Diffusion-weighted MR image analysis based on gamma distribution model for differentiating benign and malignant brain...

Area of Science:

• Neuroimaging
• Medical Physics
• Oncology

Background:

• Biopsy is invasive for brain tumor diagnosis.
• Magnetic resonance imaging (MRI) offers a less invasive alternative.
• Evaluating advanced imaging models is crucial for non-invasive tumor assessment.

Purpose of the Study:

• To assess the utility of a gamma distribution model using MRI for staging and grading brain tumors.
• To compare the diagnostic performance of the gamma model parameters against traditional diffusion metrics.
• To explore the potential of this model in differentiating benign from malignant brain tumors.

Main Methods:

• Diffusion-weighted imaging (DWI) with multiple b-values (0-2000 s/mm²) was performed on 66 patients (24 benign, 42 malignant brain tumors).
• The gamma distribution model was applied to calculate apparent diffusion coefficient (ADC), shape parameter (κ), and scale parameter (θ).
• Fractions representing intracellular, extracellular diffusion, and perfusion (ƒ1, ƒ2, ƒ3) were analyzed for staging, and specific restricted diffusion fractions (ƒ11, ƒ12, ƒ13) for grading.

Main Results:

• Gamma model parameters (κ, ƒ1, ƒ2, ƒ3) effectively differentiated meningioma from glioma.
• At b=2000 s/mm², fraction ƒ3 demonstrated the highest diagnostic performance (AUC=0.891) for differentiating tumor types.
• Fraction ƒ12 showed the best performance (AUC=0.870) for grading gliomas (high-grade vs. low-grade) at b=2000 s/mm².

Conclusions:

• The gamma distribution model combined with multi-b-value DWI shows significant potential for non-invasively staging and grading brain tumors.
• This approach could enhance diagnostic accuracy in neurooncology.
• Integration into clinical practice may improve patient outcomes through better treatment planning.