MARVEL: MR Fingerprinting with Additional micRoVascular Estimates using bidirectional LSTMs

TL;DR

This work addresses extending MR Fingerprinting to simultaneously recover relaxometry ($T_1$, $T_2$), field maps ($B_1$, $B_0$ via $\delta f$), and microvascular properties ($CBV$, $R$) from a single acquisition using a complex balance $bSSFP$ sequence. It introduces an on-the-fly vascular dictionary by convolving a Bloch-equation–based base dictionary with voxel-specific $\delta f$ distributions derived from realistic microvascular networks, paired with a Bidirectional LSTM (BiLSTM) for fast fingerprint matching. The method demonstrates rapid reconstruction (~3.5 s) with high-quality maps for $T_1$, $T_2$, $B_1$, $CBV$, and $R$ in three healthy volunteers and shows robustness to spiral undersampling. The study outlines pathways to clinical deployment, including validation against reference DSC measures and extensions to additional microstructural parameters, with potential applications in stroke and tumor assessment.

Abstract

The Magnetic Resonance Fingerprinting (MRF) approach aims to estimate multiple MR or physiological parameters simultaneously with a single fast acquisition sequence. Most of the MRF studies proposed so far have used simple MR sequence types to measure relaxation times (T1, T2). In that case, deep learning algorithms have been successfully used to speed up the reconstruction process. In theory, the MRF concept could be used with a variety of other MR sequence types and should be able to provide more information about the tissue microstructures. Yet, increasing the complexity of the numerical models often leads to prohibited simulation times, and estimating multiple parameters from one sequence implies new dictionary dimensions whose sizes become too large for standard computers and DL architectures.In this paper, we propose to analyze the MRF signal coming from a complex balance Steady-state free precession (bSSFP) type sequence to simultaneously estimate relaxometry maps (T1, T2), Field maps (B1, B0) as well as microvascular properties such as the local Cerebral Blood Volume (CBV) or the averaged vessel Radius (R).To bypass the curse of dimensionality, we propose an efficient way to simulate the MR signal coming from numerical voxels containing realistic microvascular networks as well as a Bidirectional Long Short-Term Memory network used for the matching process.On top of standard MRF maps, our results on 3 human volunteers suggest that our approach can quickly produce high-quality quantitative maps of microvascular parameters that are otherwise obtained using longer dedicated sequences and intravenous injection of a contrast agent. This approach could be used for the management of multiple pathologies and could be tuned to provide other types of microstructural information.

Figures (6)

• Figure 1: a) Simulations of intra-voxel frequency distribution. b) Creation of a vascular MRF dictionary using a 4-dimensional base dictionary and distributions of frequencies.
• Figure 1: BiLSTM network structure. For training, we used the Adam optimizer with the MSE loss and an initial learning rate of $10^{-3}$, reduced by a factor of 0.8 every 5 epochs.
• Figure 2: Parameter maps of the Cartesian acquisitions obtained with the reconstruction methods studied in this paper. (Note that the slice position slightly differs between the Spoil and bSSFP acquisitions.)
• Figure 2: BiLSTM reconstructions of the spiral acquisition for subject 2. Computed in 3.5 s.
• Figure 3: Parameter maps of one slice of a volunteer of the bSSFP spiral acquisition obtained with dictionary-matching and our BiLSTM network, and associated histograms.