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

Area of Science:

• Medical Imaging
• Radiomics
• Infectious Diseases

Background:

• Differentiating early and advanced Brucella spondylitis is critical for effective patient management.
• Current diagnostic methods may lack the precision needed for timely intervention.
• Brucellosis is a zoonotic disease that can affect the spine, leading to spondylitis.

Purpose of the Study:

• To develop and validate a magnetic resonance imaging (MRI)-based radiomics nomogram model.
• The model aims to distinguish between early and advanced stages of Brucella spondylitis.
• To provide a tool for personalized treatment based on disease stage.

Main Methods:

• Retrospective analysis of 200 patients (100 early, 100 advanced Brucella spondylitis).
• Extraction of radiomic features from lumbar MRI scans (T2-weighted, fat-suppressed).
• Development of combined clinical (Intravoxel Incoherent Motion - IVIM parameters) and radiomic nomogram model, validated with ROC, calibration, and decision curves.

Main Results:

• Eight significant radiomic features were identified.
• The integrated nomogram model achieved high diagnostic performance: AUC 0.998 (training) and 0.992 (test).
• The nomogram model outperformed standalone clinical and radiomic models.

Conclusions:

• An MRI-based radiomics nomogram model demonstrates high efficacy in differentiating early and advanced Brucella spondylitis.
• This model serves as a valuable tool for clinicians, aiding in tailored treatment strategies.
• The integration of radiomic and quantitative IVIM parameters enhances diagnostic accuracy.