Autocalibrated multiband CAIPIRINHA with through‐time encoding: Proof of principle and application to cardiac tissue phase mapping

Myocardial First-Pass Perfusion With Increased Anatomic Coverage at 3 T Using Autocalibrated Multiband Imaging

BACKGROUND

Myocardial first-pass perfusion (FPP) imaging is a useful cardiac MRI method for the diagnosis of coronary artery disease. However, conventional 2D multislice FPP acquisitions usually have gaps between myocardium slices, which limits the overall assessment of myocardial ischemia.

PURPOSE

To increase the anatomic coverage of myocardial FPP imaging at 3 T by implementing both autocalibrated multiband (MB) acquisition and k-t space acceleration with compress sensing (CS) reconstruction, without the need for additional reference scans.

STUDY TYPE

Phantom and prospective human studies.

PHANTOM/SUBJECTS

A T1MES (T1 Mapping and ECV Standardization in cardiovascular magnetic resonance) phantom and 20 subjects (12 healthy subjects and 8 patients, 10 males, age 42 ± 16 years).

FIELD STRENGTH/SEQUENCE

A 3 T/saturation recovery prepared gradient echo sequence with contrast administration.

ASSESSMENT

Phantom experiments were performed to compare the performance of autocalibrated MB-FPP with k-t acceleration using slice-GRAPPA and CS reconstructions. In vivo experiments were performed to compare the performance of conventional FPP (2.5× acceleration) with autocalibrated MB + CS-FPP (6× acceleration). In phantom experiments, the error maps were calculated. In in vivo experiments, the contrast ratio (CR) and blurring were quantitatively measured, while image quality, perceived signal-to-noise ratio (SNR), and artifact level were qualitatively graded by three cardiologists on a 4-point scale.

STATISTICAL TESTS

Wilcoxon signed-rank test, paired t-test. A P value <0.05 was considered statistically significant.

RESULTS

In phantom experiments, residual artifact was reduced using the MB + CS-FPP reconstruction method compared with using the MB + slice-GRAPPA reconstruction method. In in vivo experiments, the proposed autocalibrated MB + CS-FPP method demonstrated significantly higher CR (3.52 ± 0.78 vs 2.91 ± 0.81) and had significantly better perceived SNR (2.69 ± 0.29 vs 2.48 ± 0.31) compared to the conventional sequence. Compared with conventional FPP, MB + CS-FPP doubled the spatial coverage (MB + CS-FPP vs conventional FPP) without compromising the image quality (2.69 ± 0.26 vs 2.60 ± 0.30) or increasing the artifact level (2.60 ± 0.26 vs 2.52 ± 0.31).

CONCLUSION

Autocalibrated MB + CS-FPP improved the myocardial coverage and achieved comparable image quality with the same spatial resolution and scan time as conventional FPP and is a promising technique for clinical myocardial perfusion imaging.

EVIDENCE LEVEL

2 TECHNICAL EFFICACY: Stage 3.

Generalized self-calibrating simultaneous multi-slice MR image reconstruction from 3D Fourier encoding perspective

This work introduces a novel, k-space based one-step solution for simultaneous multi-slice MR image reconstruction from 3D Fourier encoding perspective. With undersampled SMS imaging, image reconstruction suffers from both inter-slice leakages and in-plane aliasing artifacts. Aliasing separation becomes further challenging in the presence of discrepancies between calibration and imaging. To address them, in this work a measured SMS 3D k-space with additional calibrating signals is decomposed into SMS imaging and self-calibrating data sets. Extended controlled aliasing is performed by upsampling the measured data in the k_z-direction. A slice-specific null space operator is then learned using extended self-calibration exploiting target slices and additional in-plane-shifted images. Inter-slice leakages and in-plane aliasing artifacts are jointly resolved in a single step by solving a constrained optimization problem in which null space reconstruction consistency is balanced with a Hankel-structured low rank prior while data fidelity in 3D Fourier space is enforced. Retrospective and prospective studies are performed to validate the effectiveness of the proposed method in various regions including knee and L-spine.

All-systolic first-pass myocardial rest perfusion at a long saturation time using simultaneous multi-slice imaging and compressed sensing acceleration

PURPOSE

To enable all-systolic first-pass rest myocardial perfusion with long saturation times. To investigate the change in perfusion contrast and dark rim artefacts through simulations and surrogate measurements.

METHODS

Simulations were employed to investigate optimal saturation time for myocardium-perfusion defect contrast and blood-to-myocardium signal ratios. Two saturation recovery blocks with long/short saturation times (LTS/STS) were employed to image 3 slices at end-systole and diastole. Simultaneous multi-slice balanced steady state free precession imaging and compressed sensing acceleration were combined. The sequence was compared to a 3 slice-by-slice clinical protocol in 10 patients. Quantitative assessment of myocardium-peak pre contrast and blood-to-myocardium signal ratios, as well as qualitative assessment of perceived SNR, image quality, blurring, and dark rim artefacts, were performed.

RESULTS

Simulations showed that with a bolus of 0.075 mmol/kg, a LTS of 240-470 ms led to a relative increase in myocardium-perfusion defect contrast of 34% ± 9%-28% ± 27% than a STS = 120 ms, while reducing blood-to-myocardium signal ratio by 18% ± 10%-32% ± 14% at peak myocardium. With a bolus of 0.05 mmol/kg, LTS was 320-570 ms with an increase in myocardium-perfusion defect contrast of 63% ± 13%-62% ± 29%. Across patients, LTS led to an average increase in myocardium-peak pre contrast of 59% (P < .001) at peak myocardium and a lower blood-to-myocardium signal ratio of 47% (P < .001) and 15% (P < .001) at peak blood/myocardium. LTS had improved motion robustness (P = .002), image quality (P < .001), and decreased dark rim artefacts (P = .008) than the clinical protocol.

CONCLUSION

All-systolic rest perfusion can be achieved by combining simultaneous multi-slice and compressed sensing acceleration, enabling 3-slice cardiac coverage with reduced motion and dark rim artefacts. Numerical simulations indicate that myocardium-perfusion defect contrast increases at LTS.

Minimum TR radiofrequency-pulse design for rapid gradient echo sequences

PURPOSE

A framework to design radiofrequency (RF) pulses specifically to minimize the TR of gradient echo sequences is presented, subject to hardware and physiological constraints.

METHODS

Single-band and multiband (MB) RF pulses can be reduced in duration using variable-rate selective excitation (VERSE) VERSE for a range of flip angles; however, minimum-duration pulses do not guarantee minimum TR because these can lead to a high specific absorption rate (SAR). The optimal RF pulse is found by meeting spatial encoding, peripheral nerve stimulation (PNS) and SAR constraints. A TR reduction for a range of designs is achieved and an application of this in an MB cardiac balanced steady-state free-precession (bSSFP) experiment is presented. Gradient imperfections and their imaging effects are also considered.

RESULTS

Sequence TR with low-time bandwidth product (TBP) pulses, as used in bSSFP, was reduced up to 14%, and the TR when using high TBP pulses, as used in slab-selective imaging, was reduced by up to 72%. A breath-hold cardiac exam was reduced by 46% using both MB and the TR-optimal framework. The importance of RF-based correction of gradient imperfections is demonstrated. PNS was not a practical limitation.

CONCLUSION

The TR-optimal framework designs RF pulses for a range of pulse parameters, specifically to minimize sequence TR.

Split-slice training and hyperparameter tuning of RAKI networks for simultaneous multi-slice reconstruction

PURPOSE

Simultaneous multi-slice acquisitions are essential for modern neuroimaging research, enabling high temporal resolution functional and high-resolution q-space sampling diffusion acquisitions. Recently, deep learning reconstruction techniques have been introduced for unaliasing these accelerated acquisitions, and robust artificial-neural-networks for k-space interpolation (RAKI) have shown promising capabilities. This study systematically examines the impacts of hyperparameter selections for RAKI networks, and introduces a novel technique for training data generation which is analogous to the split-slice formalism used in slice-GRAPPA.

METHODS

RAKI networks were developed with variable hyperparameters and with and without split-slice training data generation. Each network was trained and applied to five different datasets including acquisitions harmonized with Human Connectome Project lifespan protocol. Unaliasing performance was assessed through L1 errors computed between unaliased and calibration frequency-space data.

RESULTS

Split-slice training significantly improved network performance in nearly all hyperparameter configurations. Best unaliasing results were achieved with three layer RAKI networks using at least 64 convolutional filters with receptive fields of 7 voxels, 128 single-voxel filters in the penultimate RAKI layer, batch normalization, and no training dropout with the split-slice augmented training dataset. Networks trained without the split-slice technique showed symptoms of network over-fitting.

CONCLUSIONS

Split-slice training for simultaneous multi-slice RAKI networks positively impacts network performance. Hyperparameter tuning of such reconstruction networks can lead to further improvements in unaliasing performance.

CINENet: deep learning-based 3D cardiac CINE MRI reconstruction with multi-coil complex-valued 4D spatio-temporal convolutions

Cardiac CINE magnetic resonance imaging is the gold-standard for the assessment of cardiac function. Imaging accelerations have shown to enable 3D CINE with left ventricular (LV) coverage in a single breath-hold. However, 3D imaging remains limited to anisotropic resolution and long reconstruction times. Recently deep learning has shown promising results for computationally efficient reconstructions of highly accelerated 2D CINE imaging. In this work, we propose a novel 4D (3D + time) deep learning-based reconstruction network, termed 4D CINENet, for prospectively undersampled 3D Cartesian CINE imaging. CINENet is based on (3 + 1)D complex-valued spatio-temporal convolutions and multi-coil data processing. We trained and evaluated the proposed CINENet on in-house acquired 3D CINE data of 20 healthy subjects and 15 patients with suspected cardiovascular disease. The proposed CINENet network outperforms iterative reconstructions in visual image quality and contrast (+ 67% improvement). We found good agreement in LV function (bias ± 95% confidence) in terms of end-systolic volume (0 ± 3.3 ml), end-diastolic volume (− 0.4 ± 2.0 ml) and ejection fraction (0.1 ± 3.2%) compared to clinical gold-standard 2D CINE, enabling single breath-hold isotropic 3D CINE in less than 10 s scan and ~ 5 s reconstruction time.

Autocalibrated cardiac tissue phase mapping with multiband imaging and k-t acceleration

PURPOSE

To develop an autocalibrated multiband (MB) CAIPIRINHA acquisition scheme with in-plane k-t acceleration enabling multislice three-directional tissue phase mapping in one breath-hold.

METHODS

A k-t undersampling scheme was integrated into a time-resolved electrocardiographic-triggered autocalibrated MB gradient-echo sequence. The sequence was used to acquire data on 4 healthy volunteers with MB factors of two (MB2) and three (MB3), which were reconstructed using a joint reconstruction algorithm that tackles both k-t and MB acceleration. Forward simulations of the imaging process were used to tune the reconstruction model hyperparameters. Direct comparisons between MB and single-band tissue phase-mapping measurements were performed.

RESULTS

Simulations showed that the velocities could be accurately reproduced with MB2 k-t (average ± twice the SD of the RMS error of 0.08 ± 0.22 cm/s and velocity peak reduction of 1.03% ± 6.47% compared with fully sampled velocities), whereas acceptable results were obtained with MB3 k-t (RMS error of 0.13 ± 0.58 cm/s and peak reduction of 2.21% ± 13.45%). When applied to tissue phase-mapping data, the proposed technique allowed three-directional velocity encoding to be simultaneously acquired at two/three slices in a single breath-hold of 18 heartbeats. No statistically significant differences were detected between MB2/MB3 k-t and single-band k-t motion traces averaged over the myocardium. Regional differences were found, however, when using the American Heart Association model for segmentation.

CONCLUSION

An autocalibrated MB k-t acquisition/reconstruction framework is presented that allows three-directional velocity encoding of the myocardial velocities at multiple slices in one breath-hold.

Isotropic 3D Cartesian single breath-hold CINE MRI with multi-bin patch-based low-rank reconstruction

PURPOSE

To develop a novel acquisition and reconstruction framework for isotropic 3D Cartesian cardiac CINE within a single breath-hold for left ventricle (LV) and whole-heart coverage.

METHODS

A variable-density Cartesian acquisition with spiral profile ordering, out-inward sampling, and acquisition-adaptive alternating tiny golden/golden angle increment between spiral arms is proposed to provide incoherent and nonredundant sampling within and among cardiac phases. A novel multi-bin patch-based low-rank reconstruction, named MB-PROST, is proposed to exploit redundant information on a local (within a patch), nonlocal (similar patches within a spatial neighborhood), and temporal (among all cardiac phases) scale with an implicit motion alignment among patches. The proposed multi-bin patch-based low-rank reconstruction reconstruction is compared against compressed sensing reconstruction, whereas LV function parameters derived from the proposed 3D CINE framework are compared against those estimated from conventional multislice 2D CINE imaging in 10 healthy subjects and 15 patients.

RESULTS

The proposed framework provides 3D cardiac CINE images with high spatial (1.9 mm³ ) and temporal resolution (˜50 ms) in a single breath-hold of ˜20 s for LV and ˜26 s for whole-heart coverage in healthy subjects. Shorter breath-hold durations of ˜13 to 15 s are feasible for LV coverage with slightly anisotropic resolution (1.9 × 1.9 × 2.5 mm) in patients. LV function parameters derived from 3D CINE were in good agreement with 2D CINE, with a bias of -0.1 mL/0.1 mL, -0.9 mL/-1.0 mL, -0.1%/-0.8%; and confidence intervals of ±1.7 mL/±3.7 mL, ±1.2 mL/±2.6 mL, and ±1.2%/±3.6% (10 healthy subjects/15 patients) for end-systolic volume, end-diastolic volume, and ejection fraction, respectively.

CONCLUSION

The proposed framework enables 3D isotropic cardiac CINE in a single breath-hold scan of ˜20 s/˜26 s for LV/whole-heart coverage, showing good agreement with clinical 2D CINE scans in terms of LV functional assessment.

Bilateral Multiband 4D Flow MRI of the Carotid Arteries at 7T

PURPOSE

Simultaneous multislab (SMSb) 4D flow MRI was developed and implemented at 7T for accelerated acquisition of the 3D blood velocity vector field in both carotid bifurcations.

METHODS

SMSb was applied to 4D flow to acquire blood velocities in both carotid bifurcations in sagittal orientation using a local transmit/receive coil at 7T. B 1 + transmit efficiency was optimized by B 1 + shimming. SMSb 4D flow was obtained in 8 healthy subjects in single-band (SB) and multiband (MB) fashion. Additionally, MB data were retrospectively undersampled to simulate GRAPPA R = 2 (MB2_GRAPPA2), and both SB datasets were added to form an artificial MB dataset (SumSB). The band separation performance was quantified by signal leakage. Peak velocity and total flow values were calculated and compared to SB via intraclass correlation analysis (ICC).

RESULTS

Clean slab separation was achieved yielding a mean signal leakage of 13% above the mean SB noise level. Mean total flow for MB2, SumSB, and MB_GRAPPA2 deviated less than 9% from the SB values. Peak velocities averaged over all vessels and subjects were 0.48 ± 0.11 m/s for SB, 0.47 ± 0.12 m/s for SumSB, 0.50 ± 0.13 m/s for MB2, and 0.53 ± 0.13 m/s for MB2_GRAPPA2. ICC revealed excellent absolute agreement and consistency of total flow for all methods compared to SB2. Peak velocity showed good to excellent agreement and consistency for SumSB and MB2 and MB2_GRAPPA2 method showed poor to excellent agreement and good to excellent consistency.

CONCLUSION

Simultaneous multislab 4D Flow MRI allows accurate quantification of total flow and peak velocity while reducing scan times.

Simultaneous Multi-VENC and Simultaneous Multi-Slice Phase Contrast Magnetic Resonance Imaging

This work develops a novel, simultaneous multi-VENC and simultaneous multi-slice (SMV+SMS) imaging in a single acquisition for robust phase contrast (PC) MRI. To this end, the pulse sequence was designed to permit concurrent acquisition of multiple VENCs as well as multiple slices on a shared frequency encoding gradient, in which each effective echo time for multiple VENCs was controlled by adjusting net gradient area while multiple slices were simultaneously excited by employing multiband resonance frequency (RF) pulses. For VENC and slice separation, RF phase cycling and gradient blip were applied to create both inter-VENC and inter-slice shifts along phase encoding direction, respectively. With an alternating RF phase cycling that generates oscillating steady-state with low and high signal amplitude, the acquired multi-VENC k-space was reformulated into 3D undersampled k-space by generating a virtual dimension along VENC direction for modulation induced artifact reduction. In vivo studies were conducted to validate the feasibility of the proposed method in comparison with conventional PC MRI. The proposed method shows comparable performance to the conventional method in delineating both low and high flow velocities across cardiac phases with high spatial coverage without apparent artifacts. In the presence of high flow velocity that is above the VENC value, the proposed method exhibits clear depiction of flow signals over conventional method, thereby leading to high VNR image with improved velocity dynamic range.

Altered regional myocardial velocities by tissue phase mapping and feature tracking in pediatric patients with hypertrophic cardiomyopathy

BACKGROUND

Hypertrophic cardiomyopathy (HCM) is associated with heart failure, atrial fibrillation and sudden death. Reduced myocardial function has been reported in HCM despite normal left ventricular (LV) ejection fraction. Additionally, LV fibrosis is associated with elevated T1 and might be an outcome predictor.

OBJECTIVE

To systematically compare tissue phase mapping and feature tracking for assessing regional LV function in children and young adults with HCM and pediatric controls, and to evaluate structure-function relationships among myocardial velocities, LV wall thickness and myocardial T1.

MATERIALS AND METHODS

Seventeen pediatric patients with HCM and 21 age-matched controls underwent cardiac MRI including standard cine imaging, tissue phase mapping (two-dimensional cine phase contrast with three-directional velocity encoding), and modified Look-Locker inversion recovery to calculate native global LV T1. Maximum LV wall thickness was measured on cine images. LV radial, circumferential and long-axis myocardial velocity time courses, as well as global and segmental systolic and diastolic peak velocities, were quantified from tissue phase mapping and feature tracking.

RESULTS

Both tissue phase mapping and feature tracking detected significantly decreased global and segmental diastolic radial and long-axis peak velocities (by 12-51%, P<0.001-0.05) in pediatric patients with HCM vs. controls. Feature tracking peak velocities were lower than directly measured tissue phase mapping velocities (mean bias = 0.3-2.9 cm/s). Diastolic global peak velocities correlated moderately with global T1 (r = -0.57 to -0.72, P<0.01) and maximum wall thickness (r = -0.37 to -0.61, P<0.05).

CONCLUSION

Both tissue phase mapping and feature tracking detected myocardial velocity changes in children and young adults with HCM vs. controls. Associations between impaired diastolic LV velocities and elevated T1 indicate structure-function relationships in HCM.

Simultaneous multislice imaging of the heart using multiband balanced SSFP with blipped-CAIPI

Purpose

In this work, we explore the use of multiband (MB) balanced steady‐state free precession (bSSFP) with blipped‐controlled aliasing in parallel imaging (CAIPI), which avoids the issues of altered frequency response associated with RF phase cycling, and show its application to accelerating cardiac cine imaging.

Methods

Blipped and RF‐cycled CAIPI were implemented into a retrospective‐gated segmented cine multiband bSSFP sequence. The 2 methods were compared at 3T using MB2 to demonstrate the effect on frequency response. Further data (4 subjects) were acquired at both 1.5T and 3T collecting 12‐slice short axis stacks using blipped‐CAIPI with MB acceleration factors of 1–4. The impact on SNR and contrast was evaluated along with g‐factors at different accelerations.

Results

Data acquired with blipped‐CAIPI multiband bSSFP up to factor 4 yielded functional cine data with good SNR and contrast, while reliably keeping dark‐band artefacts clear of the heart at 1.5T. SAR limits the maximum MB acceleration, particularly at 3T, where minimum TR increase is problematic and leakage artefacts are more prevalent. Mean g‐factors across the heart were measured at 1.00, 1.06, and 1.12 for MB2–MB4, whereas blood‐pool SNR measures (end‐diastole) decreased by 11.8, 21.5, and 36.9%; ultimately LV‐myocardium CNR remained sufficient at 1.5T with values ranging: 15.6, 13.4, 11.9, and 9.6 (MB1–MB4).

Conclusion

Blipped‐CAIPI multiband bSSFP can be used in cardiovascular applications without affecting the frequency response because of controlled aliasing and can be readily incorporated into segmented cine acquisitions without adding any additional constraints because of phase cycling requirements. The method was used to collect full ventricular coverage within a single breath‐hold.

Non-Cartesian slice-GRAPPA and slice-SPIRiT reconstruction methods for multiband spiral cardiac MRI

PURPOSE

Spiral MRI has advantages for cardiac imaging, and multiband (MB) spiral MRI of the heart shows promise. However, current reconstruction methods for MB spiral imaging have limitations. We sought to develop improved reconstruction methods for MB spiral cardiac MRI.

METHODS

Two reconstruction methods were developed. The first is non-Cartesian slice-GRAPPA (NCSG), which uses phase demodulation and gridding operations before application of a Cartesian slice-separating kernel. The second method, slice-SPIRiT, formulates the reconstruction as a minimization problem that enforces in-plane coil consistency and consistency with the acquired MB data, and uses through-plane coil sensitivity information in the iterative solution. These methods were compared with conjugate-gradient SENSE in phantoms and volunteers. Temporal alternation of CAIPIRINHA (controlled aliasing in parallel imaging results in higher acceleration) phase and the use of a temporal filter were also investigated.

RESULTS

Phantom experiments with 3 simultaneous slices (MB = 3) showed that mean artifact power was highest for conjugate-gradient SENSE, lower for NCSG, and lowest for slice-SPIRiT. For volunteer cine imaging (MB = 3, N = 5), the artifact power was 0.182 ± 0.037, 0.148 ± 0.036, and 0.139 ± 0.034 for conjugate-gradient SENSE, NCSG, and slice-SPIRiT, respectively (P < .05, analysis of variance). Temporal alternation of CAIPIRINHA reduced artifacts for both NCSG and slice-SPIRiT.

CONCLUSION

The NCSG and slice-SPIRiT methods provide more accurate reconstructions for MB spiral cine imaging compared with conjugate-gradient SENSE. These methods hold promise for non-Cartesian MB imaging.