Alzheimer's disease has become more prevalent in our lives as the population ages. Accurate diagnosis and positive intervention can effectively delay the progress of early-stage Alzheimer’s disease. Accurate diagnosis of Alzheimer’s disease requires the combination of information from multiple regions of interest (ROIs), because the use of one single ROI may lose the connection and impact among multiple brain regions. In this paper, we firstly proposed a three-input convolutional neural network (CNN) to comprehensively utilize the information from three ROIs, hippocampus, other gray matter (without hippocampus) and white matter. In addition, as the neural network deepens, important feature information of original image will be partially lost. Therefore, we proposed a multi-output 3D CNN, which increases the connection and output of middle layers, shortens the distance between input and output, enhances feature propagation and reduces the loss of feature information. The results showed that the accuracy rate, precision rate, sensitivity, specificity and F1-score of the test set diagnosis obtained by multi-output 3D CNN model were 90.5%, 91.0%, 90.4%, 95.2% and 90.5%, respectively. The diagnostic performance was better than that of the single-output 3D CNN model.

Segmentation of tooth and alveolar bone from the cone beam computed tomography (CBCT) images provides the basic data for the three-dimensional reconstruction and visualization of bone structure. In this paper, according to the characteristics of tooth and alveolar bone, an improved potential well function was combined with the level set model for segmentation of tooth and alveolar bone, overcoming the defects of 'stop evolution' or 'too fast evolution' that might occur with the use of conventional potential well functions. Since it is difficult to effectively filter out the noises in CBCT image with the single variance Gaussian filter, a multiple small variance Gaussian filter stack was used to preprocess the image. As the contours of the same tooth in adjacent images of the image sequence showed only little changes, the segmentation result of the current layer was taken as the initial contour of the curve evolution for the next layer to reduce the times of iteration and increase the speed of segmentation. In addition, the algorithm is also used to segment a single tooth in magnetic resonance image of oral cavity successfully.

Magnetic resonance imaging (MRI) studies have documented that there were structural abnormalities in acquired immune deficiency syndrome (AIDS) patients. Although the long-term survival rate of AIDS patients can be improved by the combined antiretroviral therapy (cART), little is known about the effect of cART on brain structural alterations in AIDS patients. In this study, three-dimensional high-resolution brain structure MRI and voxel-based morphometric methods were used to explore the effects of cART on brain gray matter volume in AIDS patients. The results indicate that cART may effectively mitigate the occurrence of abnormal brain structure in AIDS patients, emphasizing the importance of early initiation of cART.

In this paper, the use of software radio technology for signal reception in magnetic resonance imaging (MRI) was attempted. A radio frequency direct band-pass sampling method was proposed for signal reception in MRI. For this method, a digital demodulation algorithm was designed and implemented with System Generator, a design and development tool of digital signal processing (DSP) developed by the Xilinx company. At the same time, a digital down conversion (DDC) system was designed and tested to flexibly implement the DDC function. The validity of the proposed receiving method was confirmed by numerical simulations and experiments.

Fast magnetic resonance imaging (MRI) has been attracting more and more research interests in recent years. With the emergence of big data and development of advanced deep learning algorithms, neural network has become a common and effective tool for image reconstruction in fast MRI. One main challenge to the deep learning-based methods for fast MRI reconstruction is the trade-off between the network performance and the network capacity. Few previous studies have used the deep learning-based methods in parallel imaging. In this work, a deep recursive cascaded convolutional network (DRCCN) architecture was designed for parallel MRI, with reduced number of network parameters while maintaining a satisfactory performance. The experimental results demonstrated that, compared to the classical methods, image reconstruction with the well-trained DRCCN networks were more accurate and less time consuming.

Diffusion tensor imaging (DTI) was used to evaluate micro-structural changes and structural connectivity in the TX mice model of Wilson's disease (WD) at 9.4 T. Region of interest (ROI) analysis showed that mean fractional anisotropy (FA) values in the hippocampus, caudate putamen and lateral globus pallidus of the TX mice decreased significantly, relative to those in the control mice. The mean diffusivity (MD) values in these regions showed no significant inter-group differences. The results of fiber tracking demonstrated that structural connectivity of in the brain of TX mice was maintained, indicating that the effects of copper accumulation in these mice were mainly regional.

Right ventricle (RV) segmentation is essential for assessing cardiac function in patients with pulmonary hypertension, tetralogy of Fallot, and so on. However, it remains difficult due to the complex structure of heart, thin myocardium and large variability of the RV, as well as the interferences from the fat nearby. Due to its high temporal and spatial resolution, cardiac cine magnetic resonance imaging is widely used for functional evaluation of the heart. This article reviews the commonly-used methods for RV segmentation from the cardiac cine magnetic resonance images. The traditional algorithms are described first, followed by the novel multi-atlas and deep learning methods. Lastly, the evaluation standards for RV segmentation are introduced. It is concluded that the deep learning-based segmentation methods may have the potential to become the method of choice in clinical settings due to their high efficiency and accuracy in the diagnosis and prognosis of the heart-related diseases.

In this study, an olfactory inhalation instrument was devised to allow the rats to inhale volatile citral or odorless air. Manganese-enhanced magnetic resonance imaging (MEMRI) was used to reveal changes of accumulative brain activities after a 24 h inhalation of volatile citral. Compared with the control group, the rats in the citral group showed increased functional activities in the core of nucleus accumbens (AcbC) and olfactory glomerular layer (GL), and decreased Mn²⁺ accumulation in the brain regions of visual cortex (VC), auditory cortex (AC) and retrosplenial cortex (RSC). Functional correlations between the GL and associated brain regions increased after citral inhalation. These results suggested that MEMRI might be used to detect brain activation associated with sustained olfactory stimulation.

High-field clinical magnetic resonance imaging (MRI) system often requires multiple-channel (i.e., 16 or 32) receivers that are capable of transmitting data at a high speed, especially during fast imaging. In this work, a data transmission module based on PowerPC processor is developed and integrated in a custom-built MRI spectrometer for high-speed data transmission between the spectrometer and the host computer. The module uses the Freescale's high-performance PowerPC processor MPC8270 as its core, and runs on the embedded Linux operating system. The processor and the computer are connected through a 100 M Ethernet network, while the sequence running the data acquisition modules (quantity scalability) is connected to the processor using a local bus. Responding to the interruption requests from the data acquisition module, the processor reads and uploads the data quickly. The speed and reliability of the response are guaranteed through the design of the driver program. The results of imaging experiments demonstrate that the design proposed can meet the high-speed data transmission requirements of multiple receiving channels.

Segmentation of prostate magnetic resonance images is of great significance in the interventional diagnosis and treatment of prostate diseases. In this work, the conventional distance regularized level set evolution (DRLSE) model is improved and applied to prostate segmentation. In magnetic resonance image, the prostate boundary near the bladder is often blurred, while that near the urethra is clear, resulting in a poor performance for the traditional gradient information indicator function. In this study, two indicator functions were used to control the evolution of boundary in the clear segment and blurred segment, respectively, to achieve better segmentation. In addition, an energy check term was added to the external energy function to prevent evolution from stopping at a false boundary. This modification could drive the level set to move to regions with large gray level fluctuation and stop evolution at a blurred boundary. Experimental results demonstrated that the performance of prostate segmentation was satisfactory, judging from the Dice similarity coefficient (DSC) which reached an average of 96%.

This paper proposes an image reconstruction method for simultaneous multi-slice imaging (SMS) by combining the virtual conjugate coil (VCC) technology and robust artificial-neural-networks for k-space interpolation (RAKI). This method can effectively improve the reconstruction quality, and is named VIRGINIA (VIRtual conjuGate coIls Neural-networks InterpolAtion). VIRGINIA utilizes the complex conjugate symmetry property of the virtual coil concept to generate virtual coil data for training, and obtains better image quality by applying the trained network to the original aliased SMS data. With experimental data, the VIRGINIA method was compared to other reconstruction methods (i.e., RAKI only and slice-GRAPPA) in terms of quantitative indices such as structural similarity index (SSIM), peak signal-to-noise ratio (PSNR) and root mean square error (RMSE). The results demonstrated that, under some certain slice-acceleration factors, VIRGINIA produced better reconstruction quality than those obtainable by Slice-GRAPPA and RAKI.

Transcranial direct current stimulation (tDCS) noninvasively modulates the excitability of the brain, and is a promising method for treating diseases in the central nervous system. tDCS is safe, effective and affordable, and has been increasingly applied in neuroscience research. Functional magnetic resonance imaging (fMRI) is a widely-used tool to evaluate the therapeutic effects and the underlying neural mechanisms of tDCS. Herein we briefly reviewed the progresses of applications of fMRI in tDCS research. Both clinical trials and animal model experiments were discussed, as well as the perspectives of future applications.

Magnetic resonance hyperthermia is a new nano-medical therapy that emerges in recent years. In the presence of external alternating magnetic fields produced by MR instruments, magnetic nanoparticles accumulated at the tumor site can generate heat through Neel relaxation and/or Brownian relaxation. Through magnetic resonance hyperthermia, magnetic nanoparticles can serve as “molecular bullets” to kill cancer cells, leaving surrounding healthy tissues unaffected. Such hyperthermic effects can also be used for thermal activation and control releasing of cancer drugs. One major challenge of magnetic resonance hyperthermia is to optimize the heating efficiency of magnetic nanoparticle suspension. Heating efficiency depends on the size, physical properties, and aggregation state of magnetic nanoparticles. In this study, the thermodynamic behavior of magnetic nanoparticles and the aggregation/disruption of monomers/clusters under different temperatures were studied by 3D Metropolis Monte Carlo method. The relationship between the critical temperature for aggregation/disruption and the frequency of external magnetic field has been established through revised Langevin function.
Simulation results show that the relative content of aggregates in colloidal magnetic nanoparticle suspension decreased with the increase in temperature, and the aggregates disrupted completely into monomers at or above the critical temperature. In addition, increasing the frequency of external alternating magnetic field significantly lowered down the critical temperature, and there existed a critical frequency where the critical temperature stabilized and became unaffected by the frequency. Preheating the suspension under critical frequency will disrupt the aggregates into monomers and thus optimize the heating efficiency of magnetic nanoparticles.

Hyperpolarized ³He or ¹²⁹Xe diffusion MRI has been demonstrated as a promising technique for the detection of microanatomical changes in chronic obstructive pulmonary disease (COPD). Compared with ³He, ¹²⁹Xe is more available for the potential clinical applications. However, the measurement of ¹²⁹Xe apparent diffusion coefficient (ADC) possesses more challenges due to the relevant low gyromagnetic ratio and spin polarization. In this present study, a single b value (b = 14 s/cm2) diffusion-weighted hyperpolarized ¹²⁹Xe MRI sequence was used to image a balloon phantom, healthy rats, and the COPD rats, respectively. All COPD rats were induced by second-hand smoke and lipopolysaccharide (LPS). The lung ¹²⁹Xe ADC maps were obtained on a 7 T MRI scanner. The mean lung parenchymal ¹²⁹Xe ADCs were 0.044 22±0.002 9 and 0.042 34±0.002 3 cm2/s (Δ = 0.8/1.2 ms) for the COPD rats, which showed significant increasements in comparison with healthy ones (0.037 7±0.002 3 and 0.036 7±0.001 3 cm2/s). Furthermore, the corresponding ADC histogram of the COPD rats exhibited a broader distribution as compared with the healthy ones. Our experiments demonstrated that the alveolar airspace
enlargement in the COPD rats are able to be quantitatively evaluated by hyperpolarized xenon diffusion-weighted MRI.

A new biocompatible gadolinium (III)-macromolecule (AP-EDA-DOTA-Gd) was developed as a magnetic resonance imaging (MRI) contrast agent. Poly (aspartic acid-cophenylalanine) was synthesized, modified via ethylenediamine, conjugated with 1,4,7,10-tetraazacyclododecane-1,4,7,10-tetraacetic acid (DOTA) and finally chelated gadolinium (III), yielding gadolinium (III)-based macromolecule (AP-EDA-DOTA-Gd). The hemolytic tests showed the hemocompatibility of this gadolinium (III)-based macromolecular conjugate. In vitro, AP-EDA-DOTA-Gd could be degraded, when it was incubated with cathepsin B in phosphate buffered solution (pH = 5.5). The T1-relaxivity (15.95 mmol–1·L·s–1) of AP-EDA-DOTA-Gd was 2.9 times of that (5.59 mmol–1·L·s–1) of the clinical MRI contrast agent (Gd-DOTA) at 1.5 T and 25 ℃. The liver enhancement of AP-EDA-DOTA-Gd was 63.5±6.1% during the maximum enhancement time (50-80 min), which was much better than that of Gd-DOTA (24.2±2.9%, 10-30 min). AP-EDA-DOTA-Gd was expected to be a potential liver MRI contrast agent.

Type 1 diabetes mellitus (T1DM) is a chronic metabolic disease characterized by insulin deficiency. Chronic T1DM causes damages to multiple organs. Numerous cross-sectional studies have shown that T1DM patients had significant cerebral atrophy, compared to normal subjects. However, few previous studies investigated progressive changes of cerebral atrophy over time in T1DM. In this study, a rat model of T1DM was established by a single dose of intraperitoneal injection of streptozotocin (STZ). Magnetic resonance imaging (MRI) was employed to measure the volumetric changes of the brain at 12 weeks and 20 weeks after STZ induction to follow the progressive brain atrophy assessment between these two time points. Histochemical staining was used assess neuropathologic changes in the brain regions showing progressive atrophy. The MRI results demonstrated that the STZ-treated rats had significantly reduced volume of grey matter (GM), white matter (WM) and whole brain, as compared to control. Voxel-wise analysis revealed significant effect of group×time interaction in multiple GM and WM regions. Results of Nissl staining and hematein-eosin staining (HE) indicate significant neuronal abnormality in the brain regions showing progressive atrophy, including somatosensory cortex, motor cortex and hippocampal CA3 region.

Motions occurring during in vivo magnetic resonance spectroscopy acquisition limit the quality of the spectra and their diagnostic value. Subtraction-based spectral editing techniques, such as J-difference editing, are especially prone to motion artifacts, causing false peaks or incomplete elimination of the background. In this study, a retrospective motion correction post-processing method was introduced to analyze GABA-edited spectra acquired by a MEGA point-resolved spectroscopy (MEGA-PRESS) sequence. Head motion during each acquisition was detected and recorded by the residual water signal; the motion-corrupted data were identified and excluded prior to averaging. The gain in spectral quality by using the proposed method was analyzed using the LCModel program. The results showed that the proposed method is an effective way to discard acquisitions corrupted by motion, and to improve spectral quality.

When performing chemical exchange saturation transfer (CEST) magnetic resonance imaging (MRI) on biological samples at high fields, nuclear Overhauser enhancement (NOE) effects may produce negative signals on the Z-spectra. The NOE-related signal intensity changes have the potential to become a new contrast mechanism. However, the presence of lipids signals within the same chemical shift range may affect the observation of true NOE effect-related signal changes. In this work, we first analyzed the influence of lipid signals on the NOE effects in tissues abundant in lipid via experiments on eggs. Similar experiments were then carried out on healthy rats and rats with brain tumors. The results showed that the presence of large lipid signals could cause pseudo NOE effects and degrade the accuracy of NOE contrast imaging.

Since the first generation ultra high field MRI systems were installed around 1990, ultra high field MRI has become one of the hottest research areas in the international community of magnetic resonance in medicine, due to the benefits of higher signal-to-noise ratio (SNR), better contrast, stronger blood oxygenation level dependent (BOLD) effect and increased spectral dispersion. In this review, the structure of latest 7.0 T whole body MRI system is briefly introduced, and the research progresses of ultra high field MRI and its future in medicine, neuroscience and cognitive science are also discussed.

Obesity has become a serious threat to human health. Magnetic resonance imaging (MRI) and in vivo magnetic resonance spectroscopy (MRS) provide powerful tools for evaluating the distribution and accumulation of adipose tissues in the body. Here, we reviewed the recent progresses in the applications of MRI and MRS technologies in basic and clinical researches on obesity.

In this study, we aimed to study the relationship between structural characteristics of the arcuate fasciculus (AF) derived from diffusion MRI and the performance of language comprehension. Constrained spherical deconvolution and tractography were used to analyze the diffusion MRI data. A total of 14 right-handed subjects with brain tumors in the left-hemisphere were recruited. Two MRI scan which one before the tumor-removing brain surgery and the other after the surgery were performed. In each scan, a short version of Boston diagnostic aphasia Examination (BDAE) test was applied to evaluate the performance of language comprehension. Four averaged diffusion indices along bilateral AF reconstructed by tractography were extrocted. The results showed that FA and RD in the left AF changed significantly (p<0.01) after the surgery, but similar changes were not found in the right AF. Diffusion indices of the left AF tract showed significantly higher (p<0.01) correlation with the BDAE results (r=0.6~0.8), relative to the right AF (r=0.3~0.4). The findings are consistent with previous reports concerning the functional lateralization of the AF.

Occlusion of lenticulostriate arteries (LSAs) often leads to lacunar infarction. Digital subtraction angiography (DSA) has high spatial resolution and superior definition of small vessels, and is the most widely used clinical routine to image LSAs. One drawback of DSA, however, is its invasiveness. Time-of-flight (TOF) magnetic resonance angiography (MRA) was recently shown to be able to image LSAs at ultra-high fields (i.e., 7 T) with specially-designed receiving coils. However, image LSAs with TOF-MRA on clinical MRI systems can be difficult due to the facts that blood flow in LSAs is slow and the luminal diameters in these vessels are small (i.e., 0.3~0.7 mm) relative to the dimension of imaging voxels. In this study, the flow-sensitive black-blood (FSBB) technique was optimized to image LSAs at 3 T to achieve a proper balance among spatial resolution, signal-to-noise ratio and scan time.

The MEGA-PRESS sequence has been widely used to measure the δ_H 3.02 γ-amino butyric acid (GABA) signal by J-difference editing. However, the sequence cannot eliminate the macromolecule signals at the same chemical shift completely. Symmetrical editing can be applied to suppress the macromolecule signals. However, the approach is rarely applied at field strength of 3 T due to insufficient frequency selectivity of the editing pulse. In this study, GABA+(i.e., GABA+macromolecules) and macromolecule-suppressed GABA signals in the occipital lobe of human subjects were measured with symmetrical editing at 3 T. The duration of the editing pulse was increased from 14 ms to 20 ms to improve frequency selectivity, and echo time (TE) from 68 ms to 80 ms. It was found that the fraction of the total signal retained following macromolecule suppression ([GABA]/[GABA+]) was 0.73. It is concluded that symmetric macromoleculesuppressed editing can be used to acquire macromolecule-suppressed GABA signals, and may serve as a better tool to assess inter-individual differences in cerebral GABA level.

Clinical magnetic resonance angiography (MRA) often involves extraction of images, which is often done manually by radiologists. The process can be tedious and time-consuming. In this study, we propose a new parallel vessel segmentation/tracking algorithm, utilizing large-scale parallel computing provided by graphics processing unit (GPU). The whole three-dimensional image volumes are first divided into small cubes, which share surface with their neighbors. Each cube is then processed separately to determine whether there are vessels passing through its surface. These results are then used for global segmentation and vessel tracking. Application of the algorithm to real MRA data showed that segmentation of a whole-brain MRA dataset could be achieved in less than 1 s.

A T₁ magnetic resonance imaging (MRI) contrast agent specific to lung cancer stem cells Gd-DOTA-HCBP-1 was synthesized by solid-phase peptide synthesis technology. Its longitudinal relaxation rate (r₁) was 6.15 mmol^-1·L·s^-1, approximately 0.6 times higher than that of the commercial T₁ contrast agents Dotarem at 11.7 T. In vitro MRI experiments indicated that the detectability of the cancer stem cells could be significantly improved to about 2 000~4 000 cells per sphere by the use of the probe.

Acquisition parameters for chemical exchange saturation transfer (CEST) imaging were optimized on a GE Signa HDe 1.5 T magnetic resonance imaging (MRI) scanner with phantoms and clinical cases. The effects of matrix size, number of averages (NEX) and flip angles on the quality of CEST images were assessed. It was shown that the signal-to-noise ratio (SNR) of the CEST images acquired on the 1.5 T scanner was relatively low, and the stability and uniformity of the B₀ field affected the outcome significantly. Reducing matrix size and increasing NEX improved the SNR of the CEST images. Optimal flip angle for magnetization transfer was found to be 105°. With a NEX of 2, usable Z spectra could be obtained. The Z spectra indicated that, with the saturation pulse frequency centered at -294~-194 Hz, signal differences could be observed for 30% Glu, I₃₂₀, H₂O, and Cr. Maximal signal differences were observed when the saturation pulse applied at -244~-214 Hz. Amide proton transfer (APT) imaging on patients showed that 25 cases of brain tumor had high CEST signals, 12 cases of cerebral infarction had low CEST signals. It was therefore possible to differentiate brain tumor from infarction with CEST imaging. There were also 12 cases which failed due to long acquisition time, patient movements, and temperature changes in the scanner room.

We synthesized TPP-Lys(Acp-DOTA-Gd)-COOH (Gd-DOTA-TPP) as a novel T₂-weighted MRI contrast agent, and used it to label human mesenchymal stem cells (hMSCs) via electroporation. hMSCs labeled with Gd-DOTA-TPP presented a rather dark T₂-weighted image at a cellular Gd content of about 9×10⁹ Gd/cell. The labeled cells were syringe-injected into a mouse brain with defined cell numbers followed by T₂-weighted MRI on an 11.7 T system, which yield an in vivo imaging sensitivity of about 10³ cells.

Magnetic resonance imaging (MRI) is a noninvasive intraocular tumor detection method without ionizing radiation. However, resolution limitation and motion artifacts are difficult to overcome in the imaging process. Conventional scanning methods inevitably introduce motion artifacts, or require the subjects to cooperate for accurate eye fixation, increasing the difficulty of imaging and giving the subject uncomfortable experiences. In this work, a new MRI method based on super-resolution theory is proposed, which uses a specialized orbit coil to scan a series of dynamic images of the eyeball, such that the acquisition resolution in different directions is complementary. High-resolution eyeball images with minimal motion artifacts could then be obtained after pre-processing, registration, super-resolution reconstruction and other operations. The study showed that the method proposed can be used to obtain clear eyeball images without the requirement of eye fixation.

Magnetic resonance imaging (MRI) is a practical technique for non-invasive temperature measurement. Thermo-sensitive MRI contrast agents, which cause temperature-dependent changes in magnetic resonance parameters of water molecules (i.e., relaxation, signal intensity and chemical shift) have been widely applied to MRI thermometry. This paper gives an overview of the types, principle, research status and application prospects of thermo-sensitive MRI contrast agent.

Abnormal metabolism γ-aminobutyric acid (GABA) is involved in many brain diseases, and in vivo GABA detection is important for the diagnosis of these diseases. This study optimized a commonly used spectral editing method for in vivo GABA detection, MEGA-PRESS. We used high-performance Shinnar-Le Roux (SLR) pulses to replace the commonly used Gaussian pulses. Single-band pulse and dual-band pulse were combined to excite the resonances of given metabolites. An optimized MEGA-PRESS sequence was implemented on the uMR780 magnetic resonance imaging system manufactured by United Imaging Healthcare (UIH). The results showed that the SLR pulse could reduce the interference of macromolecule (MM) on GABA signal. Compared with the traditional methods, combining the single-band and dual-band pulses could better suppress contamination of residual water signal to the GABA^* signal. Volunteer experiments demonstrated that the optimized MEGA-PRESS sequence improved in vivo measurement of GABA^* concentration.

To provide guidance to clinical diagnosis and develop individual-based treatment/prognostic evaluation, this study explored the correlations between morphological characteristics of glioblastoma (GBM) measured by conventional MRI and expression levels of GBM-specific molecular biomarkers. MRI data of 60 patients with histopathologically confirmed GBM were analyzed retrospectively, from which the morphological characteristics of GMB (i.e., TLD, ELD and CTE/TLD ratio) were measured. Histochemical staining was used to measure the levels of GBM-specific molecular biomarkers (i.e., EGFR, IDH-1, TP53, PTEN and NEFL), which were calculated as the product of staining index and stained area. Correlations among the morphological characteristics and the levels of molecular biomarkers were assessed using multivariate logistic regression analyses. There was a significant correlation between TLD and EGFR level (r=0.796). The correlation between TLD and PTEN level was positive (r=0.533), while the correlation between TLD and IDH-1 level was negative (r=-0.672). CTE/TLD had a positive correlation with EGFR (r=0.622) and PTEN (r=0.638) levels, along with a negative correlation with IDH-1 level (r=-0.493). ELD had no significant correlation with the levels of all five molecular biomarkers. It was concluded that the morphological characteristics of GBM derived from conventional MRI had correlations with the expression levels of EGFR, IDH-1 and PTEN, and it was possible to derive heterogeneity, molecular classification and prognosis of GBM using conventional MRI.

Cardiac right ventricle (RV) segmentation plays an essential role in the functional analysis of heart diseases, such as pulmonary hypertension. The myocardium of RV is thin and irregular-shaped, making the traditional segmentation methods less effective. To improve RV segmentation, a COLLATE (Consensus Level, Labeler Accuracy and Truth Estimation) fusion-based multi-atlas method was developed. The preprocessed target image was first registered to atlas images with a B-spline algorithm optimizing normalized mutual information. The registration coefficients obtained were then used to get a rough RV segmentation for COLLATE fusion. Shape-constrained region growing algorithm was used to correct the segmentation errors. Ten cardiac magnetic resonance datasets were blindly selected to compare the performance of RV segmentation between the method developed and a method based on deep learning. The results of manual segmentation were used as the golden standard. Ejection fraction (EF) calculated with the proposed segmentation method showed better correlation and consistency with the golden standard, relative to the results calculated with the deep learning method.

Training data enrichment is a key factor in artificial intelligence (AI) technology development. At present, the bottleneck problem is that the quantity and type of labeled training data in valid samples are unable to meet the requirements of AI+MRI aided diagnosis. In this paper, an effective approach to solve the problem was presented. High resolution isotropic multi-dimensional data of regions of interests from patients or healthy volunteers were first acquired via a series of scanning on clinical MRI scanners, including quantitative T₁, T₂, proton density (Pd) and apparent diffusion coefficient (ADC) measurements. These data were then used as the ground truth, from which different types of images associated with different imaging sequences and parameters were obtained with a virtual MRI technology. The type of the images with the best boundary resolution were then selected manually by experienced doctors, on which three-dimensional mask matrix was obtained by manual contouring and labeling, serving as the template for other types of images. This enrichment method was developed as a software platform, which could provide sufficient quantity of image data from a small number of positive cases, thus meeting the data training enrichment requirement of AI+MRI diagnosis at low cost and with high efficiency.

Diffusion-weighted magnetic resonance images contain rich information on brain white matter (WM), and have been used to construct brain templates/atlases. However, the accuracy of brain templates reconstructed with the diffusion tensor model is often limited in regions with complex WM configurations. To overcome this problem, high angular resolution diffusion imaging acquisition and reconstruction have been developed for building higher quality brain template. In this paper, the research progresses on the construction of brain template using diffusion tensor imaging (DTI) are reviewed. Firstly, the technical aspects and limitations in using DTI data to build brain templates were discussed. Secondly, diffusion spectrum imaging and high angular resolution diffusion imaging techniques were introduced, with their advantages over the DTI techniques explained. Finally, the existing problems and the perspectives of the field were discussed.

Real-time monitoring of tissue temperature is required during high intensity focused ultrasound (HIFU) tumor treatment to ensure safety and effectiveness. Magnetic resonance imaging (MRI) can be used to measure tissue temperature non-invasively during HIFU treatment. This paper examined the effects of coagulation necrosis-induced tissue phase transition on magnetic resonance thermometry (MRT) during HIFU tumor treatment. With a two-state rapid exchange model, the relationship between tissue longitudinal relaxation time (T₁) and temperature before and after HIFU radiation-induced coagulation necrosis/tissue phase transition were analyzed theoretically. Taking the effects of tissue phase transition into account, the experimental scheme and data processing procedures for MRT were optimized, and better temperature measurements were obtained. The work demonstrated the importance of considering the effects of tissue phase transition in real-time MRT during HIFU treatment.

Chemical exchange saturation transfer (CEST) is a novel molecular magnetic resonance imaging (MRI) method that makes full use of the chemical exchange between water protons and exchangeable protons from solute molecules. Through chemical exchanges, the MRI signal of bulk water decreases when saturating at specific frequency of solute protons. Important biological information related to the chemical exchange processes could be extracted from the amplitude of water signal decrease. For example, the CEST signals have been used for in vivo pH imaging since the proton exchange rate is often pH-dependent. CEST signals originated from endogenous proteins/peptides and exogenous small molecule/metal chelate probes have been used for in vivo pH imaging. Using the ratiometric methods or the amine and amide concentration-independent detection (AACID) methods, in vivo pH maps have been acquired from kidneys, ischemic brain and tumors. In this paper, we thoroughly reviewed the progresses during the past twenty years in the field of in vivo pH imaging with CEST contrast. Principles, methods and applications were discussed, as well as development trends and future directions.

Magnetic resonance spectroscopy (MRS) is useful for rapid and robust liver fat quantification. MR spectra of liver can be acquired in a single breath-hold with the high-speed multi-echo stimulated echo acquisition mode, from which T₂-corrected liver proton density fat fraction (PDFF) can be derived. However, the widespread application of this method is hampered by the cumbersome post-processing procedures for the spectroscopy data. In this work, an automated algorithm for liver fat quantification from multi-echo spectroscopy data was developed. The feasibility and accuracy of the algorithm were evaluated with fat-water phantoms experiments and human liver experiments. The results demonstrated excellent correlation and agreement between the PDFF measurements with the proposed algorithm and those obtained with the jMRUI software.