Dynamic contrast enhanced-magnetic resonance fingerprinting (DCE-MRF): A new quantitative MRI method to reliably assess tumor vascular perfusion

Abstract

Purpose

The clinical utility of conventional DCE-MRI methods is limited by the use of conventional qualitative dynamic T₁-weighted images, resulting in poor reproducibility. This study presents the initial implementation of a new DCE-magnetic resonance fingerprinting (DCE-MRF) methodology to provide reproducible, quantitative assessments of tumor vascular perfusion.

Methods

The DCE-MRF acquisition combines multiple T₁ preparations, highly undersampled spiral trajectories (R = 48), a low-rank reconstruction method, and low tip angles on a 9.4 T preclinical MRI scanner to rapidly generate dynamic T₁ maps (23-s temporal resolution). In vitro validation experiments were conducted across a range of Gadovist concentrations to assess accuracy and temporal precision in comparison to conventional methods. The DCE-MRF method was also evaluated in vivo in an orthotopic 4T1 mouse model of breast cancer (n = 25). Pharmacokinetic modeling of the in vivo data was performed using a linear reference region model (LRRM).

Results

In vitro DCE-MRF studies demonstrated good agreement with conventional MRI methods for T₁ measurements (R² $\ge$ 0.99). The iterative low-rank reconstruction method also reduced artifacts compared to conventional reconstruction methods. DCE-MRF demonstrated a 2- to 3-fold reduction in temporal variability compared to conventional DCE-MRI, and enabled effective in vivo pharmacokinetic modeling using the LRRM by generating voxelwise maps of RK^trans and k_ep,T as measures of tumor vascular perfusion.

Conclusions

DCE-MRF represents a new inherently quantitative approach to measuring tumor vascular perfusion that can be used in animal models and eventually in patients.