Quantifying heterogeneity in dynamic contrast-enhanced MRI parameter maps

Area of Science:

• Medical Imaging
• Radiology
• Quantitative MRI

Background:

• Standard Dynamic Contrast-Enhanced MRI (DCE-MRI) statistics often overlook spatial parameter distribution.
• Spatial arrangement in DCE-MRI parameter maps holds significant diagnostic and prognostic value.
• Novel statistical approaches are needed to incorporate spatial information.

Purpose of the Study:

• To introduce novel 4-D spatial statistics for DCE-MRI parameter maps.
• To assess the sensitivity of these new statistics to both parameter values and their spatial arrangement.
• To evaluate the utility of these statistics in differentiating tumor grades and treatment responses.

Main Methods:

• Creation of 4-D binary objects from 3-D DCE-MRI parameter maps by extruding voxels proportionally to their values.
• Computation of surface area, volume, surface area to volume ratio, and box counting (fractal) dimension on these 4-D objects.
• Application of these statistics to analyze glioma grades and liver metastases before/after VEGF inhibitor treatment.

Main Results:

• Significant differences were observed between low and high-grade gliomas using box counting dimension, surface area, and volume of extruded maps (p < 0.05).
• Significant differences were found in liver metastases treated with a VEGF inhibitor, particularly for the surface area to volume ratio of extruded K(trans) and v(e) maps (p = 0.0013 and p = 0.045).

Conclusions:

• Novel 4-D spatial statistics provide valuable insights into DCE-MRI data beyond conventional methods.
• These statistics demonstrate potential for improved tumor grading and assessment of treatment efficacy.
• The incorporation of spatial information significantly enhances the diagnostic power of DCE-MRI parameters.