Differentiating early and advanced Brucella spondylitis using an MRI-based radiomics nomogram model

Abstract

Background

Accurate differentiation between early and advanced Brucella spondylitis is crucial for effective treatment.

Purpose

To develop a magnetic resonance imaging (MRI)-based radiomics nomogram model for distinguishing between early and advanced stages of Brucella spondylitis.

Material and Methods

We conducted a retrospective analysis of clinical and imaging data from 100 patients with early Brucella spondylitis and 100 patients with advanced Brucella spondylitis. Regions of interest were marked on sagittal T2-weighted fat-suppressed lumbar MRI scans. Radiomic features were extracted and used to build a radiomics model. The significance of these features was evaluated using the Shapley Additive Explanations (SHAP) method. Intravoxel incoherent motion (IVIM) quantitative parameters were also included as clinical features, with key parameters selected to create a clinical model. A nomogram model was developed by combining clinical and radiomic features. The performance of the three models was compared and validated using receiver operating characteristic curves, calibration curves, and decision curves.

Results

Eight radiomic features were selected. The clinical feature's D-value showed significant differences between the training and test sets. The nomogram model integrating both clinical and radiomic features achieved an AUC of 0.998 in the training set and 0.992 in the test set, surpassing the performance of both the clinical and radiomic models alone. Calibration and decision curves confirmed the model's strong predictive performance.

Conclusion

This study shows that the MRI-based radiomics nomogram model effectively differentiates between early and advanced Brucella spondylitis, offering clinicians a valuable tool for personalized treatment across different disease stages.