Brain tissue segmentation using improved kernelized rough-fuzzy C-means with spatio-contextual information from MRI Publication date: October 2019 Source: Magnetic Resonance Imaging, Volume 62 Author(s): Anindya Halder, Nur Alom Talukdar Abstract Segmentation of brain tissues from MRI often becomes crucial to properly investigate any region of the brain in order to detect abnormalities. However, the accurate segmentation of the brain tissues is a challenging task as the different tissue regions are usually imprecise, indiscernible, ambiguous, and overlapping. Additionally, different tissue regions are non-linearly separable. Noises and other artifacts may also present in the brain MRI. Therefore, conventional segmentation techniques may not often achieve desired accuracy. To deal those challenges, a robust kernelized rough fuzzy C-means clustering with spatial constraints (KRFCMSC) is proposed in this article for brain tissue segmentation. Here, the brain tissue segmentation from MRI is considered as a clustering of pixels problem. The basic idea behind the proposed technique is the judicious integration of the fuzzy set, rough set, and kernel trick along with spatial constraints (in the form of contextual information) to increase the clustering (segmentation) performance. The use of rough and fuzzy set theory in the clustering process handles the ambiguity, indiscernibility, vagueness and overlappingness of different brain tissue regions. While, the kernel trick increases the chance of linear separability of the complex regions which are otherwise not linearly separable in its original feature space. In order to deal the noisy pixels, here in the clustering process, the spatio-contextual information is introduced from the neighbouring pixels. Experiments are carried out on different real and synthetic benchmark brain MRI datasets (publicly available from Brainweb, and IBSR) without and with added noise. The performance of the proposed method is compared with five other counterpart clustering based segmentation techniques and evaluated using various supervised as well as unsupervised validity indices such as, overall accuracy, precision, recall, kappa, Jaccard, dice, and kernelized Xie-Beni index. Experimental results justify the superiority and robustness of the proposed method over other state-of-the-art methods on both benchmark real life and synthetic brain MRI datasets with and without added noise. Statistical significance of the better segmentation accuracy can be confirmed from the paired t-test results in favour of the proposed method compared to other counterpart methods.

Gadolinium-based contrast agents toxicity in animal studies Publication date: October 2019 Source: Magnetic Resonance Imaging, Volume 62 Author(s): Miski Aghnia Khairinisa, Izuki Amano, Wataru Miyazaki, Noriyuki Koibuchi, Yoshito Tsushima

Significant correlations between human cortical bone mineral density and quantitative susceptibility mapping (QSM) obtained with 3D Cones ultrashort echo time magnetic resonance imaging (UTE-MRI) Publication date: October 2019 Source: Magnetic Resonance Imaging, Volume 62 Author(s): Saeed Jerban, Xing Lu, Hyungseok Jang, Yajun Ma, Behnam Namiranian, Nicole Le, Ying Li, Eric Y. Chang, Jiang Du Abstract Purpose Quantitative susceptibility mapping (QSM) MRI is a tool that can characterize changes in susceptibility, an intrinsic property which is associated with compositional changes in the tissue. Current QSM estimation of cortical bone is challenging because conventional clinical MRI cannot acquire signal in cortical bone. This study aimed to implement Cones 3D ultrashort echo time MRI (UTE-MRI) for ex vivo QSM measurements in human tibial cortical bone, investigating the correlations of QSM with volumetric intracortical bone mineral density (BMD). Materials and methods Nine tibial midshaft cortical bone specimens (25 mm long specimens cut at the mid-point of tibial shaft, 67 ± 20 years old, 5 women and 4 men) were scanned on a clinical 3 T MRI scanner for QSM measurement. The specimens were also scanned on a high-resolution micro-computed tomography (μCT) scanner for volumetric BMD estimation. QSM and μCT results were compared at approximately nine regions of interest (ROIs) per specimen. Results Average 3D UTE-MRI QSM showed significantly strong correlation with volumetric BMD (R = -0.82, P < 0.01) and bone porosity (R = 0.72, P < 0.01). Combining all data points together (77 ROIs), QSM showed significant moderate to strong correlation with volumetric BMD after correction for interdependencies in specimens (R = -0.70, P < 0.01). The corrections were required because the data points were not independent in each specimen. Similarly, the correlation between QSM and porosity was significant (R = 0.68, P < 0.01). Conclusions These results suggest that the Cones 3D UTE-MRI QSM technique can potentially serve as a novel and accurate tool to assess intracortical bone mineral density whilst avoiding ionizing radiation.

Modified look-locker inversion recovery (MOLLI) T1 mapping with inversion group (IG) fitting – A method for improved precision Publication date: October 2019 Source: Magnetic Resonance Imaging, Volume 62 Author(s): Marshall S. Sussman, Bernd J. Wintersperger Abstract MOLLI-based T1 mapping has been applied to a variety of cardiac pathologies. However, conventional MOLLI's requirement for rest periods between inversion groups increases scan time, and limits the choice of inversion groups. The recently developed inversion group (IG) fitting technique eliminates the rest period requirement, and permits complete flexibility of inversion groups. However, a limitation is that its T1 maps have low precision – up to 30% poorer than conventional 3-parameter methods. In the original IG method, T1 maps were derived from the first inversion group only. In the present study, a technique is presented which utilize data from all inversion groups to generate T1 maps. It is hypothesized this “composite-IG” fitting method will provided improved prevision over conventional-IG T1 mapping methods. Simulations, phantom, and in vivo experiments on nine clinical cardiac patients (congenital heart disease, ischemic- and non-ischemic cardiomyopathy) were performed. Imaging was performed on a 1.5 T Siemens scanner. Myocardial T1 mapping precision and reproducibility were calculated for conventional-IG, composite-IG, and 3-parameter techniques. Precision and reproducibility between the techniques was compared using the Wilcoxon Signed Rank test. Statistical significance was set at the 95% confidence level, with the Bonferroni correction for multiple comparisons employed. Composite-IG improves precision by 16–38% over conventional-IG (p < 0.01). Composite-IG T1 maps provided up to 5% better precision than 3-parameter fits (p < 0.01). Composite-IG had better reproducibility than conventional-IG (p < 0.01). However, there was no significant difference between composite-IG and conventional 5(3)3 3-parameter reproducibility.

Accuracy of multi-parametric breast MR imaging for predicting pathological complete response of operable breast cancer prior to neoadjuvant systemic therapy Publication date: October 2019 Source: Magnetic Resonance Imaging, Volume 62 Author(s): Hiroko Tsukada, Jitsuro Tsukada, Simone Schrading, Kevin Strobel, Takahiro Okamoto, Christiane K. Kuhl Abstract Objectives To evaluate whether multiparametric breast-MRI, obtained before the initiation of neoadjuvant systemic therapy (NST) for operable breast cancer, predicts which cancer will achieve a pathological complete response (pCR) after the completion of NST. Methods This was an IRB-approved retrospective study on 31 consecutive patients (median age, 56 years) with operable invasive breast cancer (median size: 22 mm; triple-negative: 11/31 [35%], HER2-positive: 7/31 [23%], triple-positive: 13/31 [42%]) who underwent multiparametric DCE-MRI before the initiation of NST. The MRI protocol consisted of high-resolution dynamic contrast-enhanced MRI (DCE-MRI), T2-TSE, and DWI (b-values 0, 100, 800 s/mm2). The results of surgical pathology after the completion of NST served as a standard of reference. Patient characteristics (age and menopausal status), pathological tumor characteristics (type, stage, nuclear grade, ER/PR and HER2 receptor status, and Ki-67 staining), and MRI characteristics (size, morphology, T2 signal intensity, enhancement kinetics, and ADC values) before NST were evaluated and compared between patients achieving pCR vs. non-pCR. Results Among 31 patients, 17 achieved pCR (55%) and 14 non-pCR (45%). No correlation was observed between patient- or tumor pathology-derived characteristics and pCR vs. non-pCR. Among MRI-derived tumor characteristics, tumor growth orientation parallel to Cooper's ligaments (p = 0.002) and wash-out rates (p = 0.019) correlated with pCR. Pre-NST ADC values were lower in patients achieving pCR (P = 0.086). Conclusions A tumor growth pattern parallel with Cooper's ligaments and a fast wash-out rate on pre-treatment multiparametric MRI are predictive of pCR and more closely associated with pCR than ADC values.

Phase Image Texture Analysis for Motion Detection in Diffusion MRI (PITA-MDD) Publication date: October 2019 Source: Magnetic Resonance Imaging, Volume 62 Author(s): Nahla M.H. Elsaid, Jerry L. Prince, Steven Roys, Rao P. Gullapalli, Jiachen Zhuo Abstract Purpose Pronounced spin phase artifacts appear in diffusion-weighted imaging (DWI) with only minor subject motion. While DWI data corruption is often identified as signal drop out in diffusion-weighted (DW) magnitude images, DW phase images may have higher sensitivity for detecting subtle subject motion. Methods This article describes a novel method to return a metric of subject motion, computed using an image texture analysis of the DW phase image. This Phase Image Texture Analysis for Motion Detection in dMRI (PITA-MDD) method is computationally fast and reliably detects subject motion from diffusion-weighted images. A threshold of the motion metric was identified to remove motion-corrupted slices, and the effect of removing corrupted slices was assessed on the reconstructed FA maps and fiber tracts. Results Using a motion-metric threshold to remove the motion-corrupted slices results in superior fiber tracts and fractional anisotropy maps. When further compared to a state-of-the-art magnitude-based motion correction method, PITA-MDD was able to detect comparable corrupted slices in a more computationally efficient manner. Conclusion In this study, we evaluated the use of DW phase images to detect motion corruption. The proposed method can be a robust and fast alternative for automatic motion detection in the brain with multiple applications to inform prospective motion correction or as real-time feedback for data quality control during scanning, as well as after data is already acquired.

Deep learning reveals untapped information for local white-matter fiber reconstruction in diffusion-weighted MRI Publication date: October 2019 Source: Magnetic Resonance Imaging, Volume 62 Author(s): Vishwesh Nath, Kurt G. Schilling, Prasanna Parvathaneni, Colin B. Hansen, Allison E. Hainline, Yuankai Huo, Justin A. Blaber, Ilwoo Lyu, Vaibhav Janve, Yurui Gao, Iwona Stepniewska, Adam W. Anderson, Bennett A. Landman Abstract Purpose Diffusion-weighted magnetic resonance imaging (DW-MRI) is of critical importance for characterizing in-vivo white matter. Models relating microarchitecture to observed DW-MRI signals as a function of diffusion sensitization are the lens through which DW-MRI data are interpreted. Numerous modern approaches offer opportunities to assess more complex intra-voxel structures. Nevertheless, there remains a substantial gap between intra-voxel estimated structures and ground truth captured by 3-D histology. Methods Herein, we propose a novel data-driven approach to model the non-linear mapping between observed DW-MRI signals and ground truth structures using a sequential deep neural network regression using residual block deep neural network (ResDNN). Training was performed on two 3-D histology datasets of squirrel monkey brains and validated on a third. A second validation was performed using scan-rescan datasets of 12 subjects from Human Connectome Project. The ResDNN was compared with multiple micro-structure reconstruction methods and super resolved-constrained spherical deconvolution (sCSD) in particular as baseline for both the validations. Results Angular correlation coefficient (ACC) is a correlation/similarity measure and can be interpreted as accuracy when compared with a ground truth. The median ACC of ResDNN is 0.82 and median ACC's of different variants of CSD are 0.75, 0.77, 0.79. The mean, median and std. of ResDNN & sCSD ACC across 12 subjects from HCP are 0.74, 0.88, 0.31 and 0.61, 0.71, 0.31 respectively. Conclusion This work highlights the ability of deep learning to capture linkages between ex-vivo ground truth data with feasible MRI sequences. The data-driven approach is applicable to human in-vivo data and results in intriguingly high reproducibility of orientation structure.

Neurite orientation dispersion and density imaging for evaluating the severity of neonatal hypoxic-ischemic encephalopathy in rats Publication date: October 2019 Source: Magnetic Resonance Imaging, Volume 62 Author(s): Akiko Ohki, Shigeyoshi Saito, Junichi Hata, Hirotaka James Okano, Takahiro Higuchi, Kazuki Fukuchi Abstract Purpose To evaluate the utility of neurite orientation dispersion and density imaging (NODDI) for longitudinally assessing neonatal hypoxic-ischemic (HI) encephalopathy severity with 7.0 T magnetic resonance imaging. Methods Thirteen 8-day-old Wistar rats underwent unilateral ligation of the left common carotid artery followed by mild (1 h; n = 6) or severe (2 h; n = 7) hypoxic exposure (8% O2, 34 °C). Diffusion-weighted, T2-weighted (T2W), and flow-sensitive alternating inversion recovery images were obtained with a horizontal 7.0 T scanner at 1, 24, 72, and 168 h after HI insult. The fractional anisotropy (FA), apparent diffusion coefficient (ADC), intracellular volume fraction (ICVF), isotropic volume fraction (ISO), orientation dispersion index (ODI), and cerebral blood flow (CBF) values were calculated for each group (mild and severe) at each time point (1, 24, 72, and 168 h). ICVF, ISO, and ODI were the NODDI parameters. Results Left hemisphere brain damage was identified as slight hyperintensity on T2W images after 1 h in both groups. In the severe group only, the signal hyperintensity increased time-dependently over 168 h. The ADC and CBF were not significantly different between the groups within any region. The ICVF and ODI were significantly higher in the severe vs. mild group at various points between 1 and 168 h (cortex, striatum, or white matter), whereas the FA was significantly higher in the mild vs. severe group at 168 h (cortex and white matter). The ISO was higher in the severe vs. mild group at 72 h (striatum) and 168 h (all regions), while the ISO was significantly higher in the mild vs. severe group at 24 h (all regions). Conclusion Here, ODI, a NODDI metric, identified early differences between mild and severe HI injuries. Our findings support the potential utility of NODDI for determining neonatal HI encephalopathy severity in rats.

Diffusional kurtosis imaging for differentiation of additional suspicious lesions on preoperative breast MRI of patients with known breast cancer Publication date: October 2019 Source: Magnetic Resonance Imaging, Volume 62 Author(s): Vivian Youngjean Park, Sungheon G. Kim, Eun-Kyung Kim, Hee Jung Moon, Jung Hyun Yoon, Min Jung Kim Abstract Purpose To investigate the potential of diffusional kurtosis imaging (DKI) and conventional diffusion-weighted imaging (DWI) in the evaluation of additional suspicious lesions at preoperative breast magnetic resonance imaging (MRI) in patients with breast cancer. Materials and methods Fifty-three additional suspicious lesions in 45 patients with breast cancer, which were detected on preoperative breast MRI, were examined with a 3-T MR system. DKI and DWI data were obtained using a spin-echo single-shot echo-planar imaging sequence with b-values of 0, 50, 600, 1000, and 3000 s/mm2. Histogram parameters (mean, standard deviation, minimum, maximum, 10th, 25th, 50th, 75th, 90th percentiles, kurtosis, skewness and entropy) of ADC from DWI and diffusivity (D), kurtosis (K) from DKI were calculated after postprocessing. Parameters were compared between benign vs. ductal carcinoma in situ (DCIS) vs. invasive breast lesions and diagnostic performances were evaluated by receiver operating characteristic (ROC) analysis. Correlation between the mean values of D and K was analyzed according to lesion type. Results Multiple histogram parameters of D (mean, 25th, 50th percentile, 75th percentile, and entropy) differed between benign and invasive breast lesions (all P < 0.005), but none differed between benign vs. DCIS. D-90th percentile differed between DCIS vs. invasive cancer (P = 0.040). K-10th percentile differed between benign vs. DCIS (P = 0.015). ADC-75th percentile differed between benign vs. invasive cancer and ADC-75th percentile, ADC-90th percentile differed between DCIS vs. invasive cancer, respectively (all P < 0.005). ROC curve analysis showed high specificity for discrimination between benign and invasive cancer. D-mean and K-mean showed strong correlation in benign (rs = −0.813) and invasive lesions (rs = −0.853), but no significant correlation in DCIS. Conclusion DKI may aid in the differentiation of additional suspicious lesions at preoperative breast MRI. Both ADC and DKI may have lower potential in differentiating DCIS from benign lesions.

A “flared-end” gradient coil with outer-wall direct cooling for human brain imaging: A feasibility study Publication date: October 2019 Source: Magnetic Resonance Imaging, Volume 62 Author(s): Zhi Yang, Beihan Zhao, Yong Pei, Bao Yang, Hanbing Lu Abstract Optimal gradient performance is arguably a pre-requisite to realize the full potential of ultrahigh field magnetic resonance imaging (MRI). The values of using tailored gradient coils for brain imaging have been well acknowledged. Unfortunately, conventional head-only gradient coils have two major technical limitations, i.e. limited shoulder clearance and limited cooling capacity. A design, coined “flared-end” gradient coil, combined with a cooling method, named “outer-wall direct cooling”, is proposed to address these problems. The “flared-end” design permits brain access to the center of gradient coil. The “flared end” structure is 3D-printed. It has electrical winding patterns (grooves) on one side and evenly spaced cooling channels on the opposite side. Electrical conductor (copper wire) is fixed into the grooves; coolant is in direct contact with the outer surface of the electrical conductor above each cooling channel, eliminating interfacial thermal resistance between coolant and copper wires. Heat transfer area is thus determined by the size and the number of cooling channels. This approach allows high electric current density for high gradient field strength while maintaining high cooling efficiency. Additionally, the symmetric coil geometry guarantees intrinsic torque balance. As a proof of concept, we have made a gradient coil prototype without active shielding. This coil has an inner diameter of 0.3 m, and is capable of generating 0.337, 0.225 and 0.485 mT/m/A along X, Y and Z, respectively. Active shielding was designed theoretically, but not pursued in the construction of this coil prototype. The new coil geometry and cooling method offer a novel avenue for new gradient coils tailored for human brain imaging at ultrahigh field.