A multicompartment model for intratumor tissue-specific analysis of DCE-MRI using non-negative matrix factorization

Area of Science:

• Medical Imaging
• Biophysics
• Pharmacokinetics

Background:

• Dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) is crucial for pharmacokinetic analysis.
• The partial volume effect (PVE) significantly impacts the accuracy and stability of DCE-MRI data.
• Existing methods struggle to overcome PVE, limiting precise kinetic parameter mapping.

Purpose of the Study:

• To develop a novel multicompartment model to address the PVE in DCE-MRI.
• To enable tissue-specific pharmacokinetic analysis with improved accuracy and stability.
• To generate better kinetic parameter maps for clinical applications.

Main Methods:

• Introduced fractional volumes of tissue compartments as an independent parameter per pixel.
• Developed a linear separable multicompartment model for simultaneous estimation of time-concentration curves and fractional volumes.
• Utilized minimum-volume constraint non-negative matrix factorization (MVC-NMF) to solve the convex optimization problem.

Main Results:

• The proposed model demonstrated lower bias and root mean square fitting error on synthetic data compared to state-of-the-art algorithms.
• The model showed improved robustness to varying noise levels.
• Analysis of real DCE-MRI breast cancer data indicated enhanced pharmacokinetic parameter estimation for distinguishing chemotherapy treatment response.

Conclusions:

• The new model significantly improves the accuracy and stability of tissue-specific fractional volume and kinetic parameter estimation in DCE-MRI.
• Enhanced robustness to noise leads to more precise kinetic analysis.
• This facilitates more accurate prognosis and therapeutic response evaluation using DCE-MRI data.