Esther Raithel

Six-Fold Acceleration of High-Spatial Resolution 3D SPACE MRI of the Knee Through Incoherent k-Space Undersampling and Iterative Reconstruction-First Experience.

PurposeThe aim of this study was to prospectively test the hypothesis that 6-fold acceleration of a 3-dimensional (3D) turbo spin echo (TSE) magnetic resonance imaging (MRI) pulse sequence with k-space undersampling and iterative reconstruction is feasible for fast high spatial resolution MRI of the knee, while yielding similar image quality and diagnostic performance when compared with a conventional 2-dimensional (2D) TSE MRI standard. Materials and MethodsThe study was approved by the institutional review board. A 10-minute isotropic 3D TSE knee protocol consisting of accelerated intermediate-weighted (repetition time, 900 milliseconds; echo time, 29 milliseconds; voxel size, 0.5 × 0.5 × 0.5 mm3; acquisition time, 4:45 minutes) and fat-saturated T2-weighted (repetition time, 900 milliseconds; echo time, 92 milliseconds; voxel size, 0.5 × 0.5 × 0.5 mm3; acquisition time, 5:10 minutes) SPACE (sampling perfection with application optimized contrast using different flip angle evolutions) sequence prototypes was compared against a 20-minute 2D TSE standard protocol. The accelerated SPACE sequences were equipped with an optional variable-density poisson-disc pattern as an undersampling mask. An undersampling factor of 0.17 was chosen (6-fold acceleration compared with an acquisition with full sampling). An iterative, sensitivity encoding-type reconstruction with L1 norm-based regularization term was used. The study was performed on a 3 T MRI system using a 15-channel transmit/receive knee coil. The study groups included 15 asymptomatic volunteers and 15 patient volunteers. Quantitative and qualitative assessments were performed by 2 observers. Outcome variables included signal and contrast-to-noise ratio, image quality, and diagnostic accuracy. Qualitative and quantitative measurements were statistically analyzed using nonparametric tests. P values of less than 0.01 were considered significant. ResultsThe signal-to-noise ratios of 2D and 3D MRI were similar with the exception of fluid, which was brighter on 2D MRI. Relevant contrast-to-noise ratios of 2D MRI were higher than 3D MRI; however, observer ratings for satisfaction, image quality, and visibility of anatomic structures were similar for 2D and 3D MRI. There was moderate to excellent interobserver (&kgr; = 0.54–1.00) and intermethod (&kgr; = 0.54–1.00) agreement for assessing menisci, cartilage, ligaments, cartilage, and bone. Two-dimensional and 3D MRI had similar sensitivity (100%/100%, respectively) and specificity (87%/75%, respectively) for detecting 9 meniscal tears (P = 1.00). ConclusionsWe demonstrate the successful clinical implementation of 3D TSE MRI with incoherent k-space undersampling and iterative reconstruction for 6-fold accelerated high spatial resolution isotropic 3D MRI data acquisition. Our preliminary assessments suggest similar image quality and diagnostic performance of a comprehensive 10-minute 3D TSE MRI prototype protocol and 20-minute TSE MRI standard protocol.

Compressed Sensing SEMAC: 8-fold Accelerated High Resolution Metal Artifact Reduction MRI of Cobalt-Chromium Knee Arthroplasty Implants.

ObjectiveThe aim of this study was to prospectively test the hypothesis that a compressed sensing–based slice encoding for metal artifact correction (SEMAC) turbo spin echo (TSE) pulse sequence prototype facilitates high-resolution metal artifact reduction magnetic resonance imaging (MRI) of cobalt-chromium knee arthroplasty implants within acquisition times of less than 5 minutes, thereby yielding better image quality than high-bandwidth (BW) TSE of similar length and similar image quality than lengthier SEMAC standard of reference pulse sequences. Materials and MethodsThis prospective study was approved by our institutional review board. Twenty asymptomatic subjects (12 men, 8 women; mean age, 56 years; age range, 44–82 years) with total knee arthroplasty implants underwent MRI of the knee using a commercially available, clinical 1.5 T MRI system. Two compressed sensing–accelerated SEMAC prototype pulse sequences with 8-fold undersampling and acquisition times of approximately 5 minutes each were compared with commercially available high-BW and SEMAC pulse sequences with acquisition times of approximately 5 minutes and 11 minutes, respectively. For each pulse sequence type, sagittal intermediate-weighted (TR, 3750–4120 milliseconds; TE, 26–28 milliseconds; voxel size, 0.5 × 0.5 × 3 mm3) and short tau inversion recovery (TR, 4010 milliseconds; TE, 5.2–7.5 milliseconds; voxel size, 0.8 × 0.8 × 4 mm3) were acquired. Outcome variables included image quality, display of the bone-implant interfaces and pertinent knee structures, artifact size, signal-to-noise ratio (SNR), and contrast-to-noise ratio (CNR). Statistical analysis included Friedman, repeated measures analysis of variances, and Cohen weighted k tests. Bonferroni-corrected P values of 0.005 and less were considered statistically significant. ResultsImage quality, bone-implant interfaces, anatomic structures, artifact size, SNR, and CNR parameters were statistically similar between the compressed sensing–accelerated SEMAC prototype and SEMAC commercial pulse sequences. There was mild blur on images of both SEMAC sequences when compared with high-BW images (P < 0.001), which however did not impair the assessment of knee structures. Metal artifact reduction and visibility of central knee structures and bone-implant interfaces were good to very good and significantly better on both types of SEMAC than on high-BW images (P < 0.004). All 3 pulse sequences showed peripheral structures similarly well. The implant artifact size was 46% to 51% larger on high-BW images when compared with both types of SEMAC images (P < 0.0001). Signal-to-noise ratios and CNRs of fat tissue, tendon tissue, muscle tissue, and fluid were statistically similar on intermediate-weighted MR images of all 3 pulse sequence types. On short tau inversion recovery images, the SNRs of tendon tissue and the CNRs of fat and fluid, fluid and muscle, as well as fluid and tendon were significantly higher on SEMAC and compressed sensing SEMAC images (P < 0.005, respectively). ConclusionsWe accept the hypothesis that prospective compressed sensing acceleration of SEMAC is feasible for high-quality metal artifact reduction MRI of cobalt-chromium knee arthroplasty implants in less than 5 minutes and yields better quality than high-BW TSE and similarly high quality than lengthier SEMAC pulse sequences.

Three-Dimensional CAIPIRINHA SPACE TSE for 5-Minute High-Resolution MRI of the Knee.

ObjectiveThe aim of this study was to prospectively test the hypothesis that a 2-dimensional (2D) CAIPIRINHA (controlled aliasing in parallel imaging results in higher acceleration) sampling pattern facilitates 5-minute high spatial resolution 3-dimensional (3D) sampling perfection with application optimized contrast using different flip angle evolutions (SPACE) magnetic resonance imaging (MRI) of the knee with image quality similar or better than current 2D turbo spin echo (TSE) and 3D SPACE standards. Materials and MethodsThe study was approved by our institutional review board. Twenty asymptomatic subjects (12 men, 8 women; mean age, 42 years; age range, 24–65 years) underwent 3 T MRI of the knee. A 4-fold accelerated 3D SPACE TSE prototype with 2D CAIPIRINHA sampling pattern and 5-minute acquisition time was compared with commercially available 2-fold and 4-fold accelerated 3D SPACE and 2D TSE pulse sequences with acquisition times of 11 minutes and 15 seconds, 6 minutes and 30 seconds, as well as 9 minutes and 48 seconds, respectively. Outcome variables included image quality, anatomic visibility, signal-to-noise ratio (SNR), and contrast-to-noise ratio (CNR). Statistical analysis included Friedman, repeated measures analysis of variances, and Cohens weighted &kgr; tests. Bonferroni-corrected P values of 0.005 and less were considered statistically significant. ResultsOverall, image quality, visibility of anatomic structures, SNR, and CNR of 3D CAIPIRINHA SPACE were statistically similar to 2-fold accelerated 3D SPACE and significantly better than 4-fold accelerated 3D SPACE, which exhibited degrading parallel imaging artifacts. Compared with 2.5-mm 2D TSE images, 0.5-mm 3D CAIPIRINHA SPACE images showed statistically similar good edge sharpness and very good contrast resolution, and significantly less partial volume as well as absent chemical shift and pulsatile flow artifacts. Visibility of menisci, anterior cruciate ligament, posterior cruciate ligament, medial collateral ligament, and lateral collateral ligament was good to very good on 0.5-mm 3D CAIPIRINHA SPACE images as compared with good on 2.5-mm 2D TSE image (P < 0.005). The SNR of fat, fluid, and cartilage as well as CNR between cartilage, fluid, fat, posterior cruciate ligament, and menisci were minimally higher on 2.5-mm 2D TSE image (P < 0.005). Image quality, visibility of anatomic structures, SNR, and CNR of 2.5-mm 3D CAIPIRINHA SPACE and 2.5-mm 2D TSE images were good to very good without significant differences. ConclusionsThree-dimensional SPACE with 2D CAIPIRINHA sampling pattern enables high-quality 3D TSE MRI of the knee at an acquisition time of 5 minutes and image quality, visibility of anatomic structures, SNR, and CNR similar to conventional 3D SPACE and 2D TSE, both of which require approximately 10-minute acquisition times.

Three-dimensional MR Cholangiopancreatography in a Breath Hold with Sparsity-based Reconstruction of Highly Undersampled Data

Purpose To develop a three-dimensional breath-hold (BH) magnetic resonance (MR) cholangiopancreatographic protocol with sampling perfection with application-optimized contrast using different flip-angle evolutions (SPACE) acquisition and sparsity-based iterative reconstruction (SPARSE) of prospectively sampled 5% k-space data and to compare the results with conventional respiratory-triggered (RT) acquisition. Materials and Methods This HIPAA-compliant prospective study was institutional review board approved. Twenty-nine patients underwent conventional RT SPACE and BH-accelerated SPACE acquisition with 5% k-space sampling at 3 T. Spatial resolution and other parameters were matched when possible. BH SPACE images were reconstructed by enforcing joint multicoil sparsity in the wavelet domain (SPARSE-SPACE). Two board-certified radiologists independently evaluated BH SPARSE-SPACE and RT SPACE images for image quality parameters in the pancreatic duct and common bile duct by using a five-point scale. The Wilcoxon signed-rank test was used to compare BH SPARSE-SPACE and RT SPACE images. Results Acquisition time for BH SPARSE-SPACE was 20 seconds, which was significantly (P < .001) shorter than that for RT SPACE (mean ± standard deviation, 338.8 sec ± 69.1). Overall image quality scores were higher for BH SPARSE-SPACE than for RT SPACE images for both readers for the proximal, middle, and distal pancreatic duct, but the difference was not statistically significant (P > .05). For reader 1, distal common bile duct scores were significantly higher with BH SPARSE-SPACE acquisition (P = .036). More patients had acceptable or better overall image quality (scores ≥ 3) with BH SPARSE-SPACE than with RT SPACE acquisition, respectively, for the proximal (23 of 29 [79%] vs 22 of 29 [76%]), middle (22 of 29 [76%] vs 18 of 29 [62%]), and distal (20 of 29 [69%] vs 13 of 29 [45%]) pancreatic duct and the proximal (25 of 28 [89%] vs 22 of 28 [79%]) and distal (25 of 28 [89%] vs 24 of 28 [86%]) common bile duct. Conclusion BH SPARSE-SPACE showed similar or superior image quality for the pancreatic and common duct compared with that of RT SPACE despite 17-fold shorter acquisition time. (©) RSNA, 2016.

An L1-norm phase constraint for half-Fourier compressed sensing in 3D MR imaging.

AbstractObjectiveIn most half-Fourier imaging methods, explicit phase replacement is used. In combination with parallel imaging, or compressed sensing, half-Fourier reconstruction is usually performed in a separate step. The purpose of this paper is to report that integration of half-Fourier reconstruction into iterative reconstruction minimizes reconstruction errors.Materials and methodsThe L1-norm phase constraint for half-Fourier imaging proposed in this work is compared with the L2-norm variant of the same algorithm, with several typical half-Fourier reconstruction methods. Half-Fourier imaging with the proposed phase constraint can be seamlessly combined with parallel imaging and compressed sensing to achieve high acceleration factors.ResultsIn simulations and in in-vivo experiments half-Fourier imaging with the proposed L1-norm phase constraint enables superior performance both reconstruction of image details and with regard to robustness against phase estimation errors.ConclusionThe performance and feasibility of half-Fourier imaging with the proposed L1-norm phase constraint is reported. Its seamless combination with parallel imaging and compressed sensing enables use of greater acceleration in 3D MR imaging.

Improving the robustness of 3D turbo spin echo imaging to involuntary motion

AbstractObjective3D TSE imaging is very prone to motion artifacts, especially from uncooperative patients, because of the long scan duration. The need to repeat this time-consuming 3D acquisition in the event of large motion artifacts substantially reduces patient comfort and increases the workload of the scanner. Materials and methodsA new sampling strategy enables homogenized collection of k-space data for 3D TSE imaging. It is combined with Frobenius norm-based motion-detection to enable freely stopped acquisition in 3D TSE imaging whenever excessive subject motion is detected.ResultsThe feasibility and reliability of the proposed method were demonstrated and evaluated in in-vivo experiments.ConclusionIt is shown that the additional overhead related to repeat scanning of the 3D TSE sequence as a result of patient motion can be substantially reduced by using the homogenized k-space sampling strategy with automatic scan completion as determined by Frobenius norm-based motion-detection.

Sparse-SEMAC: rapid and improved SEMAC metal implant imaging using SPARSE-SENSE acceleration.

To develop an accelerated SEMAC metal implant MRI technique (Sparse‐SEMAC) with reduced scan time and improved metal distortion correction.

Clinical Feasibility of 3-Dimensional Magnetic Resonance Cholangiopancreatography Using Compressed Sensing: Comparison of Image Quality and Diagnostic Performance

Objective The aim of this study was to evaluate the clinical feasibility of fast 3-dimensional (3D) magnetic resonance cholangiopancreatography (MRCP) using compressed sensing (CS) in comparison with conventional navigator-triggered 3D-MRCP. Materials and Methods This retrospective study was approved by our institutional review board, and the requirement of informed consent was waived. A total of 84 patients (male-to-female ratio, 41:43; mean age, 47.3 ± 18.8 years) who underwent conventional 3D navigator-triggered T2-weighted MRCP using sampling perfection with application optimized contrasts (SPACE) and fast 3D MRCP using SPACE with high undersampling combined with CS reconstruction (CS SPACE; CS-MRCP) on a 3 T scanner were included. Among them, 28 patients additionally underwent 3D breath-hold CS-MRCP (BH-CS-MRCP) with 5.7% k-space sampling. Three board-certified radiologists then independently reviewed the examinations for bile duct and pancreatic duct visualization and overall image quality on a 5-point scale, and image sharpness and background suppression on a 4-point scale, with the higher score indicating better image quality. In addition, diagnostic performance for the detection of anatomic variation and diseases of the bile duct, and pancreatic disease were assessed on a per-patient basis in the subgroup of 28 patients who underwent conventional MRCP, CS-MRCP, and BH-CS-MRCP in the same manner. Results Mean acquisition times of conventional MRCP, CS-MRCP, and BH-CS-MRCP were 7 minutes (419.7 seconds), 3 minutes 47 seconds (227.0 seconds), and 16 seconds, respectively (P < 0.0001, in all comparisons). In all patients, CS-MRCP showed better image sharpness (3.54 ± 0.60 vs 3.37 ± 0.75, P = 0.04) and visualization of the common bile duct (4.55 ± 0.60 vs 4.39 ± 0.78, P = 0.034) and pancreatic duct (3.47 ± 1.22 vs 3.26 ± 1.32, P = 0.025), but lower background suppression (3.00 ± 0.54 vs 3.37 ± 0.58, P < 0.001) than conventional MRCP. Overall image quality was not significantly different between the 2 examinations (3.51 ± 0.95 vs 3.47 ± 1.09, P = 0.75). The number of indeterminate MRCP examinations for the anatomic variation and disease of the bile duct significantly decreased on CS-MRCP, from 16.7%–22.6% to 9.5%–11.9% and 8.4%–15.6% to 3.6%–8.4% in all readers (P = 0.003–0.03). In the 28 patients who underwent BH-CS-MRCP, better image quality was demonstrated than with conventional MRCP and CS-MRCP (4.10 ± 0.84 vs 3.44 ± 1.21 vs 3.50 ± 1.11, respectively, P = 0.002, 0.001). Sensitivities for detecting bile duct disease was 88.9% to 100% on both BH-CS-MRCP and conventional MRCP (P > 0.05), and for detecting pancreatic disease was 66.7% to 83.3% on BH-CS-MRCP and 50.0% to 72.2% on conventional MRCP (P = 0.002 in reader 1, 0.06–0.47 in readers 2–3). Conclusions Compressed sensing MRCP using incoherent undersampling combined with CS reconstruction provided comparable image quality to conventional MRCP while reducing the acquisition time to within a single breath-hold (16 seconds).

Instrument visualization using conventional and compressed sensing SEMAC for interventional MRI at 3T

Interventional magnetic resonance imaging (MRI) at 3T benefits from higher spatial and temporal resolution, but artifacts of metallic instruments are often larger and may obscure target structures.

Evaluation of 2-point, 3-point, and 6-point Dixon magnetic resonance imaging with flexible echo timing for muscle fat quantification

The purpose of this study is to evaluate and compare 2-point (2pt), 3-point (3pt), and 6-point (6pt) Dixon magnetic resonance imaging (MRI) sequences with flexible echo times (TE) to measure proton density fat fraction (PDFF) within muscles. Two subject groups were recruited (G1: 23 young and healthy men, 31 ± 6 years; G2: 50 elderly men, sarcopenic, 77 ± 5 years). A 3-T MRI system was used to perform Dixon imaging on the left thigh. PDFF was measured with six Dixon prototype sequences: 2pt, 3pt, and 6pt sequences once with optimal TEs (in- and opposed-phase echo times), lower resolution, and higher bandwidth (optTE sequences) and once with higher image resolution (highRes sequences) and shortest possible TE, respectively. Intra-fascia PDFF content was determined. To evaluate the comparability among the sequences, Bland-Altman analysis was performed. The highRes 6pt Dixon sequences served as reference as a high correlation of this sequence to magnetic resonance spectroscopy has been shown before. The PDFF difference between the highRes 6pt Dixon sequence and the optTE 6pt, both 3pt, and the optTE 2pt was low (between 2.2% and 4.4%), however, not to the highRes 2pt Dixon sequence (33%). For the optTE sequences, difference decreased with the number of echoes used. In conclusion, for Dixon sequences with more than two echoes, the fat fraction measurement was reliable with arbitrary echo times, while for 2pt Dixon sequences, it was reliable with dedicated in- and opposed-phase echo timing.