Search

The successful development of the image denoising techniques for low-dose computed tomography (LDCT) was largely owing to the public-domain availability of spatially-aligned high- and low-dose CT image pairs. Even though low-dose CT scans are also highly desired in dental imaging, public-domain databases of dental CT image pairs have not been established yet. In this paper, we propose a dental CT image denoising method based on the transfer learning of a generative adversarial network (GAN) from the public-domain CT images. We trained a generative adversarial network with the Wasserstein loss function (WGAN) using 5,100 high- and low-dose medical CT image pairs of human chest and abdomen. For the generative network of GAN, we used the U-net structure of five stages to exploit its high computational efficiency. After training the proposed network, named U-WGAN, we fine-tuned the network with 3,006 dental CT image pairs of two different human skull phantoms. For the high- and low-dose scans of the phantoms, we set the tube current of the dental CT to 10 mA and 4 mA, respectively, with the tube voltage set to 90 kVp in both scans. We applied the trained network to denoising of low-dose dental CT images of dental phantoms and adult humans. The U-net processed images showed over-smoothing effects even though U-net had a good performance in the quantitative metrics. U-WGAN showed similar denoising performance to WGAN, but it reduced the computation time of WGAN by a factor of 10. The fine-tuning procedure in the transfer learning scheme enhanced the network performance in terms of the quantitative metrics, and it also improved visual appearance of the processed images. Even though the number of fine-tuning images was very limited in this study, we think the transfer learning scheme can be a good option for developing deep learning networks for dental CT image denoising.

The objective of this study was to evaluate susceptibility changes caused by iron accumulation in cognitive normal (CN) elderly, those with amnestic mild cognitive impairment (aMCI), and those with early state AD, and to compare the findings with gray matter volume (GMV) changes caused by neuronal loss. The participants included 19 elderly CN, 19 aMCI, and 19 AD subjects. The voxel-based quantitative susceptibility map (QSM) and GMV in the brain were calculated and the differences of those insides were compared among the three groups. The differences of the QSM data and GMVs among the three groups were investigated by voxel-based and region of interest (ROI)-based comparisons using a one-way analysis of covariance (ANCOVA) test with the gender and age as covariates. Finally, a receiver-operating-characteristic (ROC) curve analysis was performed. The voxel-based results showed that QSM demonstrated more areas with significant difference between the CN and AD groups compared to GMV. GMVs were decreased, but QSM values were increased in aMCI and AD groups compared with the CN group. QSM better differentiated aMCI from CN than GMV in the precuneus and allocortex regions. In the accumulation regions of iron and amyloid β, QSM can be used to differentiate between CN and aMCI groups, indicating a useful an auxiliary imaging for early diagnosis of AD., Highlights • The susceptibility difference was more sensitive than GMV change in known regions of iron and the amyloid β accumulations. • QSM values were increased in patients compared with normal elderly and better differentiated aMCI from CN than GMV values. • The QSM technology can be used as an auxiliary imaging for early diagnosis of AD.

Half-scan image reconstruction with Parker weighting can correct motion artifacts in dental CT images taken with a slow scan-based dental CT. Since the residual half-scan artifacts in the dental CT images appear much stronger than those in medical CT images, the artifacts often persist to the extent that they compromise the surface-rendered bone and tooth images computed from the dental CT images. We used a variation of generative adversarial network (GAN), so-called U-WGAN, to correct half-scan artifacts in dental CT images. For the generative network of GAN, we used a U-net structure of five stages to take advantage of its high computational efficiency. We trained the network using the Wasserstein loss function on the dental CT images of 40 patients. We tested the network with comparing its output images to the half-scan images corrected with other methods; Parker weighting and the other two popular GANs, that is, SRGAN and m-WGAN. For the quantitative comparison, we used the image quality metrics measuring the similarity of the corrected images to the full-scan images (reference images) and the noise level on the corrected images. We also compared the visual quality of the surface-rendered bone and tooth images. We observed that the proposed network outperformed Parker weighting and other GANs in all the image quality metrics. The computation time for the proposed network to process 336 × 336 × 336 3D images on a GPU-equipped personal computer was about 3 s, which was much shorter than those of SRGAN and m-WGAN, 50 s and 54 s, respectively.

High-resolution imaging is essential in three-dimensional (3D) CT image-based digital dentistry. A small amount of head motion during a CT scan can degrade the spatial resolution of the images to the extent where the efficacy of 3D image-based digital dentistry is greatly compromised. We introduce a retrospective motion artifact reduction (MAR) method for dental CTs that eliminates the necessity for any external motion tracking devices. Assuming that rigid-body motions are dominant in a dental scan of a human head, we extracted motion information from the projection data. By taking the cross-correlation between two successive projection data for every projection view, we determined the displacement of the projection data at each view. We experimentally found that any motion of the imaging object during the scan resulted in displacement of the projection data proportional to the motion amplitude. We decomposed the displacement into two components, one caused by translational motion and the other caused by rotational motion. The displacement components were used to correct the projection data before CT image reconstruction. We experimentally verified the MAR method using the projection data of a few phantoms acquired through a clinical dental CT machine. When the MAR performance was evaluated by the structural similarity index (SSIM) and the normalized absolute error (NAE) in reference to the motion-less images, the SSIM improved to 99% while the NAE was reduced by 80-90%.

A small head motion of the patient can compromise the image quality in a dental CT, in which a slow cone-beam scan is adopted. We introduce a retrospective head motion estimation method by which we can estimate the motion waveform from the projection images without employing any external motion monitoring devices. We compute the cross-correlation between every two successive projection images, which results in a sinusoid-like displacement curve over the projection view when there is no patient motion. However, the displacement curve deviates from the sinusoid-like form when patient motion occurs. We develop a method to estimate the motion waveform with a single parameter derived from the displacement curve with aid of image entropy minimization. To verify the motion estimation method, we use a lab-built micro-CT that can emulate major head motions during dental CT scans, such as tilting and nodding, in a controlled way. We find that the estimated motion waveform conforms well to the actual motion waveform. To further verify the motion estimation method, we correct the motion artifacts with the estimated motion waveform. After motion artifact correction, the corrected images look almost identical to the reference images, with structural similarity index values greater than 0.81 in the phantom and rat imaging studies.

A dual-modal approach using Raman spectroscopy and optical pH sensing was investigated to discriminate between normal and cancerous tissues. Raman spectroscopy has demonstrated the potential for in vivo cancer detection. However, Raman spectroscopy has suffered from strong fluorescence background of biological samples and subtle spectral differences between normal and disease tissues. To overcome those issues, pH sensing is adopted to Raman spectroscopy as a dual-modal approach. Based on the fact that the pH level in cancerous tissues is lower than that in normal tissues due to insufficient vasculature formation, the dual-modal approach combining the chemical information of Raman spectrum and the metabolic information of pH level can improve the specificity of cancer diagnosis. From human breast tissue samples, Raman spectra and pH levels are measured using fiber-optic-based Raman and pH probes, respectively. The pH sensing is based on the dependence of pH level on optical transmission spectrum. Multivariate statistical analysis is performed to evaluate the classification capability of the dual-modal method. The analytical results show that the dual-modal method based on Raman spectroscopy and optical pH sensing can improve the performance of cancer classification.

In this paper, two novel approaches, gradient-based and direct strain estimation techniques, are proposed for high-quality average strain imaging incorporating a cost function maximization. Stiffness typically is a continuous function. Consequently, stiffness of proximal tissues is very close to that of the tissue corresponding to a given data window. Hence, a cost function is defined from exponentially weighted neighboring pre- and post-compression RF echo normalized cross-correlation peaks in the lateral (for displacement estimation) or in both the axial and the lateral (for direct strain estimation) directions. This enforces a controlled continuity in displacement/strain and average displacement/strain is calculated from the corresponding maximized cost function. Axial stress causes lateral shift in the tissue. Therefore, a 1-D post-compression echo segment is selected by incorporating Poisson's ratio. Two stretching factors are considered simultaneously in gradient-based strain estimation that allow imaging the lesions properly. The proposed time-domain gradient-based and direct-strain-estimation-based algorithms demonstrate significantly better performance in terms of elastographic signal-to-noise ratio (SNRe), elastographic contrast-to-noise ratio (CNRe), peak signal-to-noise ratio (PSNR), and mean structural similarity (MSSIM) than the other reported time-domain gradient-based and direct-strain-estimation techniques in finite element modeling (FEM) simulation and phantom experiments. For example, in FEM simulation, it has been found that the proposed direct strain estimation method can improve up to approximately 2.49 to 8.71, 2.2 to 6.63, 1.5 to 5, and 1.59 to 2.45 dB in the SNRe, CNRe, PSNR, and MSSIM compared with the traditional direct strain estimation method, respectively, and the proposed gradient-based algorithm demonstrates 2.99 to 16.26, 18.74 to 23.88, 3 to 9.5, and 0.6 to 5.36 dB improvement in the SNRe, CNRe, PSNR, and MSSIM, respectively, compared with a recently reported time-domain gradient-based technique. The range of improvement as noted above is for low to high applied strains. In addition, the comparative results using the in vivo breast data (including malignant or benign masses) also show that the lesion size is better defined by the proposed gradient-based average strain estimation technique.

In ultrasound elastography, the strain in compressed tissue due to external deformation is estimated and is smaller in harder than softer tissue. With increased stress, the nonaxial motions of tissue elements increase and result in noisier strain images. At high strain, the envelope of the rf signal exhibits robustness to signal decorrelation. However, the precision of strain estimates using envelope signals is much worse compared to that using the rf signals. In this paper, we propose a novel approach for robust strain estimation by combining weighted rf cross-correlation and envelope cross-correlation functions. An applied strain-dependent piecewise-linear-weight is used for this purpose. In addition, we introduce nonlinear diffusion filtering to further enhance the resulting strain image. The results of our algorithm are demonstrated for up to 10% applied strain using a finite-element modelling (FEM) simulation phantom. It reveals that the elastographic signal-to-noise ratio ( SNRe) and the elastographic contrast-to-noise ratio ( CNRe) of the strain images can be improved more significantly than with other algorithms used in this paper. In addition, comparative results in terms of the mean structural similarity (MSSIM) using in vivo breast data show that the strain image quality can be improved noticeably by the proposed method than with the techniques employed in this work.

To develop an electrical properties tomography (EPT) technique that can provide in vivo electrical conductivity and permittivity images of biological tissue without performing complex-valued radiofrequency field measurements.Electrical conductivity and permittivity images are modeled as a monotonic function of tissues' water content (W) under the principle of Maxwell's mixture theory. Water content maps are estimated from two spin-echo images having different repetition times (TRs). For the modeling functions, physically measured parameters (electrical properties, water content, and TThe electrical properties (EPs) of the phantom estimated by the proposed method match well with the values measured on the bench. The conductivity and permittivity maps from all experiments show uncompromised spatial resolution without boundary artifacts and higher contrast when compared with water content maps.Human brain and phantom EP images suggest that water content is a dominating factor in determining the electrical properties of tissues. Despite possible literature inaccuracies, the proposed method offers EP maps that can provide complementary information to current approaches, to facilitate EPT scans in clinical applications. Magn Reson Med 77:1094-1103, 2017. © 2016 International Society for Magnetic Resonance in Medicine.

Strain distribution in compressed tissues gives information about elasticity of the tissues. We have measured strain from two sets of 3D micro-CT images of a breast-mimicking phantom; one obtained without compressing the phantom and the other with compressing it. To measure strain, we first calculated compression-induced displacements of high-intensity feature patterns in the image. In measuring displacement of a pixel of interest, we searched the pixel in the compressed-phantom image, whose surrounding resembles the uncompressed-phantom image most closely, using the image correlation technique. From the displacement data, we calculated average strain at a region of interest. With the calculated average strains, we could distinguish the hard inclusion in the phantom which was not distinguishable from the background body of the phantom in the ordinary micro-CT images. The calculated strains account for stiffness of the tissue of interest, one of the important parameters for diagnosing malignant tissues. We present experimental results of the displacement and strain measurement along with FEM analysis results.

We introduce a square coil system for remote magnetic navigation of a magnetic device without any physical movements of the coils. We used three square-Helmholtz coils and a square-Maxwell coil for magnetic propulsion of a small magnet along the desired path. All the square coils are mountable on a cubic frame that has an opening to accommodate a living subject. The square-Helmholtz coils control the magnetic propulsion direction by generating uniform magnetic field along the desired direction while the square-Maxwell coil controls the propulsion force by generating magnetic gradient field. We performed magnetic propulsion experiments with a down-scaled coil set and a three-channel coil driver. Experimental results demonstrate that we can use the square coil set for magnetic navigation of a magnetic device without any physical movements of the coils.

The present study examined the effects of human umbilical cord blood-derived mesenchymal stem cells (HUCB-derived MSCs) delivered through the basilar artery in a canine thromboembolic brain ischemia model. Cerebral ischemia was induced through occlusion of the middle cerebral artery by injecting thrombus emboli into 10 beagles. In the HUCBC group (n = 5), 1 × 106 HUCB-derived MSCs were transplanted through the basilar artery 1 day after ischemic induction using an endovascular interventional approach. In the control group (n = 5), phosphate-buffered saline (PBS) was injected in the same manner in as the HUCBC group. Upon neurobehavioral examination, earlier recovery was observed in the HUCBC group. The HUCBC group showed a decrease in the infarction volume at 1 week after cerebral ischemic induction, whereas the control group showed an increase in the infarction volume at 1 week, by magnetic resonance image analysis. Transplanted cells had differentiated into neurons and astrocytes and were observed in and around endothelial cells that were positive for von Willebrand factor (vWF). HUCB-derived MSCs expressed neuroprotective factors, such as brain-derived neurotrophic factor (BDNF) and vascular endothelial growth factor (VEGF), at 4 weeks after the transplantation. The transplanted cells demonstrated their efficacy by reducing the infarction lesion volume and through earlier recovery from the neurological deficit. These results suggest that intraarterial transplantation of HUCB-derived MSCs could be useful in clinical treatment of cerebral ischemia. © 2009 Wiley-Liss, Inc.

The present study introduces a new approach to determining optimal electrode positions in transcranial direct current stimulation (tDCS). Electric field and 3D conduction current density were analyzed using 3D finite element method (FEM) formulated for a dc conduction problem. The electrode positions for minimal current injection were optimized by changing the Cartesian coordinate system into the spherical coordinate system and applying the (2+6) evolution strategy (ES) algorithm. Preliminary simulation studies applied to a standard three-layer head model demonstrated that the proposed approach is promising in enhancing the performance of tDCS.

Diffusion properties of the thalamocortical pathways were compared between 30 schizophrenic patients and 22 healthy age-matched and sex-matched volunteers. T1-weighted and diffusion tensor images were used for cortical parcellation and probabilistic tractography to generate population-based thalamocortical pathway maps. The weighted fractional anisotropy and the longitudinal and transversal diffusivity of each pathway were calculated for individual participants. Patients with schizophrenia demonstrated decreased anisotropy and increased longitudinal and transversal diffusivity within the pathways from the mediodorsal nucleus and the pulvinar to the orbitofrontal and parietal-occipital-temporal lobes. Anisotropy decreases and diffusivity increases were correlated with positive and negative syndrome scale scores. Alterations in global diffusion properties within the thalamocortical pathways imply the disruption of brain circuitry in schizophrenic patients.

This study investigated the effect of 40% oxygen administration on n-back task performance, blood oxygen saturation and heart rate. Five male (25.8 +/- 1.3 years) and five female (23.0 +/- 1.0 years) college students were selected as the subjects for this study. The experiment consisted of two runs: one was an n-back task with normal air (21% oxygen) administered and the other was with hyperoxic air (40% oxygen) administered. The experimental sequence in each run consisted of Rest1 (1 min), 0-back task (1 min), 2-back task (2 min) and Rest2 (4 min). Blood oxygen saturation and heart rate were measured throughout the four phases. The results of the n-back behavioural analysis reveal that accuracy rates were enhanced with 40% oxygen administration compared to 21% oxygen. When 40% oxygen was supplied, blood oxygen saturation was increased and heart rate was decreased compared to that with 21% oxygen administration. It is suggested that 40% oxygen can stimulate brain activation by increasing actual blood oxygen concentration in the process of cognitive performance, and hyperoxia makes heart rate decrease. This result supports the hypothesis that 40% oxygen administration would lead to increases in n-back task performance.

In high field MRI capable of multi-channel radio frequency (RF) transmission, B 1 shimming is a time-consuming job because conventional B 1 shimming techniques require B 1 mapping for each channel. After acquiring the complex-numbered B 1 field maps, the optimal amplitude and phase of the driving RF pulse are determined for each channel to maximize the B 1 field uniformity in conventional B 1 shimming. However, time-consuming B 1 shimming procedures at the pre-scan may not be tolerated in the clinical imaging in which patient throughput is one of the important factors. To avoid the time-consuming B 1 mapping, the first spin echo and the stimulated echo were repeatedly acquired in the slice-selective stimulated echo sequence without imaging gradients. A cost function of the amplitudes and phases of the driving RF pulse for every channel was defined in a way that the ratio between the spin echo and stimulated echo amplitudes rapidly converged to √ 2. The amplitude and phase of the driving RF pulse were iteratively modified over the repeating RF pulse sequence so that the cost function was minimized. From the finite-difference-time-domain (FDTD) electromagnetic field simulations with a human body model placed in a birdcage coil operating at 3 T, it was observed that the RF pulse calibration with iterative cost function minimization can give improvement of B1 field uniformity as well as flip-angle calibration. The experiments at 3 T also showed improvement of RF field uniformity in the phantom imaging studies. Since the proposed RF pulse calibration is not based on B 1 mapping, the RF pulse calibration time could be much shorter than the B 1-mapping based methods. The proposed method is expected to be a practical substitute for the B 1-mapping-based B 1 shimming methods when long pre-scan time is not tolerable.

In magnetic resonance electrical impedance tomography (MREIT), we measure the induced magnetic flux density inside an object subject to an externally injected current. This magnetic flux density is contaminated with noise, which ultimately limits the quality of reconstructed conductivity and current density images. By analysing and experimentally verifying the amount of noise in images gathered from two MREIT systems, we found that a carefully designed MREIT study will be able to reduce noise levels below 0.25 and 0.05 nT at main magnetic field strengths of 3 and 11 T, respectively, at a voxel size of 3 x 3 x 3 mm(3). Further noise level reductions can be achieved by optimizing MREIT pulse sequences and using signal averaging. We suggest two different methods to estimate magnetic flux noise levels, and the results are compared to validate the experimental setup of an MREIT system.

We present cross-sectional conductivity images of two biological tissue phantoms. Each of the cylindrical phantoms with both diameter and height of 140 mm contained chunks of biological tissues such as bovine tongue and liver, porcine muscle and chicken breast within a conductive agar gelatin as the background medium. We attached four recessed electrodes on the sides of the phantom with equal spacing among them. Injecting current pulses of 480 or 120 mA ms into the phantom along two different directions, we measured the z-component Bz of the induced magnetic flux density B = (Bx, By, Bz) with a magnetic resonance electrical impedance tomography (MREIT) system based on a 3.0 T MRI scanner. Using the harmonic Bz algorithm, we reconstructed cross-sectional conductivity images from the measured Bz data. Reconstructed images clearly distinguish different tissues in terms of both their shapes and conductivity values. In this paper, we experimentally demonstrate the feasibility of the MREIT technique in producing conductivity images of different biological soft tissues with a high spatial resolution and accuracy when we use a sufficient amount of the injection current.

This study aimed to investigate the hypothesis that administration of the air with 30% oxygen compared with normal air (21% oxygen) enhances cognitive functioning through increased activation in the brain. A visuospatial task was presented while brain images were scanned by a 3 T fMRI system. The results showed that there was an improvement in performance and also increased activation in several brain areas in the higher oxygen condition. These results suggest that a higher concentration of breathed oxygen increases saturation of blood oxygen in the brain and facilitates performance.

Magnetic resonance electrical impedance tomography (MREIT) is a recently developed imaging technique that combines MRI and electrical impedance tomography (EIT). In MREIT, cross-sectional electrical conductivity images are reconstructed from the internal magnetic field density data produced inside an electrically conducting subject when an electrical current is injected into the subject. In this work the results of an electrical conductivity imaging experiment are presented, along with some practical considerations regarding MREIT. The MREIT experiment was performed with a 0.3 Tesla MRI system on a phantom made of two compartments with different electrical conductivities. The current density inside the phantom was measured by the MR current density imaging (MRCDI) technique. The measured current density was then used for conductivity image reconstruction by the J-substitution algorithm. The conductivity phantom images obtained with an injection current of 28mA showed conductivity errors of about 25.5%.

To validate the use of adaptive nonlinear filters in reconstructing conductivity and permittivity images from the noisy B₁(+) maps in electrical properties tomography (EPT).In EPT, electrical property images are computed by taking Laplacian of the B₁(+) maps. To mitigate the noise amplification in computing the Laplacian, the authors applied adaptive nonlinear denoising filters to the measured complex B₁(+) maps. After the denoising process, they computed the Laplacian by central differences. They performed EPT experiments on phantoms and a human brain at 3 T along with corresponding EPT simulations on finite-difference time-domain models. They evaluated the EPT images comparing them with the ones obtained by previous EPT reconstruction methods.In both the EPT simulations and experiments, the nonlinear filtering greatly improved the EPT image quality when evaluated in terms of the mean and standard deviation of the electrical property values at the regions of interest. The proposed method also improved the overall similarity between the reconstructed conductivity images and the true shapes of the conductivity distribution.The nonlinear denoising enabled us to obtain better-quality EPT images of the phantoms and the human brain at 3 T.

Fast B1 mapping based on short-TR sequences is prone to T1-induced errors. The purpose of this study is to develop a novel fast B1 mapping method that is less prone to T1-induced errors.The authors acquired three gradient echoes by applying three RF pulses of different flip angles in an interleaved manner. The new method, named interleaved-three-flip-angle (ITFA) method, employs a short TR for fast scan. Since the pixel intensity of the gradient echo images is dependent on both B1 and T1, the authors could compute a B1 map from the three gradient echo images with excluding the T1-effects. The authors simulated the proposed B1 mapping method for various T1 values, and they found optimal flip angles for ITFA experiments for a given repetition time. To evaluate the B1 mapping performance, the authors made a human-brain-mimicking phantom that had six compartments with different T1 and T2. The authors performed B1 mapping experiments at 3T on the phantom and a volunteer using the ITFA method, the actual flip angle imaging (AFI) method, and the double angle method (DAM), and then, the authors compared the B1 mapping results.Using a birdcage coil as a transceiver at 3T, the authors performed ITFA scans of the phantom and a volunteer with TR of 60 ms and the nominal flips angles of (25°, 70°, 80°). The authors also performed AFI scans with TR1/TR2 of 30/150 ms and the nominal flip angle of 60°. In both the phantom and human head imaging performed with the same scan times for ITFA and AFI, ITFA showed smaller average B1 errors than AFI when they were compared to DAM.ITFA excitations made it possible to reduce the T1-effects on B1 mapping of the human-brain-mimicking phantom and the human brain at 3T. The authors expect the ITFA method can be used for B1 shimming once the optimal flip angles have been predetermined for the target imaging region and for the preferred TR.

Gradient coil inductance has been remarkably reduced by the minimum-inductance design technique, which minimizes the magnetic energy stored by the gradient coil. The planar gradient coil designed by this technique, however, often has poor magnetic field linearity. Scaling the spatial frequencies of the current density function derived by this method, the magnetic field linearity of the planar gradient coil can be greatly improved with a small sacrifice of gradient coil inductance. A figure of merit of the planar gradient coil has been found to be improved by scaling the spatial frequencies.

The average cerebellar volume of Korean men (135.19 cm3) is larger than that of Korean women (123.06 cm3), and that of subjects in their twenties (134.28 cm3) is larger than that of subjects in their forties (121.83 cm3). Atrophy of the cerebellum is more markedly observed in men than in women. There is a relation between body weight and cerebellar volume for men, but not for women.

In this paper, a phase-based direct average strain estimation method is developed. A mathematical model is presented to calculate axial strain directly from the phase of the zero-lag cross-correlation function between the windowed precompression and stretched post-compression analytic signals. Unlike phase-based conventional strain estimators, for which strain is computed from the displacement field, strain in this paper is computed in one step using the secant algorithm by exploiting the direct phase-strain relationship. To maintain strain continuity, instead of using the instantaneous phase of the interrogative window alone, an average phase function is defined using the phases of the neighboring windows with the assumption that the strain is essentially similar in a close physical proximity to the interrogative window. This method accounts for the effect of lateral shift but without requiring a prior estimate of the applied strain. Moreover, the strain can be computed both in the compression and relaxation phases of the applied pressure. The performance of the proposed strain estimator is analyzed in terms of the quality metrics elastographic signal-to-noise ratio (SNRe), elastographic contrast-to-noise ratio (CNRe), and mean structural similarity (MSSIM), using a finite element modeling simulation phantom. The results reveal that the proposed method performs satisfactorily in terms of all the three indices for up to 2.5% applied strain. Comparative results using simulation and experimental phantom data, and in vivo breast data of benign and malignant masses also demonstrate that the strain image quality of our method is better than the other reported techniques.

The number of projections is a critical factor in tomographic imaging. The larger the number, the better the quality of the reconstructed image; however, it increases the radiation dose delivered to the patient. Therefore, it is important to keep the number of projections as small as possible. Traditionally, the projections are taken by moving the x-ray source around the patient at uniform angular steps. Taking projections at nonuniform steps may result in better images as compared with that obtained using uniform projections. This paper describes two different approaches that adjust the step size to adaptively select the angle of projections. The first one is based on the spectral richness of the acquired projections and the second relies on the amount of new information added by successive projections. The superior performance of the two proposed methods over the uniform projection scheme is demonstrated through simulation results using both phantom and real images.

Elasticity imaging techniques with built-in or regularization-based smoothing feature for ensuring strain continuity are not intelligent enough to prevent distortion or lesion edge blurring while smoothing. This paper proposes a novel approach with built-in lesion edge preservation technique for high quality direct average strain imaging. An edge detection scheme, typically used in diffusion filtering is modified here for lesion edge detection. Based on the extracted edge information, lesion edges are preserved by modifying the strain determining cost function in the direct-average-strain-estimation (DASE) method. The proposed algorithm demonstrates approximately 3.42–4.25 dB improvement in terms of edge-mean-square-error (EMSE) than the other reported regularized or average strain estimation techniques in finite-element-modeling (FEM) simulation with almost no sacrifice in elastographic-signal-to-noise-ratio (SNRe) and elastographic-contrast-to-noise-ratio (CNRe) metrics. The efficacy of the proposed algorithm is also tested for the experimental phantom data and in vivo breast data. The results reveal that the proposed method can generate a high quality strain image delineating the lesion edge more clearly than the other reported strain estimation techniques that have been designed to ensure strain continuity. The computational cost, however, is little higher for the proposed method than the simpler DASE and considerably higher than that of the 2D analytic minimization (AM2D) method.

Ultrasound elastography is emerging with enormous potential as a medical imaging modality for effective discrimination of pathological changes in soft tissue. It maps the tissue elasticity or strain due to a mechanical deformation applied to it. The strain image most often calculated from the derivative of the local displacement field is highly noisy because of the de-correlation effect mainly due to unstable free-hand scanning and/or irregular tissue motion; consequently, improving the SNR of the strain image is still a challenging problem in this area. In this paper, a novel approach using the nearest-neighbor weighted least-squares is presented for direct estimation of the 'mean' axial strain for high quality strain imaging. Like other time/frequency domain reported schemes, the proposed method exploits the fact that the post-compression rf echo signal is a time-scaled and shifted replica of the pre-compression rf echo signal. However, the elegance of our technique is that it directly computes the mean strain without explicitly using any post filter and/or previous local displacement/strain estimates as is usually done in the conventional approaches. It is implemented in the short-time Fourier transform domain through a nearest-neighbor weighted least-squares-based Fourier spectrum equalization technique. As the local tissue strain is expected to maintain continuity with its neighbors, we show here that the mean strain at the interrogative window can be directly computed from the common stretching factor that minimizes a cost function derived from the exponentially weighted windowed pre- and post-compression rf echo segments in both the lateral and axial directions. The performance of our algorithm is verified for up to 8% applied strain using simulation and experimental phantom data and the results reveal that the SNR of the strain image can be significantly improved compared to other reported algorithms in the literature. The efficacy of the algorithm is also tested with in vivo breast data known to have malignant or benign masses from histology.

Background Denoising with edge preservation is very important in digital x-ray imaging since it may allow us to reduce x-ray dose in human subjects without noticeable degradation of the image quality. In denoising filter design for x-ray imaging, edge preservation as well as noise reduction is of great concern not to lose detailed spatial information for accurate diagnosis. In addition to this, fast computation is also important since digital x-ray images are mostly comprised of large sized matrices. Methods We have developed a new denoising filter based on the nonlinear diffusion filter model. Rather than employing four directional gradients around the pixel of interest, we use geometric parameters derived from the local pixel intensity distribution in calculating the diffusion coefficients in the horizontal and vertical directions. We have tested the filter performance, including edge preservation and noise reduction, using low dose digital radiography and micro-CT images. Results The proposed denoising filter shows performance similar to those of nonlinear anisotropic diffusion filters (ADFs), one Perona-Malik ADF and the other Weickert's ADF in terms of edge preservation and noise reduction. However, the computation time has been greatly reduced. Conclusions We expect the proposed denoising filter can be greatly used for fast noise reduction particularly in low-dose x-ray imaging.

Background Ventricular tachycardia (VT) and ventricular fibrillation (VF) are the most serious cardiac arrhythmias that require quick and accurate detection to save lives. Automated external defibrillators (AEDs) have been developed to recognize these severe cardiac arrhythmias using complex algorithms inside it and determine if an electric shock should in fact be delivered to reset the cardiac rhythm and restore spontaneous circulation. Improving AED safety and efficacy by devising new algorithms which can more accurately distinguish shockable from non-shockable rhythms is a requirement of the present-day because of their uses in public places. Method In this paper, we propose a sequential detection algorithm to separate these severe cardiac pathologies from other arrhythmias based on the mean absolute value of the signal, certain low-order intrinsic mode functions (IMFs) of the Empirical Mode Decomposition (EMD) analysis of the signal and a heart rate determination technique. First, we propose a direct waveform quantification based approach to separate VT plus VF from other arrhythmias. The quantification of the electrocardiographic waveforms is made by calculating the mean absolute value of the signal, called the mean signal strength. Then we use the IMFs, which have higher degree of similarity with the VF in comparison to VT, to separate VF from VTVF signals. At the last stage, a simple rate determination technique is used to calculate the heart rate of VT signals and the amplitude of the VF signals is measured to separate the coarse VF from VF. After these three stages of sequential detection procedure, we recognize the two components of shockable rhythms separately. Results The efficacy of the proposed algorithm has been verified and compared with other existing algorithms, e.g., HILB [1], PSR [2], SPEC [3], TCI [4], Count [5], using the MIT-BIH Arrhythmia Database, Creighton University Ventricular Tachyarrhythmia Database and MIT-BIH Malignant Ventricular Arrhythmia Database. Four quality parameters (e.g., sensitivity, specificity, positive predictivity, and accuracy) were calculated to ascertain the quality of the proposed and other comparing algorithms. Comparative results have been presented on the identification of VTVF, VF and shockable rhythms (VF + VT above 180 bpm). Conclusions The results show significantly improved performance of the proposed EMD-based novel method as compared to other reported techniques in detecting the life threatening cardiac arrhythmias from a set of large databases.

Ventricular fibrillation (VF) is a life-threatening cardiac arrhythmia. A high impulse current is required in this stage to save lives. In this paper, an empirical mode decomposition (EMD) based algorithm is presented to separate VF from other arrhythmias. The characteristics of the VF signal has high degree of similarity with the intrinsic mode functions (IMFs) of the EMD decomposition in comparison to other ECG pathologies. This high correlation between the VF signal and its certain IMFs is exploited to separate VF from other cardiac pathologies. Reliable databases are used to verify effectiveness of our algorithm and the results demonstrate superiority of our proposed technique compared to other well-known techniques of VF discrimination.

We propose a new elastography method based on x-ray imaging. After taking two x-ray tomographic images of the breast-mimicking phantom with applying different compressing pressure to it, we calculated displacement and strain maps from the two images using a non-rigid body image registration. The strain maps showed elasticity characteristics of the phantom medium. We expect that the proposed elastography method can be incorporated into breast tomosynthesis or breast CT systems to detect early stage breast cancers.

Magnetic steering of an untethered ferromagnetic device in a living body has many advantages in the clinical fields. In this paper, the positioning and tracking methods of a ferromagnetic sphere have been compared with magnetic resonance phantom images obtained with three different imaging sequences, spin echo, gradient echo, and selected positive contrast sequences. The position of the ferromagnetic sphere has been calculated from the MRI images which have susceptibility artifacts caused by the magnetic sphere. We have compared the positioning errors among the magnetic sphere images.

This study identifies differences in the electroencephalogram (EEG) responses caused by individual sensitivity to simulator sickness. Simulator sickness was investigated by studying the changes in simulator sickness in two different subject groups (sick group and nonsick group). Subjective evaluations using the simulator sickness questionnaire and the EEG response data were gathered every 5 min while the subjects were driving at 60 km/h for 60 min in the driving graphic simulator. The response to every item of the subjective evaluation increased linearly with time; the response level in the sick group was higher than in the nonsick group. The EEG analysis showed that the sick and nonsick groups were statistically significantly different with respect to the parameter theta/total at frontal lobe and parietal lobe.

Mesenchymal stromal cells (MSCs) are multipotent cells found in several adult tissues; they have the capacity to differentiate into mesodermal, ectodermal, and endodermal tissues in vitro. There have been several reports that MSCs have therapeutic effects in a variety of diseases. Therefore, using a cell labeling technique, monitoring their temporal and spatial migration in vivo, would be useful in the clinical setting. Magnetic resonance imaging (MRI)—tracking of superparamagnetic iron oxide (SPIO)-labeled cells—is a noninvasive technique for determining the location and migration of transplanted cells. In the present study, we evaluated the influence and toxicity of SPIO (ferumoxides) labeling on multiple differentiated MSCs. To evaluate the influence and toxicity of ferumoxides labeling on differentiation of MSCs, a variety of concentrations of ferumoxides were used for labeling MSCs. We found that the cytoplasm of adherent cells was effectively labeled at low concentrations of ferumoxides. Compared with unlabeled controls, the ferumoxides-labeled MSCs exhibited a similar proliferation rate and apoptotic progression. The labeled MSCs differentiated into osteoblasts and adipocytes in an identical fashion as the unlabeled cells. However, chondrogenesis and neurogenesis were inhibited at high concentrations of ferumoxides. Our results suggest the effective concentration for ferumoxides use in tracking MSCs.

fMRI-constrained EEG/MEG source imaging can be a powerful tool in studying human brain functions with enhanced spatial and temporal resolutions. Recent studies on the combination of fMRI and EEG/MEG have suggested that fMRI prior information could be readily implemented by simply imposing different weighting factors to cortical sources overlapping with the fMRI activations. It has been also reported, however, that such a hard constraint may cause severe distortions or elimination of meaningful EEG/MEG sources when there are distinct mismatches between the fMRI activations and the EEG/MEG sources. If one wants to obtain the actual EEG/MEG source locations and uses the fMRI prior information as just an auxiliary tool to enhance focality of the distributed EEG/MEG sources, it is reasonable to weaken the strength of fMRI constraint when severe mismatches between fMRI and EEG/MEG sources are observed. The present study suggests an efficient technique to automatically adjust the strength of fMRI constraint according to the mismatch level. The use of the proposed technique rarely affects the results of conventional fMRI-constrained EEG/MEG source imaging if no major mismatch between the two modalities is detected; while the new results become similar to those of typical EEG/MEG source imaging without fMRI constraint if the mismatch level is significant. A preliminary simulation study using realistic EEG signals demonstrated that the proposed technique can be a promising tool to selectively apply fMRI prior information to EEG/MEG source imaging.

We estimated ballistocardiogram (BCG) components in EEG signals recorded inside an MRI magnet using the electro-oculogram (EOG) signals recorded simultaneously with the EEG signals. Since the EOG signals are measured near the EEG measuring points, it is thought that the BCG components in the EOG signals resemble the BCG components in the EEG signals. To estimate the BCG components in the EEG signals, we applied the Kalman filter to the EOG and EEG signals recorded inside a 3.0 T MRI magnet. After removing the estimated BCG components from the EEG signals, we extracted the visual-evoked potentials (VEPs) from the BCG-removed EEG signals. To validate the efficacy of Kalman filtering in the BCG artifact removal, we have compared three types of VEPs of eight healthy subjects: one extracted from the raw EEG signals measured outside the magnet and the others extracted from the BCG-removed EEG signals measured inside the magnet. The BCG artifacts have been removed with Kalman filtering as well as with the conventional BCG template subtraction method for the sake of comparison. No significant difference in waveforms, latencies and amplitudes has been found between the two types of VEPs extracted from the two kinds of BCG-removed EEG signals.

Tinnitus is defined as an unwanted auditory perception of internal origin, usually localized, and rarely heard by others. Persisting appearances of tinnitus are most commonly combined with diseases or damage in the inner ear or neuro-auditory pathway. Diffusion tensor magnetic resonance imaging (DTI) is a new imaging method with the capability of providing non-invasive information on tissue microstructure not available in routine clinical MRI images. Since white matter regions of the brain are an ordered structure due to the myelination and directionality of axons and have a high degree of anisotropy, the ability to detect changes in anisotropy can be extremely useful in the study of diseases such as tinnitus and multiple sclerosis, which are assumed to involve the demyelination of axons. While several studies investigated tinnitus using MRI, few studies tried to analyze neurological disorders quantitatively using DTI. In this study, the cerebral volume of white matter on DTI images of patients with tinnitus was measured using the semi-automated and intuitive menu based image processing tool (Human Analyzer, ETRI, Korea). Total number of ten patients with tinnitus including three women was examined.

We present a new medical imaging technique for breast imaging, breast MREIT, in which magnetic resonance electrical impedance tomography (MREIT) is utilized to get high-resolution conductivity and current density images of the breast. In this work, we introduce the basic imaging setup of the breast MREIT technique with an investigation of four different imaging configurations of current-injection electrode positions and pathways through computer simulation studies. Utilizing the preliminary findings of a best breast MREIT configuration, additional numerical simulation studies have been carried out to validate breast MREIT at different levels of SNR. Finally, we have performed an experimental validation with a breast phantom on a 3.0 T MREIT system. The presented results strongly suggest that breast MREIT with careful imaging setups could be a potential imaging technique for human breast which may lead to early detection of breast cancer via improved differentiation of cancerous tissues in high-resolution conductivity images.

This study investigated what effect a 30% oxygen administration had on visuospatial cognitive performance and cerebral activation and lateralization using fMRI. Eight college students were selected as the subjects for this study. An oxygen dispenser that provided 21% and 30% oxygen at a constant rate of 8L/min was developed. In order to measure the performance level of visuospatial cognition, two psychological tests were also developed. The experiment consisted of two runs, one for a visuospatial cognition task with normal air (21% oxygen) and the other for a visuospatial cognition task with hyperoxic air (30% oxygen). Functional brain images were taken with a 3T MRI using the single-shot EPI method. The results of the visuospatial behavioral analysis reveal that accuracy rates were enhanced with 30% oxygen administration when compared to 21% oxygen. There were more activations observed at the bilateral occipital, parietal, and frontal lobes with 30% oxygen administration. However, decreased cerebrum lateralization was observed with 30% oxygen administration in the same regions compared with 21% oxygen administration. Thus, it is concluded that the positive effect on the visuospatial cognitive performance level by the highly concentrated oxygen administration resulted from an increase of cerebrum activation and a decrease of cerebrum lateralization.

The purpose of this study was to assess clinical 1H MR spectroscopy (MRS) as a noninvasive method for evaluating brain tumor malignancy at 3T high-field system. Using 3T MRI/MRS system, localized water-suppressed single-voxel technique in patients with brain tumor (i.e., gliomas) was employed to evaluate spectra with peaks of N-acetyl aspartate (NAA), choline-containing compounds (Cho), creatine/phosphocreatine (Cr) and lactate. On the basis of Cr, these peak areas were quantitated as a relative ratio. The variation of metabolite measurements of the designated region in 10 normal volunteers was less than 10%. Normal ranges of NAA/Cr and Cho/Cr ratios were 1.67+/-018 and 1.16+/-0.15, respectively. NAA/Cr ratio of gliomas was significantly lower than that of the normal tissues (P= .005), but Cho/Cr ratio of gliomas was significantly higher (P= .001). Cho/Cr ratio of high-grade gliomas was significantly higher than that of low-grade gliomas. The present study demonstrated that the neuronal degradation or loss was observed in all gliomas. Higher-grade glioma was correlated with higher Cho/Cr ratio, indicating a significant dependence of Cho levels on malignancy of gliomas. Our results suggest that clinical 1H MR spectroscopy could be useful to predict tumor malignancy.

The present study was performed to determine the characteristics of the biochemical metabolites related to gastric cancer using ex vivo (1)H magnetic resonance spectroscopy (MRS), and to assess the clinical usefulness. A total of 35 gastric specimens resected during surgery for gastric cancer were used to compare MR spectra. A 1.5-T (64-MHz) clinical MR imager equipped with facilities for spectroscopy was used to obtain MR spectra from 33 gastric specimens. High-resolution (1)H nuclear magnetic resonance (NMR) spectra of the remains of two specimens were also examined with a 9.4-T (400-MHz) NMR spectrometer. Localized spectroscopic measurements were performed in two layers of gastric tissue, the proper muscle layer and the composite mucosa/submucosa layer. T(2) FSE and 3D SPGR images were used to determine the voxel size and the location for MRS data collection. MR spectra were obtained using the single-voxel PRESS technique with parameters of TR/TE = 2000/30 ms, NA = 256, and voxel size = 3 x 3 x 3 mm(3) (27 microL). Cancerous and noncancerous gastric tissues in the voxel were determined by histopathological analysis. On 9.4-T ex vivo NMR spectroscopy, the following metabolite peaks were found: lipids at 0.9 ppm (CH(3)) and 1.3 ppm (CH(2)); alanine (beta-CH(3)) at 1.58 ppm; N-Acetyl neuraminic acid (NANA: sialic acid) at 2.03 ppm; and glutathione at 2.25 ppm in normal gastric tissue layers. In the 1.5-T MR system, broad and featureless spectral peaks of the various metabolites in normal human gastric tissue were observed at 0.9 ppm, 1.3 ppm, 2.0 ppm, and 2.2 ppm regardless of gastric tissue layer. In specimens (Borrmann type III) with tubular adenocarcinoma, resonance peaks were observed at 1.26 ppm, 1.36 ppm (doublet of lactate), and 3.22 ppm (choline). Cancer lesions showed decreased levels of lipid peaks, showing the significant lactate doublet peaks, and increased intensity of the choline peak as compared with noncancerous gastric tissue. We found that decreased levels of lipids and increases in lactate and choline peaks in gastric tissue were markers for malignancy in gastric lesions. Information provided by ex vivo (1)H MRS, together with the development of in vivo (1)H MRS with high field strength and high resolution, may be very useful for the diagnosis of gastric cancer in clinical situation.

Magnetic resonance electrical impedance tomography (MREIT) is to provide cross-sectional images of the conductivity distribution sigma of a subject. While injecting current into the subject, we measure one component Bz of the induced magnetic flux density B = (Bx, By, Bz) using an MRI scanner. Based on the relation between (inverted delta)2 Bz and inverted delta sigma, the harmonic Bz algorithm reconstructs an image of sigma using the measured Bz data from multiple imaging slices. After we obtain sigma, we can reconstruct images of current density distributions for any given current injection method. Following the description of the harmonic Bz algorithm, this paper presents reconstructed conductivity and current density images from computer simulations and phantom experiments using four recessed electrodes injecting six different currents of 26 mA. For experimental results, we used a three-dimensional saline phantom with two polyacrylamide objects inside. We used our 0.3 T (tesla) experimental MRI scanner to measure the induced Bz. Using the harmonic Bz algorithm, we could reconstruct conductivity and current density images with 82 x 82 pixels. The pixel size was 0.6 x 0.6 mm2. The relative L2 errors of the reconstructed images were between 13.8 and 21.5% when the signal-to-noise ratio (SNR) of the corresponding MR magnitude images was about 30. The results suggest that in vitro and in vivo experimental studies with animal subjects are feasible. Further studies are requested to reduce the amount of injection current down to less than 1 mA for human subjects.