Ultrafast Dynamic Contrast-Enhanced MRI of the Breast: From Theory to Practice.

The development of ultrafast dynamic contrast-enhanced (UF-DCE) MRI has occurred in tandem with fast MRI scan techniques, particularly view-sharing and compressed sensing. Understanding the strengths of each technique and optimizing the relevant parameters are essential to their implementation. UF-DCE MRI has now shifted from research protocols to becoming a part of clinical scan protocols for breast cancer. UF-DCE MRI is expected to compensate for the low specificity of abbreviated MRI by adding kinetic information from the upslope of the time-intensity curve. Because kinetic information from UF-DCE MRI is obtained from the shape and timing of the initial upslope, various new kinetic parameters have been proposed. These parameters may be associated with receptor status or prognostic markers for breast cancer. In addition to the diagnosis of malignant lesions, more emphasis has been placed on predicting and evaluating treatment response because hyper-vascularity is linked to the aggressiveness of breast cancers. In clinical practice, it is important to note that breast lesion images obtained from UF-DCE MRI are slightly different from those obtained by conventional DCE MRI in terms of morphology. A major benefit of using UF-DCE MRI is avoidance of the marked or moderate background parenchymal enhancement (BPE) that can obscure the target enhancing lesions. BPE is less prominent in the earlier phases of UF-DCE MRI, which offers better lesion-to-noise contrast. The excellent contrast of early-enhancing vessels provides a key to understanding the detailed pathological structure of tumor-associated vessels. UF-DCE MRI is normally accompanied by a large volume of image data for which automated/artificial intelligence-based processing is expected to be useful. In this review, both the theoretical and practical aspects of UF-DCE MRI are summarized. EVIDENCE LEVEL: 5 TECHNICAL EFFICACY: Stage 2.

[Comparison of Image Quality Using Metal Artifact Reduction Magnetic Resonance Sequences at Different Static Magnetic Fields].

PURPOSE: This study aims to optimize the selection of the magnetic resonance imaging (MRI) scanners and metal artifact reduction magnetic resonance sequence (MARS) in patients with metallic implants by comparing the image quality of MARS at different static magnetic fields. METHOD: A titanium alloy hip prostheses stem was covered with the pork phantom. Nifedipine 10 mg was placed as a simulated lesion near the hip joint of the phantom. T1-weighted imaging (T1WI) and short tau inversion recovery (STIR) were acquired at both 1.5T and 3T. Several techniques of high-bandwidth (High BW), view angle tilting (VAT), and compressed sensing and slice encoding for metal artifacts correction (CS-SEMAC) were compared. Artifacts, sharpness, and visibility of lesions were evaluated visually by five radiological technologists using the normalized-rank approach. RESULT: CS-SEMAC reduced metal artifacts but showed bad sharpness. CS-SEMAC at 3T was the best visibility of lesions. CONCLUSION: If lesion visibility is a priority, CS-SEMAC with 3T is recommended as the first choice.

Evaluation of apparent diffusion coefficient of two-dimensional BLADE turbo gradient- and spin-echo diffusion-weighted imaging with a breast phantom.

This study aimed to evaluate the reliability of apparent diffusion coefficient (ADC) values generated with two-dimensional turbo gradient- and spin-echo with BLADE trajectory diffusion-weighted imaging (TGSE-BLADE-DWI) sequence using a breast diffusion phantom. TGSE-BLADE-DWI and single-shot spin-echo echo-planar imaging (SS-EPI-DWI) were performed using a 3.0 T magnetic resonance imaging scanner. Concordance rates of ADC values and the signal-to-noise ratio (SNR) were compared between TGSE-BLADE-DWI and SS-EPI-DWI. TGSE-BLADE-DWI provided a higher concordance rate for ADC values than SS-EPI-DWI when b-values > 2000s/mm2 and a slice thickness of 1 mm were used. TGSE-BLADE-DWI showed less image distortion than SS-EPI-DWI. The SNR of TGSE-BLADE-DWI was higher than that of SS-EPI-DWI, except at a number of excitations of 7 and a slice thickness of 1 mm. In conclusion, TGSE-BLADE-DWI can offer a better SNR, less distortion, and more reliable ADC measurements than SS-EPI-DWI in a breast phantom.

Comparison of TGSE-BLADE DWI, RESOLVE DWI, and SS-EPI DWI in healthy volunteers and patients after cerebral aneurysm clipping.

Diffusion-weighted magnetic resonance imaging is prone to have susceptibility artifacts in an inhomogeneous magnetic field. We compared distortion and artifacts among three diffusion acquisition techniques (single-shot echo-planar imaging [SS-EPI DWI], readout-segmented EPI [RESOLVE DWI], and 2D turbo gradient- and spin-echo diffusion-weighted imaging with non-Cartesian BLADE trajectory [TGSE-BLADE DWI]) in healthy volunteers and in patients with a cerebral aneurysm clip. Seventeen healthy volunteers and 20 patients who had undergone surgical cerebral aneurysm clipping were prospectively enrolled. SS-EPI DWI, RESOLVE DWI, and TGSE-BLADE DWI of the brain were performed using 3 T scanners. Distortion was the least in TGSE-BLADE DWI, and lower in RESOLVE DWI than SS-EPI DWI near air-bone interfaces in healthy volunteers (P < 0.001). Length of clip-induced artifact and distortion near the metal clip were the least in TGSE-BLADE DWI, and lower in RESOLVE DWI than SS-EPI DWI (P < 0.01). Image quality scores for geometric distortion, susceptibility artifacts, and overall image quality in both healthy volunteers and patients were the best in TGSE-BLADE DWI, and better in RESOLVE DWI than SS-EPI DWI (P < 0.001). Among the three DWI sequences, image quality was the best in TGSE-BLADE DWI in terms of distortion and artifacts, in both healthy volunteers and patients with an aneurysm clip.

Diagnostic advantage of thin slice 2D MRI and multiplanar reconstruction of the knee joint using deep learning based denoising approach.

The purpose of this study is to evaluate whether thin-slice high-resolution 2D fat-suppressed proton density-weighted image of the knee joint using denoising approach with deep learning-based reconstruction (dDLR) with MPR is more useful than 3D FS-PD multi planar voxel image. Twelve patients who underwent MRI of the knee at 3T and 13 knees were enrolled. Denoising effect was quantitatively evaluated by comparing the coefficient of variation (CV) before and after dDLR. For the qualitative assessment, two radiologists evaluated image quality, artifacts, anatomical structures, and abnormal findings using a 5-point Likert scale between 2D and 3D. All of them were statistically analyzed. Gwet's agreement coefficients were also calculated. For the scores of abnormal findings, we calculated the percentages of the cases with agreement with high confidence. The CV after dDLR was significantly lower than the one before dDLR (p < 0.05). As for image quality, artifacts and anatomical structure, no significant differences were found except for flow artifact (p < 0.05). The agreement was significantly higher in 2D than in 3D in abnormal findings (p < 0.05). In abnormal findings, the percentage with high confidence was higher in 2D than in 3D (p < 0.05). By applying dDLR to 2D, almost equivalent image quality to 3D could be obtained. Furthermore, abnormal findings could be depicted with greater confidence and consistency, indicating that 2D with dDLR can be a promising imaging method for the knee joint disease evaluation.

Evaluation of motion artifacts in brain magnetic resonance images using convolutional neural network-based prediction of full-reference image quality assessment metrics.

Purpose: Motion artifacts in magnetic resonance (MR) images mostly undergo subjective evaluation, which is poorly reproducible, time consuming, and costly. Recently, full-reference image quality assessment (FR-IQA) metrics, such as structural similarity (SSIM), have been used, but they require a reference image and hence cannot be used to evaluate clinical images. We developed a convolutional neural network (CNN) model to quantify motion artifacts without using reference images. Approach: The brain MR images were obtained from an open dataset. The motion-corrupted images were generated retrospectively, and the peak signal-to-noise ratio, cross-correlation coefficient, and SSIM were calculated. The CNN was trained using these images and their FR-IQA metrics to predict the FR-IQA metrics without reference images. Receiver operating characteristic (ROC) curves were created for binary classification, with artifact scores < 4 indicating the need for rescanning. ROC curve analysis was performed on the binary classification of the real motion images. Results: The predicted FR-IQA metric having the highest correlation with the subjective evaluation was SSIM, which was able to classify images requiring rescanning with a sensitivity of 89.5%, specificity of 78.2%, and area under the ROC curve (AUC) of 0.930. The real motion artifacts were classified with the AUC of 0.928. Conclusions: Our CNN model predicts FR-IQA metrics with high accuracy, which enables quantitative assessment of motion artifacts in MR images without reference images. It enables classification of images requiring rescanning with a high AUC, which can improve the workflow of MR imaging examinations.

Differences in apparent diffusion coefficients between normal brain echo-planar images and turbo spin-echo diffusion-weighted images with distortion correction.

PURPOSE: We performed echo-planar imaging (EPI) and turbo spin-echo (TSE) diffusion-weighted imaging (DWI) using magnetic resonance imaging (MRI) to obtain basic clinical data of the apparent diffusion coefficient (ADC) in various parts of normal brains and compared the datasets using our retrospective distortion correction technique. MATERIALS AND METHODS: The normal brains of 32 patients who underwent health check were scanned on a 1.5-T MRI instrument using EPI- and TSE-DWI. Distortion was corrected by (1) segmentation: the b0 images were segmented based on the plural threshold values; (2) edge detection: the edge was detected in the images obtained in step (1); (3) non-rigid image registration: non-rigid image registration using Demons algorithm was achieved between the b0 images of EPI-DWI and TSE-DWI, thereby, creating a displacement field; and (4) image warp: the displacement field was applied to the b1000 image to warp. Twenty-six parts of the brain were measured from the images of b0 and b1000 and the ADCs were calculated. The signal-to-noise ratio (SNR) of the cerebrospinal fluid was measured to identify the cause of the difference between the two sequences. These were compared using Wilcoxon signed-rank test (P = 0.05). RESULTS: The ADC was significantly higher measured by EPI-DWI than by TSE-DWI. The SNR of EPI-DWI was significantly higher than that of the TSE-DWI. CONCLUSION: Care must be taken when measuring ADCs near the base of the skull, such as the brain stem, where the SNR of the imaging technique is likely to decrease or distort.

Cardiovascular magnetic resonance virtual tagging with B-spline-based free-form deformation.

PURPOSE: We developed a virtual tagging technique that reconstructs tagging images using the displacement field obtained by applying B-spline free-form deformation (FFD) between diastolic images and images of other cardiac phases in cardiac cine MRI. The purpose of this study was to validate its characteristics and usefulness in phantom and patient studies. METHODS: Digital phantoms simulating uniform and non-uniform wall motion models were created, and virtual tagging images were reconstructed with various matrix sizes and tag resolutions to evaluate the accuracy of FFD and the characteristics of the tags. In the patient study, FFD's accuracy was assessed at three levels (base, middle, and apex) in healthy patients. In patients with heart failure, virtual tagging images were compared with strain maps obtained by feature tracking and virtual tagging. RESULTS: In the phantom study, blurring of tags was observed when tags were reconstructed with high resolution using a small matrix size. In the patient study, the accuracy of FFD was lower in the base than in the apex. Patients with heart failure had decreased distortion of the displacement field vector and virtual tags, indicating decreased local wall motion, consistent with areas of abnormalities found in strain maps. CONCLUSION: The virtual tagging technique does not require additional imaging and can visualize regional LV motion abnormalities via deformation of the tag as well as conventional cardiovascular magnetic resonance tagging.

Deep Learning-based Noise Reduction for Fast Volume Diffusion Tensor Imaging: Assessing the Noise Reduction Effect and Reliability of Diffusion Metrics.

To assess the feasibility of a denoising approach with deep learning-based reconstruction (dDLR) for fast volume simultaneous multi-slice diffusion tensor imaging of the brain, noise reduction effects and the reliability of diffusion metrics were evaluated with 20 patients. Image noise was significantly decreased with dDLR. Although fractional anisotropy (FA) of deep gray matter was overestimated when the number of image acquisitions was one (NAQ1), FA in NAQ1 with dDLR became closer to that in NAQ5.

Diffusion-weighted breast magnetic resonance imaging with distortion correction using non-rigid image registration: a clinical study.

Our distortion correction method for diffusion-weighted breast images was applied to clinical studies to evaluate its practicality. We used b values of 0 and 1500 s/mm2 and dynamic images captured after injecting the contrast medium for magnetic resonance imaging (MRI). This method is based on non-rigid image registration, and dispenses with multiple scans, reducing examination time and patient burden. Image correction included the following: matrix size matching; segmentation; edge detection; non-rigid image registration between T2-weighted images with fat suppression and b0 images; and image warping of b1500 images. Significant differences were noted for all combinations of cross-correlation coefficients of whole-breast images between with and without correction, and no significant differences for apparent diffusion coefficients. The percentage differences were not significant for cross-correlation coefficients, with and without correction, calculated for various sizes, positions, and types of lesions of high-intensity regions. In conclusion, our distortion correction method is effective for clinical breast MRI.

Impact of the Number of Iterations in Compressed Sensing Reconstruction on Ultrafast Dynamic Contrast-enhanced Breast MR Imaging.

PURPOSE: To assess the impact of the number of iterations of compressed sensing (CS) reconstruction on the kinetic parameters and image quality in dynamic contrast-enhanced (DCE)-MRI of the breast, with prospectively undersampled CS-accelerated scans. MATERIALS AND METHODS: Breast examinations including ultrafast DCE-MRI using CS were conducted for 21 patients. Images were reconstructed with different numbers of iterations. The peak enhancement ratio of the aorta and wash-in slope, initial area under the curve, and Ktrans of the breast lesions were measured. The root mean square error and structural similarity between the images using 50 iterations and images with a lower number of iterations were evaluated as criterion for quantitative image evaluation. RESULTS: Using an insufficient number of iterations, the contrast-enhanced effect was highly underestimated. In all semi-quantitative parameters, the number of iterations that stabilized the parameters in malignant lesions was higher than that in benign lesions. At least 15 iterations were needed for semi-quantitative parameters. For Ktrans, there were no significant differences between 10 and 50 iterations in both malignant and benign lesions. CONCLUSION: The kinetic parameters using ultrafast DCE-MRI with CS are affected by the number of iterations, especially in malignant lesions. However, if the images are reconstructed with an adequate number of iterations, ultrafast DCE-MRI with CS can be a powerful technique having high temporal and spatial resolution.

Novel distortion correction method for diffusion-weighted imaging based on non-rigid image registration between low b value image and anatomical image.

PURPOSE: This study aimed to develop a novel technique for retrospective distortion correction based on non-rigid image registration in magnetic resonance diffusion image. METHODS: A 3.0 T MRI scanner with an 18-channel dedicated breast coil and the outer shell of the original breast phantom, which provided images with non-uniform fat-suppression based on clinical data were used. The diffusion-weighted imaging with and without parallel imaging (PI) was used. The proposed study included several steps, which are FOV size matching, matrix size matching, image segmentation, edge detection, non-rigid image registration, and image wrap. We compared the results obtained using the proposed method with that obtained using TOPUP images. The correlation was assessed between T1-weighted image with fat suppression (FS-T1WI) and b1000 image with the help of cross-correlation coefficient (CCC). Shape-error analysis of tumor model and apparent diffusion coefficient (ADC) was calculated. The Steel-Dwass multiple-comparison tests were used for all comparisons and statistical analysis (P < 0.05). RESULTS: The novel method of CCC showed the highest correlation between FS-T1WI and b1000 images. In the Steel-Dwass multiple-comparison test, significant differences were found (P < 0.05) except between non-correction and TOPUP (P = 0.99). The novel method was the lowest degree of error. With PI in the right breast, no significant differences, whereas in the left breast, significant differences were observed except for between novel method and TOPUP (P = 0.73). Without PI in the right breast, significant differences were observed. In the left breast, no significant differences were observed between any combinations. The ADC value, no significant differences were observed for non-correction and novel methods. CONCLUSIONS: We developed a novel technique for retrospective distortion correction based on non-rigid image registration. The high degree of accuracy of this method combined with the lack of requirement for additional scans renders it a promising tool for application in clinical practice.

Ultrafast dynamic contrast-enhanced mri of the breast using compressed sensing: breast cancer diagnosis based on separate visualization of breast arteries and veins.

PURPOSE: To evaluate the feasibility of ultrafast dynamic contrast-enhanced (UF-DCE) magnetic resonance imaging (MRI) with compressed sensing (CS) for the separate identification of breast arteries/veins and perform temporal evaluations of breast arteries and veins with a focus on the association with ipsilateral cancers. MATERIALS AND METHODS: Our Institutional Review Board approved this study with retrospective design. Twenty-five female patients who underwent UF-DCE MRI at 3T were included. UF-DCE MRI consisting of 20 continuous frames was acquired using a prototype 3D gradient-echo volumetric interpolated breath-hold sequence including a CS reconstruction: temporal resolution, 3.65 sec/frame; spatial resolution, 0.9 × 1.3 × 2.5 mm. Two readers analyzed 19 maximum intensity projection images reconstructed from subtracted images, separately identified breast arteries/veins and the earliest frame in which they were respectively visualized, and calculated the time interval between arterial and venous visualization (A-V interval) for each breast. RESULTS: In total, 49 breasts including 31 lesions (breast cancer, 16; benign lesion, 15) were identified. In 39 of the 49 breasts (breasts with cancers, 16; breasts with benign lesions, 10; breasts with no lesions, 13), both breast arteries and veins were separately identified. The A-V intervals for breasts with cancers were significantly shorter than those for breasts with benign lesions (P = 0.043) and no lesions (P = 0.007). CONCLUSION: UF-DCE MRI using CS enables the separate identification of breast arteries/veins. Temporal evaluations calculating the time interval between arterial and venous visualization might be helpful in the differentiation of ipsilateral breast cancers from benign lesions. LEVEL OF EVIDENCE: 3 Technical Efficacy: Stage 1 J. Magn. Reson. Imaging 2018;47:97-104.