Related Experiment Video
Updated: Nov 12, 2025
15:48
Tracking the Mammary Architectural Features and Detecting Breast Cancer with Magnetic Resonance Diffusion Tensor Imaging
Published on: December 15, 2014
22.8K
Understanding diffusion-weighted MRI analysis: Repeatability and performance of diffusion models in a benign breast
Neil Peter Jerome1,2, Igor Vidić3, Liv Egnell2,3
1Department of Circulation and Medical Imaging, Norwegian University of Science and Technology (NTNU), Trondheim, Norway.
NMR in Biomedicine
|March 19, 2021
Summary
Diffusion-weighted MRI (DWI) analysis choices impact breast lesion classification. Stretched exponential models offer more repeatable diffusivity measures than biexponential models, suggesting improved biomarker potential.
Area of Science:
- Radiology
- Oncology
- Biomedical Imaging
Background:
- Diffusion-weighted MRI (DWI) is vital for oncology research and clinical breast lesion assessment.
- Analysis choices in DWI significantly influence the reliability of derived parameters.
- Understanding parameter repeatability is crucial for accurate lesion monitoring.
Purpose of the Study:
- To critically evaluate the repeatability and curve-fitting performance of common DWI signal representations.
- To assess these models in a prospective cohort of patients with benign breast lesions.
- To determine the impact of analysis choices on biomarker detectability.
Main Methods:
- Twenty patients with benign breast lesions underwent repeated 3 T DWI scans.
- Voxel-wise coefficients of variation (CoVs) were calculated for monoexponential, biexponential, and stretched exponential models.
- Goodness-of-fit was assessed using the Bayesian information criterion.
Main Results:
- Diffusivity measures (ADC, D, DDC) showed good repeatability (CoV < 10%).
- Pseudodiffusion parameters (f, D*) exhibited poor repeatability (CoV > 50%).
- The stretched exponential model demonstrated better repeatability than the biexponential model.
Conclusions:
- No single DWI signal representation consistently provided the best fit across all lesions or visits.
- Parameter robustness is critical for reliable clinical application of DWI biomarkers.
- Stretched exponential modeling shows promise for repeatable DWI biomarker analysis in breast lesions.
