Multimodal Imaging-Based Interpretable Radiomics for Differentiating Brucella and Tuberculosis Spondylitis: A Two-Center Study

Background: Spondylitis, particularly infectious forms caused by Mycobacterium tuberculosis and Brucella species, presents significant clinical challenges due to overlapping symptoms and diagnostic difficulties. Accurate differentiation is crucial for effective treatment, necessitating advanced imaging techniques and radiomics to enhance diagnostic precision and improve patient outcomes in cases of TB and brucella spondylitis. Methods This retrospective cohort study included 165 patients diagnosed with Brucella spondylitis (BS) or tuberculous spondylitis (TS) from January 2020 to December 2024, with an external validation cohort of 35 patients. Inclusion criteria consisted of relevant clinical symptoms for at least six months, positive serological tests, MRI abnormalities, and complete medical records. Imaging was performed at two centers, employing standardized protocols for T1-weighted and T2-weighted MRI. Region of Interest (ROI) segmentation and radiomics feature extraction were conducted using the Deepwise platform, yielding a total of 1049 CT and 5829 MRI features. Nine predictive models were developed and validated through nested five-fold cross-validation, assessing performance metrics such as AUC, sensitivity, and specificity. Statistical significance was set at p < 0.05. Results A total of 195 patients with 207 lesions were analyzed, comprising 124 cases of tuberculous spondylitis (TS) and 83 cases of brucellosis spondylitis (BS). An external validation cohort included 57 patients with 60 lesions. Nine predictive models were developed using selected features: the CT model utilized 80 features, while T1-weighted (T1WI), T2-weighted (T2WI), and fat-suppressed T2WI (FS T2WI) models employed 14, 33, and 32 features, respectively. Multi-modality models combined features from various sequences, achieving optimal performance with an AUC of 0.8136. Model efficacy was validated through ROC curve analysis, decision curve analyses, and calibration plots. Additionally, SHAP analysis was used to interpret model predictions, identifying key influential features, which are detailed in the supplementary materials. Conclusion Our research indicates that multimodal imaging-based radiomics hold significant potential for the differential diagnosis of BS and TS