Simultaneous fat-free isotropic 3D anatomical imaging and T2 mapping of knee cartilage with lipid-insensitive binomial off-resonant RF excitation (LIBRE) pulses

Clinical evaluation of 3D high-resolution isotropic knee MRI using Multi-Interleaved X-prepared TSE with inTUitive RElaxometry (MIXTURE) for simultaneous morphology and T2 mapping

PURPOSE

Quantitative MRI techniques such as T2 mapping are useful in comprehensive evaluation of various pathologies of the knee joint yet require separate scans to conventional morphological measurements and long acquisition times. The recently introduced 3D MIXTURE (Multi-Interleaved X-prepared Turbo-Spin Echo with Intuitive Relaxometry) technique can obtain simultaneous morphologic and quantitative information of the knee joint. To compare MIXTURE with conventional methods and to identify differences in morphological and quantitative information.

METHODS

Phantom studies were conducted, and in vivo human scans were performed (20 patients) presented with knee arthralgia. MIXTURE is based on 3D TSE without and with T2 preparation modules in an interleaved manner for both morphology with PDW and fat suppressed T2W imaging as well as quantitative T2 mapping within one single scan. Image quality and lesion depiction were visually assessed and compared between MIXTURE and conventional 2D TSE by two experienced radiologists. Contrast-to-noise ratio was used to assess the adjacent tissue contrast in a quantitative way for both obtained PDW and fat suppressed T2W images. Quantitative T2 values were measured in phantom and from in vivo knee cartilage.

RESULTS

The overall diagnostic confidence and contrast-to-noise ratio were deemed comparable between MIXTURE and 2D TSE. While the chosen T2 preparation modules for MIXTURE rendered consistent T2 values comparing to the current standard, measured cartilage T2 values ranged from 26.1 to 50.7 ms, with significant difference between the lesion and normal areas (p < 0.05).

CONCLUSIONS

MIXTURE can help to provide high-resolution information for both anatomical and pathological assessment.

Improved accuracy and precision of fat-suppressed isotropic 3D T2 mapping MRI of the knee with dictionary fitting and patch-based denoising

PURPOSE

To develop an isotropic three-dimensional (3D) T2 mapping technique for the quantitative assessment of the composition of knee cartilage with high accuracy and precision.

METHODS

A T2-prepared water-selective isotropic 3D gradient-echo pulse sequence was used to generate four images at 3 T. These were used for three T2 map reconstructions: standard images with an analytical T2 fit (AnT2Fit); standard images with a dictionary-based T2 fit (DictT2Fit); and patch-based-denoised images with a dictionary-based T2 fit (DenDictT2Fit). The accuracy of the three techniques was first optimized in a phantom study against spin-echo imaging, after which knee cartilage T2 values and coefficients of variation (CoV) were assessed in ten subjects in order to establish accuracy and precision in vivo. Data given as mean ± standard deviation.

RESULTS

After optimization in the phantom, whole-knee cartilage T2 values of the healthy volunteers were 26.6 ± 1.6 ms (AnT2Fit), 42.8 ± 1.8 ms (DictT2Fit, p < 0.001 versus AnT2Fit), and 40.4 ± 1.7 ms (DenDictT2Fit, p = 0.009 versus DictT2Fit). The whole-knee T2 CoV reduced from 51.5% ± 5.6% to 30.5 ± 2.4 and finally to 13.1 ± 1.3%, respectively (p < 0.001 between all). The DictT2Fit improved the data reconstruction time: 48.7 ± 11.3 min (AnT2Fit) versus 7.3 ± 0.7 min (DictT2Fit, p < 0.001). Very small focal lesions were observed in maps generated with DenDictT2Fit.

CONCLUSIONS

Improved accuracy and precision for isotropic 3D T2 mapping of knee cartilage were demonstrated by using patch-based image denoising and dictionary-based reconstruction.

KEY POINTS

• Dictionary T2 fitting improves the accuracy of three-dimensional (3D) knee T2 mapping. • Patch-based denoising results in high precision in 3D knee T2 mapping. • Isotropic 3D knee T2 mapping enables the visualization of small anatomical details.

Water excitation with LIBRE pulses in three-dimensional variable flip angle fast spin echo for fat-free and large field of view imaging at 3 tesla

PURPOSE

To develop and evaluate a sequence in which water excitation with lipid insensitive binomial off-resonant radio frequency excitation (LIBRE) pulses is incorporated into three-dimensional (3D) variable flip angle fast spin echo (LIBRE-vf-FSE) for fat-free and large field of view imaging at 3 Tesla (T).

MATERIALS AND METHODS

Numerical simulation was conducted to optimize the parameters of LIBRE pulses, including the flip angle, pulse duration, and frequency offset, for maximizing the fat suppression effect of the proposed LIBRE-vf-FSE sequence. The sequence was then implemented at 3 T and assessed in phantoms, lower extremity imaging of 8 healthy volunteers, and head/neck imaging of 5 healthy volunteers. Conventional water excitation (WE) and fat saturation (FatSat) were also performed for comparison. Signal-to-noise ratio (SNR) of fat and contrast-to-noise ratio (CNR) between fat and water were used to evaluate the level of fat suppression. Standard deviation (SD) of SNR was used to evaluate the uniformity of fat suppression.

RESULTS

The numerical simulation demonstrated that LIBRE-vf-FSE enables large volume imaging with uniform fat suppression, which was further confirmed by phantom and healthy volunteer experiments. LIBRE provided the lowest fat SNR and offered more uniform fat suppression compared with the WE and FatSat. Specifically, average oil SNRs obtained by LIBRE (1.10 ms, 360 Hz, and 60°), WE, and FatSat were (180.1 vs. 280.2 vs. 811.2) in phantom experiments, and average fat SNRs and SDs in legs obtained by LIBRE (1.10 ms, 360 Hz, and 60°), WE, and FatSat were (85.1 vs. 105.0 vs. 105.1) and (22.4 vs. 27.4 vs. 56.4) in vivo experiments, respectively.

CONCLUSION

The proposed LIBRE-vf-FSE sequence allows for fat suppression and large field of view imaging at 3 T. It could be an alternative approach for fat-free vf-FSE scan.

Artificial intelligence, machine learning and deep learning in musculoskeletal imaging: Current applications

Artificial intelligence is rapidly expanding in all technological fields. The medical field, and especially diagnostic imaging, has been showing the highest developmental potential. Artificial intelligence aims at human intelligence simulation through the management of complex problems. This review describes the technical background of artificial intelligence, machine learning, and deep learning. The first section illustrates the general potential of artificial intelligence applications in the context of request management, data acquisition, image reconstruction, archiving, and communication systems. In the second section, the prospective of dedicated tools for segmentation, lesion detection, automatic diagnosis, and classification of musculoskeletal disorders is discussed.

3D MRI of Articular Cartilage

Osteoarthritis, characterized by the breakdown of articular cartilage and other joint structures, is one of the most prevalent and disabling chronic diseases in the United States. Magnetic resonance imaging is a commonly used imaging modality to evaluate patients with joint pain. Both two-dimensional fast spin-echo sequences (2D-FSE) and three-dimensional (3D) sequences are used in clinical practice to evaluate articular cartilage. The 3D sequences have many advantages compared with 2D-FSE sequences, such as their high in-plane spatial resolution, thin continuous slices that reduce the effects of partial volume averaging, and ability to create multiplanar reformat images following a single acquisition. This article reviews the different 3D imaging techniques available for evaluating cartilage morphology, illustrates the strengths and weaknesses of 3D approaches compared with 2D-FSE approaches for cartilage imaging, and summarizes the diagnostic performance of 2D-FSE and 3D sequences for detecting cartilage lesions within the knee and hip joints.

Fat-suppressed T2 mapping of human knee femoral articular cartilage: comparison with conventional T2 mapping

Background

There is paucity of studies applying fat suppressed (FS) technique to T2 mapping to overcome chemical shift artifacts. The purpose of the study is to difference between FS T2 and conventional T2 mapping and reproducibility of FS T2 mapping in the femoral articular cartilage.

Methods

Eighteen patients who had normal-looking femoral cartilage and underwent knee MRI with conventional T2 and FS T2 mapping were included. T2 values of each mapping were measured by two readers independently from nine regions in the medial femoral condyle (MFC) and lateral femoral condyle (LFC). Each anatomical region was divided by lines at ± 10°, 30°, 50°, 70°, 90°, and 110°. Comparisons of T2 values between conventional and FS T2 mapping were statistically analyzed. The T2 values between FS and conventional T2 mapping in the anterior, central and posterior femoral condyles were compared.

Results

The overall femoral condyle T2 values from the FS T2 map were significantly lower than those from the conventional T2 map (48.5ms vs. 51.0ms, p < 0.001). The differences in the T2 values between the two maps were significantly different among the three divisions of the LFC (p = 0.009) and MFC (p = 0.031). The intra-class correlation coefficients indicated higher agreement in the FS T2 map than in the conventional T2 map (0.943 vs. 0.872).

Conclusions

The T2 values of knee femoral cartilage are significantly lower on FS T2 mapping than on conventional T2 mapping. FS T2 mapping is a more reproducible method for evaluating knee femoral cartilage.

A robust broadband fat-suppressing phaser T2 -preparation module for cardiac magnetic resonance imaging at 3T

PURPOSE

Designing a new T₂ -preparation (T₂ -Prep) module to simultaneously provide robust fat suppression and efficient T₂ preparation without requiring an additional fat-suppression module for T₂ -weighted imaging at 3T.

METHODS

The tip-down radiofrequency (RF) pulse of an adiabatic T₂ -Prep module was replaced by a custom-designed RF-excitation pulse that induces a phase difference between water and fat, resulting in a simultaneous T₂ preparation of water signals and the suppression of fat signals at the end of the module (a phaser adiabatic T₂ -Prep). Numerical simulations and in vitro and in vivo electrocardiogram (ECG)-triggered navigator-gated acquisitions of the human heart were performed. Blood, myocardium, and fat signal-to-noise ratios and right coronary artery vessel sharpness were compared against previously published adiabatic T₂ -Prep approaches.

RESULTS

Numerical simulations predicted an increased fat-suppression bandwidth and decreased sensitivity to transmit magnetic field inhomogeneities using the proposed approach while preserving the water T₂ -Prep capabilities. This was confirmed by the tissue signals acquired in the phantom and the in vivo images, which show similar blood and myocardium signal-to-noise ratio, contrast-to-noise ratio, and significantly reduced fat signal-to-noise ratio compared with the other methods. As a result, the right coronary artery conspicuity was significantly increased.

CONCLUSION

A novel fat-suppressing T₂ -Prep method was developed and implemented that showed robust fat suppression and increased vessel sharpness compared with conventional techniques while preserving its T₂ -Prep capabilities.

Applied machine learning and artificial intelligence in rheumatology

Machine learning as a field of artificial intelligence is increasingly applied in medicine to assist patients and physicians. Growing datasets provide a sound basis with which to apply machine learning methods that learn from previous experiences. This review explains the basics of machine learning and its subfields of supervised learning, unsupervised learning, reinforcement learning and deep learning. We provide an overview of current machine learning applications in rheumatology, mainly supervised learning methods for e-diagnosis, disease detection and medical image analysis. In the future, machine learning will be likely to assist rheumatologists in predicting the course of the disease and identifying important disease factors. Even more interestingly, machine learning will probably be able to make treatment propositions and estimate their expected benefit (e.g. by reinforcement learning). Thus, in future, shared decision-making will not only include the patient’s opinion and the rheumatologist’s empirical and evidence-based experience, but it will also be influenced by machine-learned evidence.

Natively fat-suppressed 5D whole-heart MRI with a radial free-running fast-interrupted steady-state (FISS) sequence at 1.5T and 3T

PURPOSE

To implement, optimize, and test fast interrupted steady-state (FISS) for natively fat-suppressed free-running 5D whole-heart MRI at 1.5 tesla (T) and 3T.

METHODS

FISS was implemented for fully self-gated free-running cardiac- and respiratory-motion-resolved radial imaging of the heart at 1.5T and 3T. Numerical simulations and phantom scans were performed to compare fat suppression characteristics and to determine parameter ranges (number of readouts [NR] per FISS module and TR) for effective fat suppression. Subsequently, free-running FISS data were collected in 10 healthy volunteers and images were reconstructed with compressed sensing. All acquisitions were compared with a continuous balanced steady-state free precession version of the same sequence, and both fat suppression and scan times were analyzed.

RESULTS

Simulations demonstrate a variable width and location of suppression bands in FISS that were dependent on TR and NR. For a fat suppression bandwidth of 100 Hz and NR ≤ 8, simulations demonstrated that a TR between 2.2 ms and 3.0 ms is required at 1.5T, whereas a range of 3.0 ms to 3.5 ms applies at 3T. Fat signal increases with NR. These findings were corroborated in phantom experiments. In volunteers, fat SNR was significantly decreased using FISS compared with balanced steady-state free precession (P < 0.05) at both field strengths. After protocol optimization, high-resolution (1.1 mm³ ) 5D whole-heart free-running FISS can be performed with effective fat suppression in under 8 min at 1.5T and 3T at a modest scan time increase compared to balanced steady-state free precession.

CONCLUSION

An optimal FISS parameter range was determined enabling natively fat-suppressed 5D whole-heart free-running MRI with a single continuous scan at 1.5T and 3T, demonstrating potential for cardiac imaging and noncontrast angiography.

Noncontrast free-breathing respiratory self-navigated coronary artery cardiovascular magnetic resonance angiography at 3 T using lipid insensitive binomial off-resonant excitation (LIBRE)

BACKGROUND

Robust and homogeneous lipid suppression is mandatory for coronary artery cardiovascular magnetic resonance (CMR) imaging since the coronary arteries are commonly embedded in epicardial fat. However, effective large volume lipid suppression becomes more challenging when performing radial whole-heart coronary artery CMR for respiratory self-navigation and the problem may even be exacerbated at increasing magnetic field strengths. Incomplete fat suppression not only hinders a correct visualization of the coronary vessels and generates image artifacts, but may also affect advanced motion correction methods. The aim of this study was to evaluate a recently reported lipid insensitive CMR method when applied to a noncontrast self-navigated coronary artery CMR acquisitions at 3 T, and to compare it to more conventional fat suppression techniques.

METHODS

Lipid insensitive binomial off resonant excitation (LIBRE) radiofrequency excitation pulses were included into a self-navigated 3D radial GRE coronary artery CMR sequence at 3 T. LIBRE was compared against a conventional CHESS fat saturation (FS) and a binomial 1-180°-1 water excitation (WE) pulse. First, fat suppression of all techniques was numerically characterized using Matlab and experimentally validated in phantoms and in legs of human volunteers. Subsequently, free-breathing self-navigated coronary artery CMR was performed using the LIBRE pulse as well as FS and WE in ten healthy subjects. Myocardial, arterial and chest fat signal-to-noise ratios (SNR), as well as coronary vessel conspicuity were quantitatively compared among those scans.

RESULTS

The results obtained in the simulations were confirmed by the experimental validations as LIBRE enabled near complete fat suppression for 3D radial imaging in vitro and in vivo. For self-navigated whole-heart coronary artery CMR at 3 T, fat SNR was significantly attenuated using LIBRE compared with conventional FS. LIBRE increased the right coronary artery (RCA) vessel sharpness significantly (37 ± 9% (LIBRE) vs. 29 ± 8% (FS) and 30 ± 8% (WE), both p < 0.05) and led to a significant increase in the measured RCA vessel length to (83 ± 31 mm (LIBRE) vs. 56 ± 12 mm (FS) and 59 ± 27 (WE) p < 0.05).

CONCLUSIONS

Applied to a respiratory self-navigated noncontrast 3D radial whole-heart sequence, LIBRE enables robust large volume fat suppression and significantly improves coronary artery image quality at 3 T compared to the use of conventional FS and WE.