Intravoxel Incoherent Motion applied to Cardiac diffusion weighted MRI using breath-hold acquisitions in healthy volunteers

Bénédicte M. A. Delattre, Magalie Viallon, Hongjiang Wei, Yuemin Zhu, Vinay Pai, Han Wen, Pierre CroisilleJournal of Cardiovascular Magnetic Resonance 2012 · DOI

Abstract

This study demonstrates that Intravoxel Incoherent Motion (IVIM) MRI can successfully map heart perfusion and diffusion in humans without using contrast agents. By employing a specialized filtering and reconstruction technique called PCATMIP, researchers were able to overcome the significant signal loss typically caused by the heart's motion and heart-rate variability.

Introduction

Measuring blood flow (perfusion) in the heart usually requires injecting contrast agents like Gadolinium. While effective, there is a growing interest in "non-contrast" methods. Diffusion-Weighted Imaging (DWI) and the Intravoxel Incoherent Motion (IVIM) model offer a way to measure both the microscopic movement of water (diffusion) and the flow of blood in the capillary network (perfusion) simultaneously.

However, applying IVIM to the heart is notoriously difficult. The heart’s constant beating and the variability in heart rate (RR variability) cause "signal loss"—essentially, the motion of the heart mimics the diffusion of water, making the data noisy and inaccurate. This paper tests whether a specific image processing strategy can "rescue" this signal and provide reliable measurements in healthy volunteers.

Methods

The researchers performed breath-hold DWI scans on 12 healthy volunteers. To tackle the motion problem, they used two main strategies:

  1. Multi-Phase Acquisition: Instead of taking one image, they acquired images at 10 different "trigger-delay" points during the diastolic (resting) phase of the heart cycle.
  2. PCATMIP Processing: They applied Principal Component Analysis (PCA) to filter out noise and temporal Maximum Intensity Projection (tMIP). This technique essentially looks across all 10 time points and selects/reconstructs the signal that is least affected by motion.

They compared the results of this PCATMIP method against a standard approach of simply picking the single "best" diastolic time point (1TD). The data was then fitted to the IVIM model to extract three parameters: D (true diffusion coefficient), f (perfusion fraction), and D* (pseudo-diffusion coefficient).

Key Results and Insights

The study successfully produced high-quality maps of perfusion and diffusion across the left ventricle, proving that IVIM is a viable tool for human cardiac imaging.

Signal intensity decay curves across different trigger delays (TD), illustrating the impact of cardiac motion on signal stability and the variability that PCATMIP aims to address.
  • Motion Recovery: The most significant finding was the confirmation that PCATMIP could recover signal lost to heart-rate variability. As shown in the signal decay analysis, even when a volunteer has an unstable heart rhythm, the algorithm can "find" the true signal that would otherwise be buried in noise.
  • Perfusion Underestimation: The study found that using a single time point (1TD) significantly underestimates the perfusion fraction (f). The PCATMIP method yielded a significantly higher perfusion fraction (0.163) compared to the single-point method (0.118), with a p-value of 0.0018. This suggests that previous attempts at cardiac IVIM might have been reporting lower-than-actual blood flow values.
  • Diffusion Stability: Interestingly, the true diffusion coefficient (D) remained consistent regardless of the method used (~2.35 ± 1.12 x 10⁻³ mm²/s). This indicates that while perfusion measurements are highly sensitive to motion, the measurement of water diffusion in the heart muscle is relatively robust.
Diffusion-weighted images of the heart at various b-values, demonstrating the visual quality and signal attenuation used for IVIM modeling.

Limitations and Future Work

The study was limited to a cohort of 12 healthy volunteers. Patients with heart disease may have more erratic heart rhythms or different tissue properties that could challenge the PCATMIP algorithm. Furthermore, the protocol requires multiple breath-holds, which might be difficult for patients with heart failure or respiratory distress. The next step is clinical validation to determine if these IVIM parameters can detect ischemia in patients with coronary artery disease without the need for contrast injection.

Summary updated

Abstract

Poster presentation
BibTeX
@article{delattre2012intravoxel,
  author = {Bénédicte M. A. Delattre and Magalie Viallon and Hongjiang Wei and Yuemin Zhu and Vinay Pai and Han Wen and Pierre Croisille},
  title = {{Intravoxel Incoherent Motion applied to Cardiac diffusion weighted MRI using breath-hold acquisitions in healthy volunteers}},
  journal = {Journal of Cardiovascular Magnetic Resonance},
  year = {2012},
  doi = {10.1186/1532-429x-14-s1-p261},
}
Figure