of Bootstrap internal validation was 0.823.

Conclusion: Age > 60 years, disease duration > 1 year, multiple stenosis segments, small median sagittal diameter, small cross-sectional area of the spinal canal, and shorter segment length are risk factors for secondary cauda equina injury in patients with lumbar spinal stenosis. The risk prediction model based on this nomogram has good clinical application value.