Fast multi‐component analysis using a joint sparsity constraint for MR fingerprinting

Abstract

This is an open access article under the terms of the Creat ive Commo ns Attri butio n-NonCo mmerc ial-NoDerivs License, which permits use and distribution in any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made. Purpose: To develop an efficient algorithm for multi-component analysis of magnetic resonance fingerprinting (MRF) data without making a priori assumptions about the exact number of tissues or their relaxation properties. Methods: Different tissues or components within a voxel are potentially separable in MRF because of their distinct signal evolutions. The observed signal evolution in each voxel can be described as a linear combination of the signals for each component with a non-negative weight. An assumption that only a small number of components are present in the measured field of view is usually imposed in the interpretation of multi-component data. In this work, a joint sparsity constraint is introduced to utilize this additional prior knowledge in the multi-component analysis of MRF data. A new algorithm combining joint sparsity and non-negativity constraints is proposed and compared to state-of-the-art multi-component MRF approaches in simulations and brain MRF scans of 11 healthy volunteers. Results: Simulations and in vivo measurements show reduced noise in the estimated tissue fraction maps compared to previously proposed methods. Applying the proposed algorithm to the brain data resulted in 4 or 5 components, which could be attributed to different brain structures, consistent with previous multi-component MRF publications. Conclusions: The proposed algorithm is faster than previously proposed methods for multi-component MRF and the simulations suggest improved accuracy and precision of the estimated weights. The results are easier to interpret compared to voxel-wise methods, which combined with the improved speed is an important step toward clinical evaluation of multi-component MRF. K E Y W O R D S joint sparsity constraint, MR fingerprinting, multi-component analysis, NNLS, partial volume effect, Sparsity Promoting Iterative Joint NNLS (SPIJN) 522 | NAGTEGAAL ET AL. | 523 NAGTEGAAL ET AL.

Variable flip angle echo planar time-resolved imaging (vFA-EPTI) for fast high-resolution gradient echo myelin water imaging

“…In addition, cutting-edge MWI models can be easily combined with EPTI to obtain more parameters of interest such as frequency offsets, since these models are developed based on more conventional sequences of which we can directly incorporate EPTI into. Although it is a challenge to directly apply such advanced MWI models to MRF, several studies have also been conducted recently to combine MRF with a multi-compartment dictionary matching for MWF estimation ( Chen et al, 2019 ; Hamilton et al, 2016 ; Nagtegaal et al, 2020 ) showing promising results.…”

Slice‐selective extended phase graphs in gradient‐crushed, transient‐state free precession sequences: An application to MR fingerprinting

“…The simulations with ssEPG indicate that it accurately captures the highly variable magnitudes of slice-selective profiles that result from unbalanced gradients. Such slice profile modulations may be relevant in the context of partial volume effects, and possibly in multicompartment MRF parameter estimation, 23,24 F I G U R E 5 The magnitude of the slice profiles from the second excitation of the MRF sequences for a TBW of 4 RF pulse and 1 cycle per nominal slice thickness gradient crusher at the listed ∆B 0 s. (A) The magnitude of the MRF signals modeled at the listed ∆B 0 s. (B) The slice profiles and signal magnitudes of ∆B 0 = 1/4 and 1/2 overlap with 5/4 and 3/2, respectively, in (A) and (B). The phase of the MRF signals are shown in (C).…”

“…The simulations with ssEPG indicate that it accurately captures the highly variable magnitudes of slice‐selective profiles that result from unbalanced gradients. Such slice profile modulations may be relevant in the context of partial volume effects, and possibly in multicompartment MRF parameter estimation, 23,24 depending on the length scale of the heterogeneity of the tissue. The ssEPG model has the lowest RMSE (Figure 3) and highest concordance correlation coefficient (Figure 4) compared to other EPG methods.…”

Magnetic resonance fingerprinting: from evolution to clinical applications

“…Partial volume artefacts, which also occur in other volumetric acquisition methods such as computed tomography and conventional MRI, can be diminished with multicompartment models. 2,45,46 In particular, Nagtegaal et al 46 used compressed sensing optimisation and sparsity techniques to model voxels of multicompartment tissue without making restrictive assumptions. This resulted in a robustness to noise that enabled tighter classification bounds on compartment fractions.…”

“…However, Siemens have provided many in the original MRF team with research grants, 16 and have been involved in recent work. 15,30,40,59,63 Philips 46,48,49,69 and GE Healthcare 41,62 researchers have published work relating to MRF, indicating wider industry interest.…”