Segmentation of human brain using structural MRI

Segmentation of human brain using structural MRI is a key step of processing in imaging neuroscience. The methods have undergone a rapid development in the past two decades and are now widely available. This non-technical review aims at providing an overview and basic understanding of the most common software. Starting with the basis of structural MRI contrast in brain and imaging protocols, the concepts of voxel-based and surface-based segmentation are discussed. Special emphasis is given to the typical contrast features and morphological constraints of cortical and sub-cortical grey matter. In addition to the use for voxel-based morphometry, basic applications in quantitative MRI, cortical thickness estimations, and atrophy measurements as well as assignment of cortical regions and deep brain nuclei are briefly discussed. Finally, some fields for clinical applications are given.     

Fast three-dimensional sodium imaging of human brain

A three-dimensional sodium imaging technique with a minimum echo time of 0.9 ms is described in a 2.0 Tesla whole-body system. The relaxation behaviour in vivo of sodium was analysed: a lastT ₂ ^* relaxation component between 1.2 and 1.6 ms and a slowT ₂ ^* relaxation component between 7.1 ms and 8.4 ms were quantified in brain tissue of three volunteers. Three-dimensional sodium images of the human brain were acquired in 8.5 min with a resolution of 4.7 × 4.7 × 10 mm (0.2 cc voxel size) and a signal-to-noise ratio of 20 in brain tissue and 30 in cerebrospinal fluid.     

Subjective acceptance of 7 Tesla MRI for human imaging

Objective One prerequisite for transferring ultra-high-field magnetic resonance imaging (MRI) (>3T) into clinical diagnostic workup is a low rate of side effects. To our knowledge, publications of subjective acceptance and willingness to undergo examinations at >3T are rare. We present first results from our research site. Materials and methods Exposure to 7T whole-body MRI of head, extremities, or breast was assessed in 102 subjects. They judged sources of discomfort (examination duration, room temperature) and physiological sensations (vertigo, light flashes) on a 10-point scale, differentiating between examination phases: table stationary or moving. For comparison, the same questionnaire was completed by 43 of these subjects after undergoing a 1.5T examination. Vertigo was the most pronounced sensation at 7T with 5% rating it as very unpleasant (none at 1.5T). This should be compared with the fact that the lengthy exam duration was regarded as even more uncomfortable. Compared to 1.5T, average study duration at 7T was roughly doubled, and 7T elicited a wider range of complaints. Conclusion Although the number of side effects is increased at 7T compared to 1.5T, 7T was well tolerated by the majority of subjects. Further data collection is necessary for better understanding of these effects.     

In vitro imaging of single living human umbilical vein endothelial cells with a clinical 3.0-T MRI scanner

Iron oxide-labelled, single, living human umbilical vein endothelial cells (HUVECs) were imaged over time in vitro using a clinical 3.0-T magnetic resonance (MR) microscopy system. Labelling efficiency, toxicity, cell viability, proliferation and differentiation were assessed using flow cytometry, magnetic cell sorting and a phenanthroline assay. MR images were compared with normal light and fluorescence microscopy. Efficient uptake of iron oxide into HUVECs was shown, although with higher label uptake dose-dependent cytotoxic effects were observed, affecting cell viability. For MR imaging, a T2* weighted three-dimensional protocol was used with in-plane resolution of 39×48μm² and 100-μm slices with a scan time of 13 min. MRI could detect living cells in standard culture dishes at single-cell resolution, although label loss was observed that corresponded with the intracellular iron measurements. MR microscopy using iron oxide labels is a promising tool for studying HUVEC migration and cell biology in vitro and in vivo, but possible toxic effects of label uptake and loss of label over time should be taken into account.     

MDEFT imaging of the human brain at 8 T

T ₁-weighted images of the human brain obtained with the MDEFT sequence at 8 T are presented. These images are characterized by an excellent contrast and good signal to noise ratio. Importantly, results were obtained with adiabatic spin inversion and demonstrate that such pulses can be used event in the ultra high frequency (>300 MHz) range. It is thus possible to obtain high quality results at this field strength without violating SAR guidelines.     

Image artifacts from MR-based attenuation correction in clinical, whole-body PET/MRI

Purpose

Integrated whole-body PET/MRI tomographs have become available. PET/MR imaging has the potential to supplement, or even replace combined PET/CT imaging in selected clinical indications. However, this is true only if methodological pitfalls and image artifacts arising from novel MR-based attenuation correction (MR-AC) are fully understood.

Results

Here we present PET/MR image artifacts following routine MR-AC, as most frequently observed in clinical operations of an integrated whole-body PET/MRI system.

Conclusion

A clinical adoption of integrated PET/MRI should entail the joint image display and interpretation of MR data, MR-based attenuation maps and uncorrected plus attenuation-corrected PET images in order to recognize potential pitfalls from MR-AC and to ensure clinically accurate image interpretation.     

PET/MRI in cancer patients: first experiences and vision from Copenhagen

Combined PET/MRI systems are now commercially available and are expected to change the medical imaging field by providing combined anato-metabolic image information. We believe this will be of particular relevance in imaging of cancer patients. At the Department of Clinical Physiology, Nuclear Medicine & PET at Rigshospitalet in Copenhagen we installed an integrated PET/MRI in December 2011. Here, we describe our first clinical PET/MR cases and discuss some of the areas within oncology where we envision promising future application of integrated PET/MR imaging in clinical routine. Cases described include brain tumors, pediatric oncology as well as lung, abdominal and pelvic cancer. In general the cases show that PET/MRI performs well in all these types of cancer when compared to PET/CT. However, future large-scale clinical studies are needed to establish when to use PET/MRI. We envision that PET/MRI in oncology will prove to become a valuable addition to PET/CT in diagnosing, tailoring and monitoring cancer therapy in selected patient populations.     

Challenges of imaging structure and function with MRI

Deals with current image reconstruction and processing issues in MRI for the study of tissue structure and function. In image reconstruction, the authors discuss the need for new algorithms that can produce high-resolution images with good signal-to-noise ratio from reduced amounts of data. In image processing, the authors describe outstanding problems in automatic image registration and segmentation     

Optically imaging cardiac activation with a laser system

The authors have developed a laser imaging system that: 1) depicts potential distribution with high spatial and temporal resolution; 2) can image large areas; 3) does not require filtering for signal or image enhancement; 4) is flexible, allowing for the acquisition of images of programmable size and shape at varying frame rates; and 5) has a large depth of field, thus minimizing the loss of focus when simultaneously imaging areas of the heart at different distances from the prime focal plane. The high duality of the images obtained with this technique allows quantitative measurements to be made of both the potential distribution and the change in this distribution over time. Here, the authors discuss the components of the system and the present their experimental results     

MR perfusion imaging: correlation with PET and quantitative angiography

4. Conclusions The presented MR approach reliably identifies patients with anatomically and hemodynamically signiticant coronary artery stenoses. This is due to the fact, that the pulse sequence used produces a substantial change in signal intensity in the perfused versus poorly perfused myocardial regions. Analysis of upslope in this setting rather than of other parameters provides a very sensitive and specific measure of myocardial ischemia. As upslope is a semiquantitative measure of absolute perfusion, even patients with triple vessel disease may be evaluated using this method. This is not the case when using conventional nuclear techniques. Furthermore, the spatial resolution of the MR images permits one to resolve the subendocardial layers of the myocardium, which thus can be evaluated separately from the entire wall. Again, this is not possible using nuclear cardiology perfusion imaging. The robustness of this MR perfusion imaging approach and the fact, that most of the heart can be covered may qualify for its clinical application in the management of coronary artery disease.     

Approaches for the optimization of MR protocols in clinical hybrid PET/MRI studies

Magnetic resonance imaging (MRI) is the examination method of choice for the diagnosis of a variety of diseases. MRI allows us to obtain not only anatomical information but also identification of physiological and functional parameters such as networks in the brain and tumor cellularity, which plays an increasing role in oncologic imaging, as well as blood flow and tissue perfusion. However, in many cases such as in epilepsy, degenerative neurological diseases and oncological processes, additional metabolic and molecular information obtained by PET can provide essential complementary information for better diagnosis. The combined information obtained from MRI and PET acquired in a single imaging session allows a more accurate localization of pathological findings and better assessment of the underlying physiopathology, thus providing a more powerful diagnostic tool. Two hundred and twenty-one patients were scanned from April 2011 to January 2012 on a Philips Ingenuity TF PET/MRI system. The purpose of this review article is to provide an overview of the techniques used for the optimization of different protocols performed in our hospital by specialists in the following fields: neuroradiology, head and neck, breast, and prostate imaging. This paper also discusses the different problems encountered, such as the length of studies, motion artifacts, and accuracy of image fusion including physical and technical aspects, and the proposed solutions.     

Feasibility of diffusion-weighted imaging with DWIBS in staging Hodgkin lymphoma in pediatric patients: comparison with PET/CT

Objective

The aim of the study was to evaluate feasibility of diffusion-weighted whole-body imaging with background body signal suppression (DWIBS) method in diagnosing Hodgkin lymphoma in pediatric patients and to compare it with 18F-FDG PET/CT as a gold standard.

Materials and methods

Eleven patients (median age 14) with newly diagnosed Hodgkin lymphoma were examined with 18F-FDG PET/CT and MRI including whole-body DWIBS sequence (b = 0, 800 s/mm²), before the oncologic treatment. About 26 locations of lymphatic tissues were evaluated visually and quantitatively using ADC_mean (DWIBS) and SUV_max (18F-FDG PET/CT), respectively.

Results

All affected lymph node regions (n = 134) diagnosed in 18F-FDG PET/CT were found with DWIBS, presenting decreased diffusion. Significant correlation was found between ADC and SUV values (R² = − 0.37; p = 0.0001). Nevertheless, additional 33 regions were recognized only by DWIBS. They were significantly smaller than regions diagnosed by both methods.

Discussion

Agreement between DWIBS and 18F-FDG PET/CT for detection and staging of malignant lymphoma is high. DWIBS can be used for the evaluation of pediatric Hodgkin lymphoma.

     

Optical imaging of the intact human brain [Guest Editorial]

This special issue includes a collection of articles on the use of optical devices to measure brain activity in humans in a noninvasive manner and is intended to provide a sample of different aspects of this area of research.     

High-resolution imaging of the excised porcine heart at a whole-body 7 T MRI system using an 8Tx/16Rx pTx coil

Introduction

MRI of excised hearts at ultra-high field strengths (\({\mathrm{B}}_{0}\)≥7 T) can provide high-resolution, high-fidelity ground truth data for biomedical studies, imaging science, and artificial intelligence. In this study, we demonstrate the capabilities of a custom-built, multiple-element transceiver array customized for high-resolution imaging of excised hearts.

Method

A dedicated 16-element transceiver loop array was implemented for operation in parallel transmit (pTx) mode (8Tx/16Rx) of a clinical whole-body 7 T MRI system. The initial adjustment of the array was performed using full-wave 3D-electromagnetic simulation with subsequent final fine-tuning on the bench.

Results

We report the results of testing the implemented array in tissue-mimicking liquid phantoms and excised porcine hearts. The array demonstrated high efficiency of parallel transmits characteristics enabling efficient pTX-based B₁⁺-shimming.

Conclusion

The receive sensitivity and parallel imaging capability of the dedicated coil were superior to that of a commercial 1Tx/32Rx head coil in both SNR and T₂*-mapping. The array was successfully tested to acquire ultra-high-resolution (0.1 × 0.1 × 0.8 mm voxel) images of post-infarction scar tissue. High-resolution (isotropic 1.6 mm³ voxel) diffusion tensor imaging-based tractography provided high-resolution information about normal myocardial fiber orientation.

     

Characterization of the impact to PET quantification and image quality of an anterior array surface coil for PET/MR imaging

Object

The aim of this study was to determine the impact to PET quantification, image quality and possible diagnostic impact of an anterior surface array used in a combined PET/MR imaging system.

Materials and methods

An extended oval phantom and 15 whole-body FDG PET/CT subjects were re-imaged for one bed position following placement of an anterior array coil at a clinically realistic position. The CT scan, used for PET attenuation correction, did not include the coil. Comparison, including liver SUV_mean, was performed between the coil present and absent images using two methods of PET reconstruction. Due to the time delay between PET scans, a model was used to account for average physiologic time change of SUV.

Results

On phantom data, neglecting the coil caused a mean bias of ?8.2 % for non-TOF/PSF reconstruction, and ?7.3 % with TOF/PSF. On clinical data, the liver SUV neglecting the coil presence fell by ?6.1 % (±6.5 %) for non-TOF/PSF reconstruction; respectively ?5.2 % (±5.3 %) with TOF/PSF. All FDG-avid features seen with TOF/PSF were also seen with non-TOF/PSF reconstruction.

Conclusion

Neglecting coil attenuation for this anterior array coil results in a small but significant reduction in liver SUV_mean but was not found to change the clinical interpretation of the PET images.     

Study of kinetics of 19F-MRI using a fluorinated imaging agent (19FIT) on a 3T clinical MRI system

Magnetic Resonance Materials in Physics, Biology and Medicine - To use 19F imaging tracer (19FIT-27) to evaluate kinetics in major organs. Kinetics studies using proton MRI are difficult because of...     

Sodium (23Na) ultra-short echo time imaging in the human brain using a 3D-Cones trajectory

Object

Sodium magnetic resonance imaging (²³Na-MRI) of the brain has shown changes in ²³Na signal as a hallmark of various neurological diseases such as stroke, Alzheimer’s disease, Multiple Sclerosis and Huntington’s disease. To improve scan times and image quality, we have implemented the 3D-Cones (CN) sequence for in vivo ²³Na brain MRI.

Materials and methods

Using signal-to-noise (SNR) as a measurement of sequence performance, CN is compared against more established 3D-radial k-space sampling schemes featuring cylindrical stack-of-stars (SOS) and 3D-spokes kooshball (KB) trajectories, on five healthy volunteers in a clinical setting. Resolution was evaluated by simulating the point-spread-functions (PSFs) and experimental measures on a phantom.

Results

All sequences were shown to have a similar SNR arbitrary units (AU) of 6–6.5 in brain white matter, 7–9 in gray matter and 17–18 AU in cerebrospinal fluid. SNR between white and gray matter were significantly different for KB and CN (p = 0.046 and <0.001 respectively), but not for SOS (p = 0.1). Group mean standard deviations were significantly smaller for CN (p = 0.016). Theoretical full-width at half-maximum linewidth of the PSF for CN is broadened by only 0.1, compared to 0.3 and 0.8 pixels for SOS and KB respectively. Actual image resolution is estimated as 8, 9 and 6.3 mm for SOS, KB and CN respectively.

Conclusion

The CN sequence provides stronger tissue contrast than both SOS and KB, with more reproducible SNR measurements compared to KB. For CN, a higher true resolution in the same amount of time with no significant trade-off in SNR is achieved. CN is therefore more suitable for ²³Na-MRI in the brain.