Biomedical subjects
Publications and source records attributed to S J Riederer.
We present a statistical analysis of the maximum intensity projection (MIP) algorithm, which is commonly used for MR angiography (MRA). The analysis explains why MIP projection images display as much as a twofold increase in signal- and contrast-to-noise ratios over those of the source image set. This behavior is demonstrated with simulations and in phantom and MRA image sets.
Despite the fact that respiratory motion is a major factor limiting the image quality of MR examinations in the upper abdomen, little quantitative information is available about the kinematics of visceral motion during respiration. The objective of this study was to obtain a measure of the relative longitudinal and transverse displacements of the upper abdominal organs during breathing using an MR line scan technique.
The incorporation of the phase offset multi planar (POMP) technique into breathheld magnetization-prepared gradient echo imaging is discussed as a means for improving imaging efficiency without sacrificing resolution, contrast, or SNR improvement. The phase encoding order necessary to preserve the centric approach is described. When combined with interleaving, the POMP technique enables four 256 x 256 images to be acquired in a 12-s breathhold, doubling the efficiency of the original technique. This scan efficiency is compared with that of other T1-weighted 2DFT methods.
A technique for estimation of the longitudinal relaxation time of a large homogeneous object with an acquisition time of 4 s or less was developed by combining spatially selective rf tagging pulses with a T1-weighted magnetization-prepared gradient-echo sequence. Multiple 5-mm-wide tagged areas are laid orthogonal to the imaging section of interest. The contrast between each tag and the untagged regions differs because each tag is produced at a different time. The T1 value is determined from the nulling time at which tagged and untagged areas have no contrast.
Two hypotheses were tested: (a) that view-to-view variations in bulk phase and modulus of magnetization in vascular volume elements can indicate the presence of disordered blood flow, and (b) that a substantial loss of signal intensity on magnetic resonance (MR) angiograms of poststenotic regions is due to view-to-view changes in magnetization. To test these hypotheses, a technique was developed in which view-to-view variations in transverse magnetization were used to create angiographic projection images, which showed only disordered flow (disordered flow maps) in vitro and in vivo. In phantom studies, this technique recovered signal intensity downstream from stenoses. A combination of disordered flow maps with morphologic images improved visualization of stenotic regions and provided information on characteristics of local flow. These results show that view-to-view variations in transverse magnetization occur in regions of disordered flow and are an important cause of loss of signal intensity. This technique can provide information about dynamic blood flow and improve depiction of anatomic structures on MR angiograms.
It has been shown that magnetic resonance image data can be corrected for the effects of motion by using retrospective adaptive techniques which employ navigator (NAV) echoes. We demonstrate the accuracy with which NAV echoes can measure motion, as well as the independent nature of the respective view-to-view and intraview corrections.
Retrospective correction of magnetic resonance (MR) image data to eliminate the effects of patient motion is possible with use of adaptive correction techniques. These methods require an accurate record of the motion that occurs during imaging. The authors evaluated whether motion information suitable for adaptive correction could be obtained from phase-encoded image data alone rather than from separate navigator echoes. Once such displacements were estimated from the image data, motion correction proceeded with use of the same algorithm used for the navigator echoes. The results show that image data alone can be used to effectively measure view-to-view displacements in phantoms, but external markers are required for accurate measurement during axial head imaging of patients.
Use of intraacquisition modification of pulse-sequence parameters to reduce acquisition time for conventional T2-weighted spin-echo images was evaluated. With this technique (variable-rate spin-echo pulse sequence), the repetition time and echo time (TR msec/TE msec) were reduced during imaging as a function of the phase-encoding view. To maintain T2-based contrast, TR and TE for the low-spatial-frequency views were left at their prescribed values (eg, 2,000/80). TR and TE for the high-spatial-frequency views were progressively reduced during imaging (eg, to 1,000/20). Acquisition time was reduced by as much as 25%. In one pulse sequence, the duration of multisection imaging nominally performed at TR 2,000 and with 256 phase-encoding views was reduced from 9 minutes 30 seconds to 6 minutes 30 seconds. In all sequences, edges and small structures were enhanced, and T2 contrast was somewhat decreased in high spatial frequencies. Filtering of the raw data before reconstruction can suppress these effects and provide a net increase in contrast-to-noise ratio.
Magnetization-prepared ultrashort-repetition-time (snapshot) gradient-echo imaging is a technique of magnetic resonance (MR) imaging with many potential applications. In the application of this technique to abdominal imaging, the effects on contrast of phase-encoding order, resolution, preparation-phase inversion time, and data-acquisition flip angle were predicted and then demonstrated with images obtained in examinations of 22 patients. In the analysis of 36 liver lesions, snapshot images were compared with corresponding T1-weighted spin-echo images on the basis of signal-to-noise ratio (S/N) of liver and contrast-to-noise ratio (C/N) between liver and lesion. Snapshot MR imaging produced abdominal images with 192 (or 256) x 256 resolution, negligible motion artifact, and C/N 1.29 times (+/- 0.48) higher than that in T1-weighted spin-echo imaging. Acquisition times were 13 seconds or less, short enough for imaging during suspended respiration. Also, use of a phased-array multicoil further improves the S/N in snapshot images without acquisition-time penalty.
Real-time interactive color flow magnetic resonance (MR) imaging is a combination of real-time MR imaging and color encoding of velocity-induced phase angle. Flow-compensated (FC) and flow-encoded (FE) images are acquired continuously by using gradient echoes and a 12-msec repetition time. Each image is reconstructed within 200 msec of acquisition, and the FC magnitude image is displayed in gray-scale format. The phase difference between the reconstructed FC and FE images, a difference proportional to velocity along the flow-encoding direction, is encoded in color and superimposed on the gray-scale FC image. Magnitude and phase information are thus presented simultaneously. The viewer may interactively adjust many acquisition parameters during data acquisition. Experimental results of phantom and in vivo human studies validate the method. Characteristics of the color flow MR imaging technique are compared with those of duplex color ultrasound.
Given several images of the same slice, a linear filter can produce an image in which the contrast-to-noise ratio (CNR) between pathological and normal tissues is greater than in any of the initial images. To distinguish the pathology from more than one tissue, the filter should optimize the set of CNRs between the pathology and each of the interfering tissues. We define the optimal filter as the one which provides the largest value for the minimum CNR in the set and show how it is selected from a field of only four possibilities. The filter is demonstrated with both experimental phantom studies and clinical cases. Filter performance is compared with that of other techniques for distinguishing a desired feature from more than one interfering process.
Several aspects of blipped echo-planar imaging (EPI) are treated mathematically. An expression relating the necessary readout gradient strength and sampling time to the spatial resolution and readout duration is derived. It is shown how the net spatial resolution may be limited by the object's T2 characteristics and B0 field homogeneity, irrespective of the number of sampled points. Additionally, off-resonance effects result in a loss of spatial resolution and image distortion to a considerably greater degree than in conventional two-dimensional Fourier transform imaging. The extent of these effects is directly related to the time required to acquire the data matrix, and is therefore amplified when EPI is implemented on a standard commercial whole-body system which because of limited gradient performance uses necessarily longer sampling durations. Specific hardware modifications to a standard commercial imager are considered to allow successful EPI implementation. EPI image characteristics are compared quantitatively with those of conventional methods.
We describe a system for performing interactive MRI in real time. Using a TR/TE 7.1/3.5 ms sequence, the operator may alter a scan parameter and observe the effects of the alteration on the image within a few hundred milliseconds. With this system, we can interactively control the oblique scan slice orientation and, using inversion pulses, the image contrast.
We present a spoiler sequence which substantially eliminates residual transverse magnetization in rapid magnetic resonance imaging. Such magnetization is due to the repetition time TR being on the order of or shorter than the transverse relaxation time T2. This sequence was designed using computer simulation methods to model the slice profile and signal. The spoiler pulse is applied in the slice-selecting direction and the magnitude of the gradient spoiler is incremented with each repetition as an arithmetic series. A finite number of different spoiler amplitudes is repeated with a cycle time several times greater than the anticipated T2. Each spoiler is shown to nullify the residual transverse magnetization from the just-measured signal of the experiment as well as that from preceding repetitions. It is shown experimentally that the spoiler pulse can be as short as 2 ms. This is considerably shorter than existing methods using constant spoiling or phase-tagged spoiling. This sequence was used to acquire T1-weighted images free of the central line artifact of FLASH imaging. Furthermore, images formed with the new spoiled technique are shown to have contrast-to-noise comparable if not superior to standard T1-weighted spin-echo images while taking only a fifth of the scan time.
We describe a system which forms MR angiographic images at high speed. Multiple axial sections are imaged sequentially using a 2DFT GRASS sequence with TR/TE 50/15 ms, 64 phase encodings per image. Reconstruction and projection of each image are performed immediately (within 220 ms) after data for that image are acquired. The projection angiogram is constructed line by line as the scan progresses, thereby totally eliminating any additional time required for reconstruction and projection.
We show the importance of phase-encoding order in ultra-short TR (less than 10 ms) snapshot MR imaging. Contrast is more heavily dictated by the contrast preparation phase early in the measurement phase than later. When the low spatial frequencies are measured first, the resultant image contrast more closely resembles that established by the preparation phase. The transient nature of the data acquisition causes a kappa-space filter to be applied to the data. The effects of this filter can be minimized by appropriate choice of scan parameters.
To assess blood flow rapidly, a limited-flip-angle, gradient recalled pulse sequence was modified to acquire two views at the same phase-encoding step in successive repetitions. One view is obtained with first-moment flow compensation, while the second view is obtained with selectable flow encoding (non-zero first moment) along one direction. Blood flowing along the encoded direction acquires a phase difference between the two views, resulting in signal dependent on both direction and speed of flow. Stationary tissues undergo no phase change. Therefore, the phase shift between the two views produces an image that spatially renders flow direction and velocity. With a 24-msec repetition time, a 256 X 128 matrix, and two excitations, data acquisition is completed in 13 seconds per location (both a magnitude image and a flow image are produced at each location). Images generated with flow phantoms confirmed the accuracy of this method. Preliminary clinical evidence in 23 human subjects suggests that this method is useful in evaluating portal hypertension, distinguishing arterial from venous flow, distinguishing between slow flow and clot, and confirming the presence of clot. This method appears to be a fast, easy way to assess blood flow in large vessels.