Evaluation of calculation processes of apparent diffusion coefficient subtraction method (ASM) imaging.

A number of restricted diffusion (RD) imaging techniques, such as diffusion kurtosis (DK) imaging and Q space imaging, have been developed and proven to be useful for the diagnosis of diseases, including cerebral gliomas and cerebrovascular infarction. In particular, apparent diffusion coefficient (ADC) subtraction method (ASM) imaging has become available recently as a novel RD imaging technique. ASM is based on the difference between the ADC values in an image pair of two ADC maps, ADC basic (ADCb) and ADC modify (ADCm), which are created from diffusion-weighted images taken using short and long effective diffusion times, respectively. The present study aimed to assess the potential of different types of ASM imaging by comparing them with DK imaging which is the gold-standard RD imaging technique. In the present basic study using both polyethylene glycol phantom and cell-containing bio-phantom, three different types of ASM images were created using different calculation processes. ASM/A is an image calculated by dividing the absolute difference between ADCb and ADCm by ADCb several times. By contrast, ASM/S is an image created by dividing the absolute difference between ADCb and ADCm by the standard deviation of ADCb several times. As for positive ASM/A image (PASM/A), the positive image, which was resultant after subtracting ADCb from ADCm, was divided by ADCb several times. A comparison was made between the types of ASM and DK images. The results showed the same tendency between ASM/A in addition to both ASM/S and PASM/A. By increasing the number of divisions by ADCb from three to five times, ASM/A images transformed from DK-mimicking to more RD-sensitive images compared with DK images. These observations suggest that ASM/A images may prove useful for future clinical applications in RD imaging protocols for the diagnosis of diseases.

Evaluation of Fast Diffusion Kurtosis Imaging Using New Software Designed for Widespread Clinical Use.

Clinical research using restricted diffusion-weighted imaging, especially diffusion kurtosis (DK) imaging, has been progressing, with reports on its effectiveness in the diagnostic imaging of cerebral infarctions, neurodegenerative diseases, and tumors, among others. However, the application of DK imaging in daily clinical practice has not spread because of the long imaging time required and the use of specific software for image creation. Herein, with the aim of promoting clinical research using DK imaging at any medical facility, we evaluated fast DK imaging using a new software program. We developed a new macro program that produces DK images using general-purpose, inexpensive software (Microsoft Excel and ImageJ), and we evaluated fast DK imaging using bio-phantoms and a healthy volunteer in clinical trials. The DK images created by the new software with diffusion-weighted images captured with short-time imaging sequences were similar to the original DK images captured with long-time imaging sequences. The DK images using three b-values, which can reduce the imaging time by 43%, were equivalent to the DK images using five b-values. The DK imaging technique developed herein might allow any medical facility to increase its daily clinical use of DK imaging and easily conduct clinical research.

Evaluation of the Imaging Process for a Novel Subtraction Method Using Apparent Diffusion Coefficient Values.

Diffusion-weighted imaging may be used to obtain the apparent diffusion coefficient (ADC), which aids the diagnosis of cerebral infarction and tumors. An ADC reflects elements of free diffusion. Diffusion kurtosis imaging (DKI) has attracted attention as a restricted diffusion imaging technique. The ADC subtraction method (ASM) was developed to visualize restricted diffusion with high resolution by using two ADC maps taken with different diffusion times. We conducted the present study to provide a bridge between the reported basic ASM research and clinical research. We developed new imaging software for clinical use and evaluated its performance herein. This software performs the imaging process automatically and continuously at the pixel level, using ImageJ software. The new software uses a macro or a plugin which is compatible with various operating systems via a Java Virtual Machine. We tested the new imaging software's performance by using a Jurkat cell bio-phantom, and the statistical evaluation of the performance clarified that the ASM values of 99.98% of the pixels in the bio-phantom and physiological saline were calculated accurately (p<0.001). The new software may serve as a useful tool for future clinical applications and restricted diffusion imaging research.

Development of a novel phantom using polyethylene glycol for the visualization of restricted diffusion in diffusion kurtosis imaging and apparent diffusion coefficient subtraction method.

The present study aimed to investigate whether polyethylene glycol (PEG) phantoms have the potential to be used as standard phantoms for magnetic resonance imaging (MRI) in order to visualize restricted diffusion in diffusion kurtosis imaging (DKI), the ADC subtraction method (ASM) and the apparent diffusion coefficient (ADC). Diffusion-weighted images of 0-120 mM PEG phantoms were captured to create ADC, DKI and ASM images with post-processing. ASM is a recently developed method for restricted diffusion imaging using the readout segmentation of long variable echo-train sequences. As the PEG concentration increases, the ADC value decreases. Conversely, an increase in DKI and ASM values is associated with increasing PEG concentration. Formulae were constructed to represent the association between PEG concentrations and ADC, DKI and ASM values. These formulae can be used to determine the required PEG concentrations to mimic arbitrary ADC, DKI and ASM values of certain diseases, including tumors and infarctions. Validation experiments were conducted using bio-phantoms and clarified that the PEG phantoms cover the range of ADC and DKI values reported in previous clinical research using 3T MRI. PEG phantoms may be useful for future MRI research involving restricted diffusion.

Development of a novel method for visualizing restricted diffusion using subtraction of apparent diffusion coefficient values.

In order to visualize restricted diffusion, the present study developed a novel method called 'apparent diffusion coefficient (ADC) subtraction method (ASM)' and compared it with diffusion kurtosis imaging (DKI). The diffusion-weighted images of physiological saline, in addtion to bio-phatoms of low cell density and the highest cell density were obtained using two sequences with different effective diffusion times. Then, the calculated ADC values were subtracted. The mean values and standard deviations (SD) of the ADC values of physiological saline, low cell density and the highest cell density phantoms were 2.95±0.08x10­3, 1.90±0.35x10­3 and 0.79±0.05x10­3 mm2/sec, respectively. The mean kurtosis values and SD of DKI were 0.04±0.01, 0.44±0.13 and 1.27±0.03, respectively. The ASM and SD values were 0.25±0.20x104, 0.51±0.41x104 and 4.80±4.51x104 (sec/mm2)2, respectively. Using bio­phantoms, the present study demonstrated that DKI exhibits restricted diffusion in the extracellular space. Similarly, ASM may reflect the extent of restricted diffusion in the extracellular space.

[Improvement of Brain Contrast Using Spin Echo T1-weighted Image at 3 T MRI].

Brain T1-weighted images using spin echo (SE) sequence has poor contrast at 3.0 Tesla magnetic resonance imaging (3.0 T MRI) systems from the influence of crosstalk and magnetized transfer (MT) effect, and prolongation of the T1 value. Therefore, improving of scan parameters has been reported such as excitation flip angle (FA) and interleave data acquisition. The purpose of this study was to show the effects of alterations of presaturation pulse amplitude and chemical shift selective (CHESS) pulse amplitude. Gray-to-white matter contrast increased with decreasing amplitude of presaturation pulse in whole brain imaging. Presaturation and CHESS pulse consist of radio frequency pulse. Therefore, both pulses have a similar effect on MT pulse. Manual alteration of presaturation pulse amplitude for each scan lacks versatility on clinical use. However, decreasing amplitude of presaturation pulse is equal to decreasing thickness of presaturation pulse. About CHESS pulse, it requires no manual alteration for each scan. For example, switching fat suppression mode from strong to weak increase T1 contrast. Our study demonstrated that using not only low excitation FA and interleave date acquisition but also low amplitude of presaturation and CHESS pulse increase the contrast in T1 SE brain scans at 3.0 T MRI.

The Usefulness of Readout-Segmented Echo-Planar Imaging (RESOLVE) for Bio-phantom Imaging Using 3-Tesla Clinical MRI.

Readout-segmented echo-planar imaging (RESOLVE) is a multi-shot echo-planar imaging (EPI) modality with k-space segmented in the readout direction. We investigated whether RESOLVE decreases the distortion and artifact in the phase direction and increases the signal-to-noise ratio (SNR) in phantoms image taken with 3-tesla (3T) MRI versus conventional EPI. We used a physiological saline phantom and subtraction mapping and observed that RESOLVE's SNR was higher than EPI's. Using RESOLVE, the combination of a special-purpose coil and a large-loop coil had a higher SNR compared to using only a head/neck coil. RESOLVE's image distortioas less than EPI's. We used a 120 mM polyethylene glycol phantom to examine the phase direction artifact.vThe range where the artifact appeared in the apparent diffusion coefficient (ADC) image was shorter with RESOLVE compared to EPI. We used RESOLVE to take images of a Jurkat cell bio-phantom: the cell-region ADC was 856×10-6mm2/sec and the surrounding physiological saline-region ADC was 2,951×10-6mm2/sec. The combination of RESOLVE and the 3T clinical MRI device reduced image distortion and improved SNR and the identification of accurate ADC values due to the phase direction artifact reduction. This combination is useful for obtaining accurate ADC values of bio-phantoms.

Neural Responses to Central and Peripheral Objects in the Lateral Occipital Cortex.

Human object recognition and classification depend on the retinal location where the object is presented and decrease as eccentricity increases. The lateral occipital complex (LOC) is thought to be preferentially involved in the processing of objects, and its neural responses exhibit category biases to objects presented in the central visual field. However, the nature of LOC neural responses to central and peripheral objects remains largely unclear. In the present study, we used functional magnetic resonance imaging (fMRI) and a wide-view presentation system to investigate neural responses to four categories of objects (faces, houses, animals, and cars) in the primary visual cortex (V1) and the lateral visual cortex, including the LOC and the retinotopic areas LO-1 and LO-2. In these regions, the neural responses to objects decreased as the distance between the location of presentation and center fixation increased, which is consistent with the diminished perceptual ability that was found for peripherally presented images. The LOC and LO-2 exhibited significantly positive neural responses to all eccentricities (0-55°), but LO-1 exhibited significantly positive responses only to central eccentricities (0-22°). By measuring the ratio relative to V1 (RRV1), we further demonstrated that eccentricity, category and the interaction between them significantly affected neural processing in these regions. LOC, LO-1, and LO-2 exhibited larger RRV1s when stimuli were presented at an eccentricity of 0° compared to when they were presented at the greater eccentricities. In LOC and LO-2, the RRV1s for images of faces, animals and cars showed an increasing trend when the images were presented at eccentricities of 11 to 33°. However, the RRV1s for houses showed a decreasing trend in LO-1 and no difference in the LOC and LO-2. We hypothesize, that when houses and the images in the other categories were presented in the peripheral visual field, they were processed via different strategies in the lateral visual cortex.

Retinotopy and attention to the face and house images in the human visual cortex.

Attentional modulation of the neural activities in human visual areas has been well demonstrated. However, the retinotopic activities that are driven by face and house images and attention to face and house images remain unknown. In the present study, we used images of faces and houses to estimate the retinotopic activities that were driven by both the images and attention to the images, driven by attention to the images, and driven by the images. Generally, our results show that both face and house images produced similar retinotopic activities in visual areas, which were only observed in the attention + stimulus and the attention conditions, but not in the stimulus condition. The fusiform face area (FFA) responded to faces that were presented on the horizontal meridian, whereas parahippocampal place area (PPA) rarely responded to house at any visual field. We further analyzed the amplitudes of the neural responses to the target wedge. In V1, V2, V3, V3A, lateral occipital area 1 (LO-1), and hV4, the neural responses to the attended target wedge were significantly greater than those to the unattended target wedge. However, in LO-2, ventral occipital areas 1 and 2 (VO-1 and VO-2) and FFA and PPA, the differences were not significant. We proposed that these areas likely have large fields of attentional modulation for face and house images and exhibit responses to both the target wedge and the background stimuli. In addition, we proposed that the absence of retinotopic activity in the stimulus condition might imply no perceived difference between the target wedge and the background stimuli.

Decreased resting-state connections within the visuospatial attention-related network in advanced aging.

Advanced aging is accompanied by a decline in visuospatial attention. Previous neuroimaging and electrophysiological studies have demonstrated dysfunction in specific brain areas related to visuospatial attention. However, it is still unclear how the functional connectivity between brain regions causes the decline of visuospatial attention. Here, we combined task and rest functional magnetic resonance imaging (fMRI) to investigate the age-dependent alterations of resting-state functional connectivity within the task-related network. Twenty-three young subjects and nineteen elderly subjects participated in this study, and a modified Posner paradigm was used to define the region of interest (ROI). Our results showed that a marked reduction in the number of connections occurred with age, but this effect was not uniform throughout the brain: while there was a significant loss of communication in the anterior portion of the brain and between the anterior and posterior cerebral cortices, communication in the posterior portion of the brain was preserved. Moreover, the older adults exhibited weakened resting-state functional connectivity between the supplementary motor area and left anterior insular cortex. These findings suggest that, the disrupted functional connectivity of the brain network for visuospatial attention that occurs during normal aging may underlie the decline in cognitive performance.

A new phantom and empirical formula for apparent diffusion coefficient measurement by a 3 Tesla magnetic resonance imaging scanner.

The aim of this study was to create a new phantom for a 3 Tesla (3T) magnetic resonance imaging (MRI) device for the calculation of the apparent diffusion coefficient (ADC) using diffusion-weighted imaging (DWI), and to mimic the ADC values of normal and tumor tissues at various temperatures, including the physiological body temperature of 37°C. The phantom was produced using several concentrations of sucrose from 0 to 1.2 M, and the DWI was performed using various phantom temperatures. The accurate ADC values were calculated using the DWIs of the phantoms, and an empirical formula was developed to calculate the ADC values of the phantoms from an arbitrary sucrose concentration and arbitrary phantom temperature. The empirical formula was able to produce ADC values ranging between 0.33 and 3.02×10-3 mm2/sec, which covered the range of ADC values of the human body that have been measured clinically by 3T MRI in previous studies. The phantom and empirical formula developed in this study may be available to mimic the ADC values of the clinical human lesion by 3T MRI.

In vitro assessment of factors affecting the apparent diffusion coefficient of Jurkat cells using bio-phantoms.

It is well known that many tumor tissues show lower apparent diffusion coefficient (ADC) values, and that several factors are involved in the reduction of ADC values. The aim of this study was to clarify how much each factor contributes to decreases in ADC values. We investigate the roles of cell density, extracellular space, intracellular factors, apoptosis and necrosis in ADC values using bio-phantoms. The ADC values of bio-phantoms, in which Jurkat cells were encapsulated by gellan gum, were measured by a 1.5-Tesla magnetic resonance imaging device with constant diffusion time of 30sec. Heating at 42℃ was used to induce apoptosis while heating at 48℃ was used to induce necrosis. Cell death after heating was evaluated by flow cytometric analysis and electron microscopy. The ADC values of bio-phantoms including non-heated cells decreased linearly with increases in cell density, and showed a steep decline when the distance between cells became less than 3µm. The analysis of ADC values of cells after destruction of cellular structures by sonication suggested that approximately two-thirds of the ADC values of cells originate from their cellular structures. The ADC values of bio-phantoms including necrotic cells increased while those including apoptotic cells decreased. This study quantitatively clarified the role of the cellular factors and the extracellular space in determining the ADC values produced by tumor cells. The intermediate diffusion time of 30msec might be optimal to distinguish between apoptosis and necrosis.

Regional neural response differences in the determination of faces or houses positioned in a wide visual field.

In human visual cortex, the primary visual cortex (V1) is considered to be essential for visual information processing; the fusiform face area (FFA) and parahippocampal place area (PPA) are considered as face-selective region and places-selective region, respectively. Recently, a functional magnetic resonance imaging (fMRI) study showed that the neural activity ratios between V1 and FFA were constant as eccentricities increasing in central visual field. However, in wide visual field, the neural activity relationships between V1 and FFA or V1 and PPA are still unclear. In this work, using fMRI and wide-view present system, we tried to address this issue by measuring neural activities in V1, FFA and PPA for the images of faces and houses aligning in 4 eccentricities and 4 meridians. Then, we further calculated ratio relative to V1 (RRV1) as comparing the neural responses amplitudes in FFA or PPA with those in V1. We found V1, FFA, and PPA showed significant different neural activities to faces and houses in 3 dimensions of eccentricity, meridian, and region. Most importantly, the RRV1s in FFA and PPA also exhibited significant differences in 3 dimensions. In the dimension of eccentricity, both FFA and PPA showed smaller RRV1s at central position than those at peripheral positions. In meridian dimension, both FFA and PPA showed larger RRV1s at upper vertical positions than those at lower vertical positions. In the dimension of region, FFA had larger RRV1s than PPA. We proposed that these differential RRV1s indicated FFA and PPA might have different processing strategies for encoding the wide field visual information from V1. These different processing strategies might depend on the retinal position at which faces or houses are typically observed in daily life. We posited a role of experience in shaping the information processing strategies in the ventral visual cortex.

Influence of permittivity and electrical conductivity on image pattern of MRI.

In proton density-weighted (PDW) MR imaging, the patterns of signal intensity vary depending on the imaged material, and change with the flip angle (FA) applied to the imaged material. The correlation between the pre-determined FA and the actual FA applied to imaged objects was investigated using 4 types of phantoms having different dielectric properties. PDW images were acquired using the spin-echo (SE) method and different pre-determined FA. Dependency of the signal intensity distribution in the phantom on the pre-determined FA differed among phantoms: patterns for water and 0.402 w/w% saline solution phantoms changed with the pre-determined FA, whereas those for olive oil and 4.02 w/w% saline solution phantoms were barely affected by the pre-determined FA. Causes of these phenomena were considered to be the differences between the pre-determined FA and the actual FA among the phantoms; differences were also influenced by the positioning of the phantom. Our study showed that the actual FA in the phantom is greater than the pre-determined FA in high permittivity media, whereas it is reduced by an increased conductivity of the media.

Development of MRI phantom equivalent to human tissues for 3.0-T MRI.

PURPOSE: A 3.0-T MRI phantom (called the CAGN-3.0T phantom) having human-equivalent relaxation times and human-equivalent conductivity was developed. METHODS: The ingredients of the phantom are carrageenan (as a gelatinizer), agarose (as a T2-relaxation modifier), GdCl3 (as a T1-relaxation modifier), NaCl (as a conductivity modifier), and NaN3 (as an antiseptic). Numerous samples with varying concentrations of agarose, GdCl3, and NaCl were prepared, and T1 and T2 values were measured using 3.0-T MRI. RESULTS: The T1 values of the CAGN-3.0T phantom were unaffected by NaCl, while the T2 values were only slightly affected. Based on the measured data, empirical formulae were devised to express the relationships between the concentrations of agarose, GdCl3, and NaCl and the relaxation times. The formula for expressing the conductivity of the CAGN-3.0T phantom was obtained. CONCLUSIONS: By adjustments to the concentrations of agarose, GdCl3, and NaCl, the relaxation times and conductivity of almost all types of human tissues can be simulated by CAGN-3.0T phantoms. The phantoms have T1 values of 395-2601 ms, T2 values of 29-334 ms, and conductivity of 0.27-1.26 S/m when concentrations of agarose, GdCl3, and NaCl are varied from 0 to 2.0 w/w%, 0 to 180 µmol/kg, and 0 to 0.7 w/w%, respectively. The CAGN-3.0T phantom has sufficient strength to replicate the torso without using reinforcing agents, and can be cut with a knife into any shape.

Development of a method to present wide-view visual stimuli in MRI for peripheral visual studies.

We developed a novel wide-view visual presentation system for fMRI studies. Computer-generated images were projected onto a hemispheric, translucent screen inside the MRI bore and were then back-projected onto a 52mm diameter screen. To achieve a wide field view, a spherical screen with a curvature radius of 30mm was placed 30mm away from the subjects' eyes. The subjects wore contact lenses that enabled them to focus on the screen, and the resulting visual field reached 120°. To evaluate the clarity and quality of the MRI images, a signal-to-noise ratio valuation experiment was performed. In addition, we successfully applied this visual presentation system to studies of visual retinotopic mapping and object perception neural function in the peripheral visual field. Our study demonstrated that the system is compatible with the MRI environment. Based on the wide-field mapping results, this system was more effective at mapping a checkerboard stimuli in V1-V3 from the central to peripheral visual fields. In higher-level visual areas, we successfully located several classical category-selective areas, including the face-selective area (FFA), occipital face area (OFA), house-selective area (PPA), transverse occipital sulcus (TOS), lateral occipital complex (LOC) and posterior fusiform area (pFs). In these areas, we found that the response amplitudes exhibited different decreasing trends with increasing eccentricity. In conclusion, we developed a simple, effective method for presenting wide-view visual stimuli within the MRI environment that can be applied to many kinds of fMRI studies of peripheral vision.

Visualization of stent lumen in MR imaging: relationship with stent design and RF direction.

Magnetic resonance imaging (MRI) visualization of metallic stent lumens is possible if the stent structure counteracts eddy currents in the lumen induced by the radio frequency magnetic field, B(1). To examine the effectiveness of various stent designs in counteracting eddy currents, we anchored eight copper stent models and 2 commercially available nickel-titanium alloy (Nitinol) stents in a gel phantom, perpendicular or parallel to the direction of B(1). A mesh stent lumen showed hypointensity irrespective of its alignment relative to B(1). A solenoid stent lumen showed hypointensity with the stent axis parallel to B(1), but it had the same signal intensity as outside the lumen when perpendicular to B(1). A Moebius stent lumen showed no signal reduction, irrespective of alignment relative to B(1). Lumens of the commercially available stents showed hypointensity regardless of alignment relative to B(1). Computer simulation revealed that the signal intensities of the stents corresponded to magnetic flux densities of B(1) in the stents, which are modified by the structure of the stent. While in vivo MRI viewing of a Moebius stent lumen is likely possible regardless of axis alignment, inherent structural weakness may be problematic. As a more practical choice, the solenoid stent is easier to manufacture and generates no hypointensive signal when the axis is parallel to B(0).

In vitro assessment of factors affecting the apparent diffusion coefficient of Ramos cells using bio-phantoms.

The roles of cell density, extracellular space, intracellular factors, and apoptosis induced by the molecularly targeted drug rituximab on the apparent diffusion coefficient (ADC) values were investigated using bio-phantoms. In these bio-phantoms, Ramos cells (a human Burkitt's lymphoma cell line) were encapsulated in gellan gum. The ADC values decreased linearly with the increase in cell density, and declined steeply when the extracellular space became less than 4 µm. The analysis of ADC values after destruction of the cellular membrane by sonication indicated that approximately 65% of the ADC values of normal cells originate from the cell structures made of membranes and that the remaining 35% originate from intracellular components. Microparticles, defined as particles smaller than the normal cells, increased in number after rituximab treatments, migrated to the extracellular space and significantly decreased the ADC values of bio-phantoms during apoptosis. An in vitro study using bio-phantoms was conducted to quantitatively clarify the roles of cellular factors and of extracellular space in determining the ADC values yielded by tumor cells and the mechanism by which apoptosis changes those values.

Development of a human-tissue-like phantom for 3.0-T MRI.

PURPOSE: A 3.0-T MRI phantom having human-tissue-equivalent relaxation times was developed. METHODS: The ingredients of the phantom are carrageenan (for gelatinization), GdCl(3) (as a T(1)-relaxation modifier), agarose (as a T(2)-relaxation modifier), and NaN(3) (as an antiseptic agent). Numerous samples with varying concentrations of GdCl(3) and agarose were prepared, and T(1) and T(2) were measured using 3.0-T MRI. RESULTS: Relaxation times of the phantom samples ranged from 395 to 2601 ms for T(1) values and 29 to 334 ms for T(2) values. Based on the measured results, empirical formulae were devised to express the relationships between the concentrations of relaxation modifiers and relaxation times. CONCLUSIONS: Adjustment of GdCl(3) and agarose concentrations allows arbitrary setting of relaxation times, and the creation of a phantom that can mimic relaxation times of human-tissue. Carrageenan is considered the most suitable as a gelling agent for an MRI phantom, as it permits the relatively easy and inexpensive production of a large phantom such as for the human torso, and which can be easily shaped with a knife.

Production of a human-tissue-equivalent MRI phantom: optimization of material heating.

PURPOSE: We conceived a 2-stage heating method to dissolve the ingredients of magnetic resonance (MR) imaging phantoms to overcome issues of uneven quality in conventional MR imaging phantoms, and we evaluated uniformity and the reproducibility of our method. METHODS: We used a 3-liter capacity, column-shaped, enamel-coated porcelain container to produce a muscle-equivalent phantom (diameter, 160 mm; height, 100 mm; volume, 2 liters). The phantom contained: 1) carrageenan as a gelling agent; 2) agarose as a T2 modifier; 3) GdCl3 as a T1 modifier; 4) NaN3 as an antiseptic; and 5) distilled water. We applied both direct heating and 2-stage heating of pre-soaked materials. We placed powdered materials directly into hot water for direct heating but soaked them in water one day before use (post-swelling) in 2-stage heating. The materials in the container were melted in a silicone oil bath of 120 or 140 degrees C under various conditions, then allowed to gel by natural cooling. We observed the resulting gel phantoms macroscopically using a CCD camera and evaluated their uniformity by microscopy and MR imaging. RESULTS: We found it necessary to raise the temperature inside the phantom to 100.0 degrees C, to produce a uniform gel with stable homogeneity and few bubbles. Use of an enamel-coated porcelain container required setting the temperature of the oil bath at 140 degrees C. CONCLUSION: A uniform and reproducible human tissue-equivalent phantom with few bubbles can be manufactured using our 2-stage heating method, which employs pre-soaking in a silicone oil bath at 140 degrees C for 30 min. We then added the swollen carrageenan to the agarose solution, which heating the temperature to 140 degrees C for 30 min while continuously stirring at 120 rpm, following with natural cooling.