MRI-based Texture Analysis in Differentiation of Benign and Malignant Vertebral Compression Fractures

Area of Science:

• Radiology
• Medical Imaging Analysis
• Oncology

Background:

• Vertebral compression fractures require accurate diagnosis for effective clinical management.
• Conventional imaging may not always provide typical or definitive findings.
• Texture analysis offers a novel approach to extract non-visual information from diagnostic images.

Purpose of the Study:

• To evaluate magnetic resonance imaging (MRI) in vertebral compression fractures using texture analysis.
• To compare texture analysis findings with histopathological data.
• To assess the efficacy of texture analysis in distinguishing benign from malignant fractures.

Main Methods:

• Fifty-five patients with 56 vertebral compression fractures were included.
• MRI data were analyzed using Local Image Feature Extraction (LIFEx) software for texture analysis.
• Results were correlated with histopathological diagnoses.

Main Results:

• The Decision Tree algorithm achieved approximately 95% accuracy, precision, and recall.
• Texture analysis demonstrated significant potential in fracture characterization.

Conclusions:

• Integrating texture analysis parameters with conventional MRI can improve diagnostic accuracy for radiologists.
• This approach may help avoid unnecessary invasive procedures.
• Further large-scale prospective studies are needed to validate texture analysis in vertebral compression fractures.