The study developed a radiomic model based on machine learning from CT images for preoperative screening of lumbar spine osteoporosis, as undiagnosed osteoporosis increases the risk of complications. They enrolled 166 patients with DEXA, spine CT, and MRI, with data split into training and validation cohorts in an 8:2 ratio. They extracted 851 radiomic features from CT using 3D Slicer PyRadiomics, feature selection was done by mRMR and LASSO regression, classifiers were LR, SVM, XGBoost and RF. The XGBoost radiomic model was the best with an AUC of 0.89 in the training and 0.91 in the test cohort. DeLong's test confirmed statistically significant differences compared to VBQ and HU (p < 0.05). Analysis of the decision curve showed a higher net benefit of the radiomic model. The model used nine CT radiomictic features and significantly outperforms VBQ and HU in diagnostic accuracy.