HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Figures Only Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Rybicki, F. J. Articles by Ma, J. Search for Related Content PubMed PubMed Citation Articles by Rybicki, F. J. Articles by Ma, J. Social Bookmarking                      What's this?

Fast Three-Point Dixon MR Imaging Using Low-Resolution Images for Phase Correction

A Comparison with Chemical Shift Selective Fat Suppression for Pediatric Musculoskeletal Imaging

¹ Department of Radiology, Children's Hospital and Harvard Medical School, 330 Longwood Ave., Boston, MA 02115.
² Department of Radiology, Brigham and Women's Hospital and Harvard Medical School, 75 Francis St., Boston, MA 02115.
³ Edward B. Singleton Department of Diagnostic Imaging, Texas Children's Hospital and Baylor College of Medicine, 6621 Fannin St., Houston, TX 77030.
⁴ Department of Radiology, Children's Hospital, The Cleveland Clinic Foundation, 9500 Euclid Ave., Cleveland, OH 44195.
⁵ Department of Radiology, Massachusetts General Hospital and Harvard Medical School, 55 Fruit St., Boston, MA 02114.
⁶ Department of Radiology, The University of Texas M. D. Anderson Cancer Center, 1515 Holcombe Blvd., Houston TX 77030.

Received March 22, 2001; accepted after revision May 15, 2001.

 
Address correspondence to F. J. Rybicki.

Abstract

Top Abstract Introduction Subjects and Methods Results Discussion References

 
OBJECTIVE. The purpose of this study is to describe and to implement a new fast three-point Dixon MR imaging sequence with online image reconstruction, and to compare this sequence with conventional chemical shift selective (CHESS) suppression of fat in pediatric musculoskeletal imaging.

SUBJECTS AND METHODS. A three-point Dixon technique using a fast spin-echo sequence with a new phase-correction algorithm providing online image reconstruction was implemented on a 1.5-T scanner. Twelve pediatric patients and young adults were imaged with both the new three-point Dixon and conventional CHESS sequences. Three radiologists un-aware of imaging parameters and clinical information independently scored the homogeneity of fat suppression and conspicuity of abnormality using a four-point system. An additional comparison between the two techniques was made using a phantom.

RESULTS. The three-point Dixon method showed superior fat suppression and lesion conspicuity (p < 0.001), particularly in the hands and feet, where CHESS is prone to inconsistent fat suppression. The phantom study showed no significant difference in the ratio of suppressed fat signal to background noise and more homogeneous fat suppression using the three-point Dixon method.

CONCLUSION. Compared with CHESS, the new fast three-point Dixon sequence with online image reconstruction provides superior fat suppression and lesion conspicuity and can be routinely used in pediatric musculoskeletal imaging.

Introduction

Top Abstract Introduction Subjects and Methods Results Discussion References

 
Fast spin-echo T2-weighted MR imaging sequences with chemical shift selective (CHESS) fat suppression are routinely used in pediatric musculoskeletal MR imaging. Because both fat and water have high signal intensity in fast spin-echo T2-weighted imaging, robust fat suppression is particularly crucial for the detection of disease. Using CHESS, the homogeneity of the fat suppression over the entire field of view is highly dependent on the magnetic field homogeneity, which is invariably perturbed by the presence of the patient [1]. Multipoint Dixon techniques [1,2,3,4] can achieve high-quality fat suppression [5, 6] and produce three images per slice: pure water, pure fat, and water plus fat. However, multipoint Dixon techniques are less commonly used, because they require multiple data acquisitions and lengthy image reconstruction algorithms. The multiple acquisitions lengthen the imaging time; the reconstruction algorithms typically require off-line image reconstruction. The three-point Dixon sequence described in this study allows water and fat separation with a reduced imaging time and a reduced reconstruction time. The reduced imaging time is achieved by incorporating a fast spin-echo sequence [7] with an echo-train length of 8. The reduced reconstruction time is based on the fact that the phase errors are slowly varying on the scale of image pixels. Therefore, the phase error can be corrected with low-resolution images that have the advantage of decreased acquisition time and increased signal-to-noise ratio. This principle is combined with an asymmetric data acquisition [8,9,10]. The result is scan times comparable to those of CHESS and online reconstruction. The objectives of this study are to implement this sequence in routine pediatric musculoskeletal imaging and to compare the new three-point Dixon and conventional CHESS sequences in two tasks: homogeneity of fat suppression and lesion conspicuity.

Subjects and Methods

Top Abstract Introduction Subjects and Methods Results Discussion References

 
Three-Point Dixon Technique
In the original Dixon technique [2], water and fat are separated using two image acquisitions (a two-point technique) for which the magnetization vectors for water and fat are parallel and anti-parallel. The summation of the images in the complex form yields a pure water image, and the subtraction of the images yields a pure fat image. The main limitation of this method is that field inhomogeneity—for example, inhomogeneity introduced from the presence of the patient—yields a phase error that leads to an incorrect solution for the water and fat images. The phase error is commonly corrected using the information in a third data acquisition; the methods are therefore termed "three-point Dixon." In a commonly used approach [3], the magnetization vectors are offset by 0, , and 2 in the three acquisitions; this is referred to as a symmetric technique. One limitation of symmetric techniques is that for each pixel the phase error is absolutely determined over the range only [-, ]. Thus there is potential for an incorrect assignment of the phase, so-called "phase aliasing." The correction of phase aliasing, called phase unwrapping, is challenging. Failure in phase unwrapping results in reversal between the water and fat signal.

An alternative three-point strategy uses asymmetric data sampling [8,9,10]. That is, for at least one of the three data acquisitions, the fat and water magnetization vectors are neither parallel nor anti-parallel. The main advantage of asymmetric data sampling is the ability to separate water and fat without direct phase unwrapping. The three-point Dixon technique in this study uses a modified form of this asymmetric sampling with three acquisitions separated by /2 [11, 12].

The data-processing portion of our modification is based on the realization that the elusive phase error is, in general, spatially slow varying. Therefore the phase error can be adequately and more readily determined from a set of low-resolution images for which the signal-to-noise ratio is substantially elevated. In postprocessing, three low-resolution images are reconstructed and the phase factors for all pixels with adquate water and fat signal are determined directly on a pixel-by-pixel basis. For pixels that either have low signal-to-noise ratio or contain only water or only fat, the phase factor is obtained by a region-growing process designed to ensure spatial phase continuity. The low-resolution phase factors are then used directly for correcting the phase errors. The final output of the pure water, pure fat, and water-plus-fat images do not have resolution loss because only the phase errors are removed from the images with high resolution.

The entire reconstruction is transparent to the operators and is automatically initiated after each data collection. Output images are installed directly into the image database for viewing and archiving. The software package has been implemented on the LX platform operating at the 8.2 hardware/software level (General Electric Medical Systems, Milwaukee, WI).

Patients
Thirteen musculoskeletal MR studies were performed in 12 patients (median age, 4.8 years; range, 7 months-20 years). Both fast spin-echo T2-weighted images with CHESS pulse fat suppression (TR range/effective TE, 3000-5000/96-105; echo-train length, 8; bandwidth, 16 kHz; phase, 160-256; number of signal averages, 2; scanning time, 2:47-5:06 min) and three-point Dixon fast spin-echo T2-weighted images (2500-3400/78-150; echo-train length, 8; bandwidth, 16-31.2 kHz; phase, 192; scanning time, 3:48-6:10 min) were acquired during the same imaging session at Children's Hospital, Boston, MA. Imaging was performed with the following coils: quadrature receive-transmit head, receive-only shoulder, and receive-only extremity.

Phantom
A water-fat-air phantom [13] containing a com oil sample was scanned (Fig. 1A,1B) with both the CHESS (TR/effective TE, 3000/80; echo-train length, 8) and three-point Dixon sequences (3000/87; echo-train length, 8). For each sequence, the slice thickness was 5 mm, the field of view was 20 cm, and the matrix size was 256 x 192. The CHESS sequence was performed with two acquisitions (compared with the single acquisition for the three-point Dixon sequence) to obtain comparable scanning times and to mimic a clinical scenario.

View larger version (76K): [in this window] [in a new window] [as a PowerPoint slide]   Fig. 1A. —Images of corn-oil-and-water phantom acquired with three-point Dixon and chemical shift selective (CHESS) suppression of fat. CHESS fat suppression of corn oil in central vial shows inhomogeneity at top of vial at its interface with air (arrow).

View larger version (65K): [in this window] [in a new window] [as a PowerPoint slide]   Fig. 1B. —Images of corn-oil-and-water phantom acquired with three-point Dixon and chemical shift selective (CHESS) suppression of fat. Three-point Dixon pure-water image illustrates more homogeneous suppression of corn oil in central vial.

Data Collection and Analysis
Three pediatric radiologists who were unaware of the imaging parameters and clinical information independently and retrospectively scored the homogeneity of fat suppression and conspicuity of abnormalities in the clinical images using a four-point system (1 = poor, 4 = excellent). The scores of each observer were averaged, and differences between scores were evaluated using Wilcoxon's signed rank test.

The same three pediatric radiologists then chose among the following for each clinical case: CHESS was superior to three-point Dixon (score = 1), three-point Dixon was superior to CHESS (score = 2), or both sequences were equal (score = 3). The reviewers were paired (A vs B, A vs C, and B vs C), and computed unweighted kappa statistics on the scores were obtained. (A kappa value of 1.0 reveals perfect agreement, zero is no agreement beyond chance, and -1.0 shows perfect disagreement.)

The phantom images were analyzed using regions of interest to measure the signal intensity of both fat and background noise. A quantitative comparison between the three-point Dixon pure water images and the CHESS images was performed by applying the Student's t test to the ratio of suppressed fat to background noise.

Results

Top Abstract Introduction Subjects and Methods Results Discussion References

 
Patient Images
The qualitative improvement in fat suppression and lesion conspicuity offered by the three-point Dixon pure-water images was statistically significant at the 0.001 level (Table 1). Regarding the fat suppression, all reviewers were in perfect agreement ( = 1) that the three-point Dixon pure-water images were superior to the CHESS images; the agreement that the three-point Dixon pure-water images provided better lesion conspicuity was fair ( = 0.38, 0.46, and 0.20). The advantage of the three-point Dixon sequence proved most dramatic in imaging body parts with irregular contours, such as the feet (Fig. 2A,2B), where CHESS typically provides limited fat suppression.

View larger version (102K): [in this window] [in a new window] [as a PowerPoint slide]   Fig. 2A. —MR examination of both feet in 2.5-year-old girl with 6-month history of limp favoring left foot. Feet are positioned together in quadrature head coil. Three-point Dixon pure-water images shows excellent conspicuity of marrow edema of left first metatarsal bone (arrow).

View larger version (112K): [in this window] [in a new window] [as a PowerPoint slide]   Fig. 2B. —MR examination of both feet in 2.5-year-old girl with 6-month history of limp favoring left foot. Feet are positioned together in quadrature head coil. Chemical shift selective image corresponding to A does not reveal finding caused by inhomogeneous fat suppression.

The three-point Dixon pure-water images proved useful in answering questions raised by the CHESS images (Fig. 3A,3B), and the additional information from the three-point Dixon pure-fat images clarified the chemical composition of tissues in regions where the CHESS fat suppression was inhomogeneous (Fig. 4A,4B,4C).

View larger version (108K): [in this window] [in a new window] [as a PowerPoint slide]   Fig. 3A. —4-year-old girl with multiple venous malformations. Examination of knee is performed using extremity coil. Chemical shift selective image raises possibility of venous malformation in subcutaneous tissues (arrow) over patella.

View larger version (100K): [in this window] [in a new window] [as a PowerPoint slide]   Fig. 3B. —4-year-old girl with multiple venous malformations. Examination of knee is performed using extremity coil. Three-point Dixon pure-water image unequivocally excludes abnormality in this region.

View larger version (80K): [in this window] [in a new window] [as a PowerPoint slide]   Fig. 4A. —Follow-up MR imaging in 8-month-old girl after chemotherapy for rhabdomyosarcoma of right leg. Quadrature head coil is used. Three-point Dixon pure-water image shows nearly complete replacement of fat in marrow space with longer T1 substance, likely edema, red-marrow, or both.

View larger version (89K): [in this window] [in a new window] [as a PowerPoint slide]   Fig. 4B. —Follow-up MR imaging in 8-month-old girl after chemotherapy for rhabdomyosarcoma of right leg. Quadrature head coil is used. Three-point Dixon pure-fat image shows minimal, if any, fat in marrow space.

View larger version (83K): [in this window] [in a new window] [as a PowerPoint slide]   Fig. 4C. —Follow-up MR imaging in 8-month-old girl after chemotherapy for rhabdomyosarcoma of right leg. Quadrature head coil is used. Chemical shift selective image offers less robust characterization of tissue chemical composition.

Phantom Images
The water-fat phantom experiment yielded no quantitative difference (p > 0.7) between the ratios of suppressed fat signal to background noise between the three-point Dixon pure-water images and the CHESS images. In addition, the three-point Dixon pure-water image showed more homogeneous signal, particularly near the interface between air and fat (Fig. 1A,1B).

Discussion

Top Abstract Introduction Subjects and Methods Results Discussion References

 
The most common methods to suppress the signal from fat in clinical imaging are CHESS and the inversion-recovery preparatory pulse. CHESS relies on a homogeneous magnetic field over the entire field of view to achieve homogeneous fat suppression. However, when imaging the extremities, the field is often inhomogeneous because the anatomy of interest can be off-center in the magnet. This is particularly true in the pediatric population; for example, a child may not tolerate being examined in the prone position with the upper extremity positioned above the head. In this scenario, fat saturation in the elbow, forearm, and wrist is difficult to achieve. Furthermore, the anatomy of interest may be small or irregularly shaped. In these instances, an unconventional use of a coil becomes necessary; for example, an adult head coil imaging both ankles or both knees of a young child, or a small surface coil imaging an infant's foot [14]. In these situations, inhomogeneous CHESS fat suppression may obscure abnormalities on T2-weighted images. One alternative to CHESS is an inversion-recovery-prepared fast spin-echo sequence. However, these images have inherently less signal [11], and the homogeneity of fat suppression relies on the homogeneity of the radiofrequency field associated with the inversion pulse.

The primary purpose of this study was to show that the new fast three-point Dixon sequence with a fully online image reconstruction could be implemented for routine pediatric musculoskeletal imaging. The fundamental challenge for Dixon techniques is to robustly correct the phase errors of the complex images using a short acquisition time and a short reconstruction time. To minimize acquisition time, the three-point Dixon sequence uses fast spin echo with an echo-train length of 8 [7]. The novel component of the current sequence is that the image processing time is reduced because, in image processing, the images used have reduced matrix size, increased signal-to-noise ratio, and fewer pixels that contain only water or fat that would require additional analysis. The reduction in image processing time enables online reconstruction, which makes this sequence practical for routine clinical use. Although the usefulness of Dixon techniques and variations of Dixon techniques has been shown in imaging the musculoskeletal system [15] and in pediatric imaging [16], these reconstruction algorithms are generally off-line. Moreover, imaging time for T1 weighting is typically greater than 7 min, and the longer imaging times for T2 weighting are usually prohibitive [17].

Three-point Dixon MR imaging using low-resolution images for phase correction showed excellent reliability of water and fat separation in initial experiments [11, 12], and the phantom study showed that fat suppression was quantitatively equal to CHESS and appeared more homogeneous. Thus, the second purpose of this study was to confirm these observations in the clinical setting. In all cases, the water- and fat-containing structures were well separated. Qualitative analysis revealed more homogeneous fat suppression and increased lesion conspicuity from the three-point Dixon pure-water images. Moreover, in some cases, the robust separation of water and fat using the three-point Dixon sequence made possible the exclusion of disease in areas where the corresponding CHESS images yielded suboptimal fat suppression. Although the significance of the pure-fat images was not rigorously examined, these images were clinically helpful for tissue characterization.

One limitation of this study is the small variability of the parameters in the comparison between CHESS and three-point Dixon sequences. Although the uniformity of fat saturation should not be affected, variability in parameters could lead to differences in image contrast and image signal-to-noise ratio, which in turn could affect the detection of disease. With respect to image contrast, both the CHESS and three-point Dixon sequences were performed with a long TR. Although the TR was on average slightly decreased for the three-point Dixon sequence to maintain short scanning times, in 11 of 13 cases the TR for the three-point Dixon sequence was greater than or equal to 3000 msec, thus minimizing T1 weighting. All scans were also performed with an effective TE to achieve reasonably heavy T2 weighting. Consequently, the differences in image contrast are not expected to dramatically affect the detection of lesion abnormality. The signal-to-noise ratio for each sequence is proportional to the square root of the ratio of the number of signal averages to the bandwidth [18]. Using an estimate for the effective number of signal averages in a three-acquisition Dixon sequence [3], the signal-to-noise ratio among the CHESS and three-point Dixon sequence varied by less than 25% for each of the 13 comparisons. Thus, despite the small variability between parameters, the contrast and signal-to-noise ratio among patients is expected to have a small impact, if any, on the improved conspicuity of disease obtained by using the three-point Dixon pure-water images compared with CHESS.

In summary, this work describes a three-point Dixon sequence suitable for routine pediatric musculoskeletal MR imaging. The sequence uses fast spin echo to shorten the scanning time and a method to reduce the image processing time that incorporates low-resolution images for phase correction. In a clinical comparison with CHESS, the new three-point Dixon sequence provides superior fat suppression and lesion conspicuity.

References

Top Abstract Introduction Subjects and Methods Results Discussion References

 

1. Glover GH. Multipoint Dixon technique for water and fat proton and susceptibility imaging. J Magn Reson Imaging 1991;1:521 -530[Medline]
2. Dixon WT. Simple proton spectroscopic imaging. Radiology 1984;153:189 -194[Abstract/Free Full Text]
3. Glover GH, Schneider E. Three-point Dixon technique for true water/fat decomposition with B0 field inhomogeneity correction. Magn Reson Med 1991;18:371 -383[Medline]
4. Wang Y, Li D, Haacke EM, Brown JJ. A three-point Dixon method for water and fat separation using 2D and 3D gradient-echo techniques. J Magn Reson Imaging 1998;8:703 -710[Medline]
5. Tamler B, Sommer FG, Glover GH, Schneider E. Prostatic MR imaging performed with the three-point Dixon technique. Radiology 1991;179:43 -47[Abstract/Free Full Text]
6. Daniel BL, Butts K, Glover GH, Cooper C, Herfkens RJ. Breast cancer: gadolinium-enhanced MR imaging with a 0.5-T open imager and three-point Dixon technique. Radiology 1998;207:183 -190[Abstract/Free Full Text]
7. Hardy PA, Hinks RS, Tkach JA. Separation of fat and water in fast spin-echo MR imaging with the three-point Dixon technique. J Magn Reson Imaging 1995;5:181 -185[Medline]
8. Xiang QS, An L. Water-fat imaging with direct phase encoding. (abstr) In: Proceedings of the 3rd annual scientific meeting of the Society of Magnetic Resonance. Berkeley, CA: International Society for Magnetic Resonance in Medicine (ISMRM), 1995:650
9. Xiang QS, An L. General 3-point water-fat imaging with optimized SNR. (abstr) In: Proceedings of the 4th annual scientific meeting of the Society of Magnetic Resonance. Berkeley, CA: ISMRM, 1996: 1544
10. Xiang QS, An L. Water-fat imaging with direct phase encoding. J Magn Reson Imaging 1997;7:1002 -1015[Medline]
11. Rybicki FJ, Mulkern RV, Robertson RL, et al. Fast three-point Dixon MR imaging of the retrobulbar space using low-resolution images for phase correction: a comparison with fast spin echo inversion recovery. Am J Neuroradiol 2001;22:1798 -1802[Abstract/Free Full Text]
12. Ma J. Multipoint Dixon imaging with reduced time and increased reliability. (abstr) In: Proceedings of 6th annual scientific meeting of the International Society for Magnetic Resonance in Medicine. Berkeley, CA: ISMRM, 1998:622
13. Kurodak K, Oshio K, Mulkern RV, Jolesz FA. Optimization of chemical shift selective suppression of fat. Magn Reson Med 1998;40:505 -510[Medline]
14. Barnewolt CE, Chung T. Techniques, coils, pulse sequences, and contrast enhancement in pediatric musculoskeletal MR imaging. Magn Reson Imaging Clin North Am 1998;6:441 -453
15. Maas M, Dijkstra PF, Akkerman EM. Uniform fat suppression in hands and feet through the use of two-point Dixon chemical shift MR imaging. Radiology 1999;210:189 -193[Abstract/Free Full Text]
16. Sargent MA, Poskitt KJ, Xiang QS, et al. Application of a three-point method for water-fat MR imaging in children. Pediatr Radiol 1999;29:444 -448[Medline]
17. Carson BW, Xiang QS. Fat suppression using direct phase encoding: musculoskeletal applications using MR imaging. AJR 1999;173:230 -233[Free Full Text]
18. Edelstein WA, Glover GH, Hardy CJ, et al. The intrinsic signal-to-noise ratio in NMR imaging. Magn Reson Med 1986;3:604 -618[Medline]

CiteULike

Complore

Connotea

Del.icio.us

Digg

Reddit

Technorati

This article has been cited by other articles:

				L. F. Donnelly, K. J. O'Brien, B. J. Dardzinski, S. A. Poe, J. A. Bean, S. K. Holland, and S. R. Daniels Using a Phantom to Compare MR Techniques for Determining the Ratio of Intraabdominal to Subcutaneous Adipose Tissue Am. J. Roentgenol., April 1, 2003; 180(4): 993 - 998. [Abstract] [Full Text] [PDF]

				S. B. Reeder, N. J. Pelc, M. T. Alley, and G. E. Gold Rapid MR Imaging of Articular Cartilage with Steady-State Free Precession and Multipoint Fat-Water Separation Am. J. Roentgenol., February 1, 2003; 180(2): 357 - 362. [Abstract] [Full Text] [PDF]

This Article Abstract Figures Only Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Rybicki, F. J. Articles by Ma, J. Search for Related Content PubMed PubMed Citation Articles by Rybicki, F. J. Articles by Ma, J. Social Bookmarking                      What's this?