Focus Session
SCMR 22nd Annual Scientific Sessions
Xenios Milidonis, PhD, MSc
Research Associate
King's College London
Ludovica Beraldi, BEng
Student
King's College London
Tobias Schaeffter, PhD
Head of Division Medical Physics and Metrological Information Technology
Physikalisch-Technische Bundesanstalt (PTB), Braunschweig and Berlin, Germany
Amedeo Chiribiri, MD, PhD
Reader in Cardiovascular Imaging; Consultant Cardiologist
King's College London
Background:
Quantitative assessment of dynamic contrast-enhanced (DCE) MRI for myocardial perfusion using tracer kinetic modelling is gaining popularity over more traditional methods such as visual assessment and semi-quantitative measures. Validation and translation of such techniques is hampered by the lack of clinically relevant standards. We developed a novel synthetic multi-compartmental phantom allowing controlled and reproducible extravasation of contrast agents for the validation and translation of new CMR perfusion techniques into the clinic.
Methods:
A two-compartmental myocardium was designed, allowing diffusion of contrast agent from a set of capillaries (vascular compartment) into an outer shell (interstitial compartment) through microscopic pores (Figure 1). The myocardium is compatible with an existing hardware phantom allowing independent control of the outflow from each compartment via means of calibrated peristaltic pumps and ultrasonic flow meters [1]. The myocardium was scanned with a 1.5T Philips Ingenia system and an ECG-gated turbo field echo sequence (TR/TE = 2.61/1.25 ms, 20° flip angle, 1 average, 128×128 matrix, 1.5×1.5×12 mm3 resolution) after injection of contrast boluses. The system’s flow behavior was assessed using phase contrast (PC) MRI. Scans were repeated twice for three different ground truth perfusion rates (1.3, 2.6, 3.9 mL/g/min) and three vascular-interstitial flow ratios (100-0%, 75-25%, 50-50%). DCE-MRI data was quantified using Fermi model-based deconvolution [2] and two-compartmental tracer kinetic modeling [3]. Linear regression and repeated measures analysis of variance (ANOVA) were used to compare perfusion estimates from both methods with ground truth.
Results:
Figure 2 shows the DCE-MRI myocardial signal intensity-time curves and corresponding velocity maps from PC-MRI. Without extravasation, the total signal is characteristic to a one-compartment system where the contrast agent rapidly traverses the capillaries. Extravasation leads to a proportional increase in the signal in the interstitial compartment that prolongs the clearance of the contrast agent. Equivalently, velocity maps demonstrate a flow increase in the interstitial compartment and a flow decrease in the vascular compartment with increasing extravasation. Global perfusion estimates were not significantly different between methods (F(1.339, 22.764) = 0.166, p = 0.760), despite a better linear correlation between two-compartmental modeling and ground truth (R2=0.647) compared with Fermi model-based deconvolution (R2=0.452).
Conclusion:
A synthetic myocardium simulating two-compartmental exchange was developed. The myocardium is MR-compatible, can be manufactured reproducibly and can be used to validate novel perfusion techniques based on either intravascular or extracellular contrast dynamics. Future work will focus on optimizing the design to facilitate voxel-wise validation.