(Radiographics. 1999;19:1057-1067.)
© RSNA, 1999
IMAGING & THERAPEUTIC TECHNOLOGY |
Effect of Field Strength on MR Images: Comparison of the Same Subject at 0.5, 1.0, and 1.5 T1
Antoine J. Maubon, MD, PhD,
Jean-Michel Ferru, MD,
Vincent Berger, MD,
Marie Colette Soulage, MD,
Marc DeGraef, MD,
Pierre Aubas, MD,
Patrice Coupeau,
Erik Dumont, PhD and
Jean-Pierre Rouanet, MD
1 From the Department of Radiology, Centre Medico-chirurgical Beausoleil, 119 Avenue de Lodève, 34000 Montpellier, France (A.J.M., J.M.F., V.B., M.C.S., M.D., J.P.R.); the Department of Medical Information, Centre Hospitalo Universitaire Montpellier, France (P.A.); and Philips Medical Systems, Paris, France (P.C., E.D.). Presented as a scientific exhibit at the 1997 RSNA scientific assembly. Received May 1, 1998; revision requested June 23; final revision received February 8, 1999; accepted February 12. Address reprint requests to A.J.M.
 |
Abstract
|
|---|
To assess the effect of field strength on magnetic resonance (MR) images, the same healthy subject was imaged at three field strengths: 0.5, 1.0, and 1.5 T. Imaging was performed with three similarly equipped MR imagers of the same generation and from the same manufacturer. The same imaging sequences were used with identical parameters and without repetition time correction for field strength. Imaging was performed in four anatomic locations: the brain, lumbar spine, knee, and abdomen. Quantitative image analysis involved calculation of signal-to-noise ratio, contrast-to-noise ratio, and relative contrast; qualitative image analysis was performed by four readers blinded to field strength. The results of all of the examinations were considered to be of diagnostic value. In general, signal-to-noise ratio and contrast-to-noise ratio were lowest at 0.5 T and highest at 1.5 T; relative contrast was not related to field strength. At qualitative analysis, images obtained at 1.0 and 1.5 T were superior to images obtained at 0.5 T; qualitative differences were less important in locations where there is motion or high magnetic susceptibility differences between tissues (eg, the spine and abdomen). However, excellent image quality was obtained with all three field strengths.
Index Terms: Magnetic resonance (MR), high-field-strength imaging, **.121412 • Magnetic resonance (MR), low-field-strength imaging, **.12141
|
INTRODUCTION
|
|---|
Is stronger really better? The ideal field strength for clinical magnetic resonance (MR) imaging has been an issue since the introduction of MR imaging (1,2). Nonetheless, in most previous studies, comparisons between field strengths were performed with MR imaging units from different companies or with different imaging sequences and different software (3–5).
We performed a study to assess the effect of field strength on image quality when all other parameters remain equal. The results of MR imaging studies of the same subject were compared quantitatively and qualitatively; these studies were performed with identical sequences and with similarly equipped MR imaging units of the same generation and from the same manufacturer at three field strengths: 0.5, 1.0, and 1.5 T.
|
MATERIALS AND METHODS
|
|---|
A healthy 36-year-old male volunteer (height, 173 cm; weight, 79 kg) underwent three MR imaging studies performed with different units at different sites within 1 week. The MR imaging units were from the Gyroscan series of Philips Medical Systems (Best, The Netherlands): the Gyroscan T5 NT (0.5 T), Gyroscan T10 NT (1.0 T), and Gyroscan ACS NT (1.5 T). These are superconducting, short-bore magnets with 15 mT/m gradients; they are implemented with Release 4.5 software and have echo-planar imaging capability and gradient-echo (GRE) and spin-echo (SE) imaging capability. The gradient coils and amplifiers used were identical for all units. Two intrinsic differences between the units were the radio frequency and the maximum radio-frequency power, which were 21.3 MHz and 5 kW at 0.5 T, 42.6 MHz and 15 kW at 1.0 T, and 64 MHz and 25 kW at 1.5 T.
The imaging sequences were strictly identical for all units without repetition time corrections for field strength. Receiver bandwidths were kept constant on all units due to the constraints of constant gradient performance and constant echo time and echo train length. The sequences were chosen according to the most frequent sites of MR imaging: the brain, lumbar spine, knee, and abdomen. They were performed with the same body coil or dedicated surface coils on all units. We chose the sequence types to represent everyday practice as closely as possible and to represent a wide range of imaging sequences (conventional and fast SE, GRE, echo-planar, inversion-recovery, GRE and SE) and imaging techniques (fat saturation, respiratory triggering, breath holding, three-dimensional imaging) (Table 1).
Quantitative Analysis
On each image, automated measurements were made with an electronic cursor on a remote console (Easyvision [Release 4.3]; Philips). We compared the signal intensity in regions of interest and calculated the following: signal-to-noise ratio (SNR) (signal intensity of tissue/standard deviation of noise measured outside the body), contrast-to-noise ratio (CNR) ([signal intensity of tissue A - signal intensity of tissue B]/standard deviation of noise measured outside the body), and relative contrast ([signal intensity of tissue A - signal intensity of tissue B]/signal intensity of tissue A, where tissue A has higher signal intensity than tissue B). The regions of interest were of the same size and in the same locations for each of the three sets of sequences.
For comparison purposes, for each sequence relative scores were given to the values of the three field strengths on a scale from 1 (lowest SNR, CNR, or relative contrast) to 3 (highest SNR, CNR, or relative contrast). The scores were pooled to obtain a global quantitative score for each field strength. Differences in the scores were compared by means of the Friedman test.
Qualitative Analysis
Four independent readers (A.J.M., V.B., M.C.S., M.D.) blinded to the field strength individually rated sets of three images obtained with the same sequence at different field strengths (0.5, 1.0, and 1.5 T); the image sets were read in random order. The optimal window parameters for each image were subjectively chosen by an experienced MR imager (J.M.F.) not involved in the readings and were not modified by the readers. Image quality and the demonstration and conspicuity of small anatomic structures selected according to location (cortical vessels, nerve roots, lumbar veins, menisci, cartilage, aortic branches, etc) were subjectively rated by the readers on a scale from 1 (worst sequence) to 3 (best sequence).
The scores were compared with the nonparametric Kruskal-Wallis test; differences were considered statistically significant when the P value was less than .05. When a difference was statistically significant, we tried to determine the origin of the difference by pairing the results for field strength by means of the Wilcoxon test with the Bonferroni correction for multiple comparisons.
|
RESULTS
|
|---|
The results of all of the examinations were considered to be of diagnostic value.
Quantitative Analysis
Figures 1–6 show MR images obtained with selected sequences (two sets of brain images, one set of spine images, one set of knee images, and two sets of abdominal images) at different field strengths (0.5, 1.0, and 1.5 T). These figures also show the corresponding graphs of SNR, CNR, and relative contrast in selected tissues according to field strength. In general, SNR and CNR increased steeply between 0.5 and 1.0 T but less steeply between 1.0 and 1.5 T.
View larger version (129K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 1a. (a-c) Axial T1-weighted SE MR images (562/14) obtained at 0.5 T (a), 1.0 T (b), and 1.5 T (c) show the basal nuclei. (d-f) Corresponding graphs show SNR (d), CNR (e), and relative contrast (RC) (f) according to field strength. CSF = cerebrospinal fluid, GM = gray matter, WM = white matter.
|
|
View larger version (130K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 1b. (a-c) Axial T1-weighted SE MR images (562/14) obtained at 0.5 T (a), 1.0 T (b), and 1.5 T (c) show the basal nuclei. (d-f) Corresponding graphs show SNR (d), CNR (e), and relative contrast (RC) (f) according to field strength. CSF = cerebrospinal fluid, GM = gray matter, WM = white matter.
|
|
View larger version (130K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 1c. (a-c) Axial T1-weighted SE MR images (562/14) obtained at 0.5 T (a), 1.0 T (b), and 1.5 T (c) show the basal nuclei. (d-f) Corresponding graphs show SNR (d), CNR (e), and relative contrast (RC) (f) according to field strength. CSF = cerebrospinal fluid, GM = gray matter, WM = white matter.
|
|
View larger version (11K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 1d. (a-c) Axial T1-weighted SE MR images (562/14) obtained at 0.5 T (a), 1.0 T (b), and 1.5 T (c) show the basal nuclei. (d-f) Corresponding graphs show SNR (d), CNR (e), and relative contrast (RC) (f) according to field strength. CSF = cerebrospinal fluid, GM = gray matter, WM = white matter.
|
|
View larger version (10K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 1e. (a-c) Axial T1-weighted SE MR images (562/14) obtained at 0.5 T (a), 1.0 T (b), and 1.5 T (c) show the basal nuclei. (d-f) Corresponding graphs show SNR (d), CNR (e), and relative contrast (RC) (f) according to field strength. CSF = cerebrospinal fluid, GM = gray matter, WM = white matter.
|
|
View larger version (10K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 1f. (a-c) Axial T1-weighted SE MR images (562/14) obtained at 0.5 T (a), 1.0 T (b), and 1.5 T (c) show the basal nuclei. (d-f) Corresponding graphs show SNR (d), CNR (e), and relative contrast (RC) (f) according to field strength. CSF = cerebrospinal fluid, GM = gray matter, WM = white matter.
|
|
View larger version (123K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 2a. (a-c) Axial T2-weighted GRE and SE MR images (5,374/110) obtained at 0.5 T (a), 1.0 T (b), and 1.5 T (c) show the cerebellar peduncles. (d-f) Corresponding graphs show SNR (d), CNR (e), and relative contrast (RC) (f) according to field strength. CSF = cerebrospinal fluid, GM = gray matter, WM = white matter.
|
|
View larger version (115K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 2b. (a-c) Axial T2-weighted GRE and SE MR images (5,374/110) obtained at 0.5 T (a), 1.0 T (b), and 1.5 T (c) show the cerebellar peduncles. (d-f) Corresponding graphs show SNR (d), CNR (e), and relative contrast (RC) (f) according to field strength. CSF = cerebrospinal fluid, GM = gray matter, WM = white matter.
|
|
View larger version (98K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 2c. (a-c) Axial T2-weighted GRE and SE MR images (5,374/110) obtained at 0.5 T (a), 1.0 T (b), and 1.5 T (c) show the cerebellar peduncles. (d-f) Corresponding graphs show SNR (d), CNR (e), and relative contrast (RC) (f) according to field strength. CSF = cerebrospinal fluid, GM = gray matter, WM = white matter.
|
|
View larger version (10K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 2d. (a-c) Axial T2-weighted GRE and SE MR images (5,374/110) obtained at 0.5 T (a), 1.0 T (b), and 1.5 T (c) show the cerebellar peduncles. (d-f) Corresponding graphs show SNR (d), CNR (e), and relative contrast (RC) (f) according to field strength. CSF = cerebrospinal fluid, GM = gray matter, WM = white matter.
|
|
View larger version (13K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 2e. (a-c) Axial T2-weighted GRE and SE MR images (5,374/110) obtained at 0.5 T (a), 1.0 T (b), and 1.5 T (c) show the cerebellar peduncles. (d-f) Corresponding graphs show SNR (d), CNR (e), and relative contrast (RC) (f) according to field strength. CSF = cerebrospinal fluid, GM = gray matter, WM = white matter.
|
|
View larger version (10K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 2f. (a-c) Axial T2-weighted GRE and SE MR images (5,374/110) obtained at 0.5 T (a), 1.0 T (b), and 1.5 T (c) show the cerebellar peduncles. (d-f) Corresponding graphs show SNR (d), CNR (e), and relative contrast (RC) (f) according to field strength. CSF = cerebrospinal fluid, GM = gray matter, WM = white matter.
|
|
View larger version (143K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 3a. (a-c) Sagittal T2-weighted fast SE MR images (2,924/130) obtained at 0.5 T (a), 1.0 T (b), and 1.5 T (c) show the midline of the lumbar spine. (d-f) Corresponding graphs show SNR (d), CNR (e), and relative contrast (RC) (f) according to field strength. BM = bone marrow, CSF = cerebrospinal fluid, Med = medulla.
|
|
View larger version (130K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 3b. (a-c) Sagittal T2-weighted fast SE MR images (2,924/130) obtained at 0.5 T (a), 1.0 T (b), and 1.5 T (c) show the midline of the lumbar spine. (d-f) Corresponding graphs show SNR (d), CNR (e), and relative contrast (RC) (f) according to field strength. BM = bone marrow, CSF = cerebrospinal fluid, Med = medulla.
|
|
View larger version (131K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 3c. (a-c) Sagittal T2-weighted fast SE MR images (2,924/130) obtained at 0.5 T (a), 1.0 T (b), and 1.5 T (c) show the midline of the lumbar spine. (d-f) Corresponding graphs show SNR (d), CNR (e), and relative contrast (RC) (f) according to field strength. BM = bone marrow, CSF = cerebrospinal fluid, Med = medulla.
|
|
View larger version (16K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 3d. (a-c) Sagittal T2-weighted fast SE MR images (2,924/130) obtained at 0.5 T (a), 1.0 T (b), and 1.5 T (c) show the midline of the lumbar spine. (d-f) Corresponding graphs show SNR (d), CNR (e), and relative contrast (RC) (f) according to field strength. BM = bone marrow, CSF = cerebrospinal fluid, Med = medulla.
|
|
View larger version (20K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 3e. (a-c) Sagittal T2-weighted fast SE MR images (2,924/130) obtained at 0.5 T (a), 1.0 T (b), and 1.5 T (c) show the midline of the lumbar spine. (d-f) Corresponding graphs show SNR (d), CNR (e), and relative contrast (RC) (f) according to field strength. BM = bone marrow, CSF = cerebrospinal fluid, Med = medulla.
|
|
View larger version (18K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 3f. (a-c) Sagittal T2-weighted fast SE MR images (2,924/130) obtained at 0.5 T (a), 1.0 T (b), and 1.5 T (c) show the midline of the lumbar spine. (d-f) Corresponding graphs show SNR (d), CNR (e), and relative contrast (RC) (f) according to field strength. BM = bone marrow, CSF = cerebrospinal fluid, Med = medulla.
|
|
View larger version (144K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 4a. (a-c) Frontal T1-weighted echo-planar SE MR images (630/20) obtained at 0.5 T (a), 1.0 T (b), and 1.5 T (c) show the intercondylar processes. The different appearance at 0.5 T (a) is due to a difference in patient positioning (30° flexion). (d-f) Corresponding graphs show SNR (d), CNR (e), and relative contrast (RC) (f) according to field strength. Cart = cartilage, cartil = cartilage, Menisc = meniscus.
|
|
View larger version (138K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 4b. (a-c) Frontal T1-weighted echo-planar SE MR images (630/20) obtained at 0.5 T (a), 1.0 T (b), and 1.5 T (c) show the intercondylar processes. The different appearance at 0.5 T (a) is due to a difference in patient positioning (30° flexion). (d-f) Corresponding graphs show SNR (d), CNR (e), and relative contrast (RC) (f) according to field strength. Cart = cartilage, cartil = cartilage, Menisc = meniscus.
|
|
View larger version (147K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 4c. (a-c) Frontal T1-weighted echo-planar SE MR images (630/20) obtained at 0.5 T (a), 1.0 T (b), and 1.5 T (c) show the intercondylar processes. The different appearance at 0.5 T (a) is due to a difference in patient positioning (30° flexion). (d-f) Corresponding graphs show SNR (d), CNR (e), and relative contrast (RC) (f) according to field strength. Cart = cartilage, cartil = cartilage, Menisc = meniscus.
|
|
View larger version (17K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 4d. (a-c) Frontal T1-weighted echo-planar SE MR images (630/20) obtained at 0.5 T (a), 1.0 T (b), and 1.5 T (c) show the intercondylar processes. The different appearance at 0.5 T (a) is due to a difference in patient positioning (30° flexion). (d-f) Corresponding graphs show SNR (d), CNR (e), and relative contrast (RC) (f) according to field strength. Cart = cartilage, cartil = cartilage, Menisc = meniscus.
|
|
View larger version (20K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 4e. (a-c) Frontal T1-weighted echo-planar SE MR images (630/20) obtained at 0.5 T (a), 1.0 T (b), and 1.5 T (c) show the intercondylar processes. The different appearance at 0.5 T (a) is due to a difference in patient positioning (30° flexion). (d-f) Corresponding graphs show SNR (d), CNR (e), and relative contrast (RC) (f) according to field strength. Cart = cartilage, cartil = cartilage, Menisc = meniscus.
|
|
View larger version (17K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 4f. (a-c) Frontal T1-weighted echo-planar SE MR images (630/20) obtained at 0.5 T (a), 1.0 T (b), and 1.5 T (c) show the intercondylar processes. The different appearance at 0.5 T (a) is due to a difference in patient positioning (30° flexion). (d-f) Corresponding graphs show SNR (d), CNR (e), and relative contrast (RC) (f) according to field strength. Cart = cartilage, cartil = cartilage, Menisc = meniscus.
|
|
View larger version (112K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 5a. (a-c) Axial T2-weighted ultrafast SE MR images (1,800/100) obtained with respiratory triggering at 0.5 T (a), 1.0 T (b), and 1.5 T (c) show the body of the pancreas. (d-f) Corresponding graphs show SNR (d), CNR (e), and relative contrast (RC) (f) according to field strength. Kid = kidney.
|
|
View larger version (114K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 5b. (a-c) Axial T2-weighted ultrafast SE MR images (1,800/100) obtained with respiratory triggering at 0.5 T (a), 1.0 T (b), and 1.5 T (c) show the body of the pancreas. (d-f) Corresponding graphs show SNR (d), CNR (e), and relative contrast (RC) (f) according to field strength. Kid = kidney.
|
|
View larger version (121K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 5c. (a-c) Axial T2-weighted ultrafast SE MR images (1,800/100) obtained with respiratory triggering at 0.5 T (a), 1.0 T (b), and 1.5 T (c) show the body of the pancreas. (d-f) Corresponding graphs show SNR (d), CNR (e), and relative contrast (RC) (f) according to field strength. Kid = kidney.
|
|
View larger version (14K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 5d. (a-c) Axial T2-weighted ultrafast SE MR images (1,800/100) obtained with respiratory triggering at 0.5 T (a), 1.0 T (b), and 1.5 T (c) show the body of the pancreas. (d-f) Corresponding graphs show SNR (d), CNR (e), and relative contrast (RC) (f) according to field strength. Kid = kidney.
|
|
View larger version (16K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 5e. (a-c) Axial T2-weighted ultrafast SE MR images (1,800/100) obtained with respiratory triggering at 0.5 T (a), 1.0 T (b), and 1.5 T (c) show the body of the pancreas. (d-f) Corresponding graphs show SNR (d), CNR (e), and relative contrast (RC) (f) according to field strength. Kid = kidney.
|
|
View larger version (14K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 5f. (a-c) Axial T2-weighted ultrafast SE MR images (1,800/100) obtained with respiratory triggering at 0.5 T (a), 1.0 T (b), and 1.5 T (c) show the body of the pancreas. (d-f) Corresponding graphs show SNR (d), CNR (e), and relative contrast (RC) (f) according to field strength. Kid = kidney.
|
|
View larger version (133K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 6a. (a-c) Axial T1-weighted GRE MR images (120/4.2) obtained with breath holding at 0.5 T (a), 1.0 T (b), and 1.5 T (c) show the superior hepatic veins. Note the out-of-phase appearance at 1.0 T (b) and the in-phase appearance at 1.5 T (c). Also note the aortic artifact in the left lobe of the liver on all three images (arrowhead). (d-f) Corresponding graphs show SNR (d), CNR (e), and relative contrast (RC) (f) according to field strength. Kid = kidney.
|
|
View larger version (108K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 6b. (a-c) Axial T1-weighted GRE MR images (120/4.2) obtained with breath holding at 0.5 T (a), 1.0 T (b), and 1.5 T (c) show the superior hepatic veins. Note the out-of-phase appearance at 1.0 T (b) and the in-phase appearance at 1.5 T (c). Also note the aortic artifact in the left lobe of the liver on all three images (arrowhead). (d-f) Corresponding graphs show SNR (d), CNR (e), and relative contrast (RC) (f) according to field strength. Kid = kidney.
|
|
View larger version (115K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 6c. (a-c) Axial T1-weighted GRE MR images (120/4.2) obtained with breath holding at 0.5 T (a), 1.0 T (b), and 1.5 T (c) show the superior hepatic veins. Note the out-of-phase appearance at 1.0 T (b) and the in-phase appearance at 1.5 T (c). Also note the aortic artifact in the left lobe of the liver on all three images (arrowhead). (d-f) Corresponding graphs show SNR (d), CNR (e), and relative contrast (RC) (f) according to field strength. Kid = kidney.
|
|
View larger version (13K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 6d. (a-c) Axial T1-weighted GRE MR images (120/4.2) obtained with breath holding at 0.5 T (a), 1.0 T (b), and 1.5 T (c) show the superior hepatic veins. Note the out-of-phase appearance at 1.0 T (b) and the in-phase appearance at 1.5 T (c). Also note the aortic artifact in the left lobe of the liver on all three images (arrowhead). (d-f) Corresponding graphs show SNR (d), CNR (e), and relative contrast (RC) (f) according to field strength. Kid = kidney.
|
|
View larger version (16K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 6e. (a-c) Axial T1-weighted GRE MR images (120/4.2) obtained with breath holding at 0.5 T (a), 1.0 T (b), and 1.5 T (c) show the superior hepatic veins. Note the out-of-phase appearance at 1.0 T (b) and the in-phase appearance at 1.5 T (c). Also note the aortic artifact in the left lobe of the liver on all three images (arrowhead). (d-f) Corresponding graphs show SNR (d), CNR (e), and relative contrast (RC) (f) according to field strength. Kid = kidney.
|
|
View larger version (14K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 6f. (a-c) Axial T1-weighted GRE MR images (120/4.2) obtained with breath holding at 0.5 T (a), 1.0 T (b), and 1.5 T (c) show the superior hepatic veins. Note the out-of-phase appearance at 1.0 T (b) and the in-phase appearance at 1.5 T (c). Also note the aortic artifact in the left lobe of the liver on all three images (arrowhead). (d-f) Corresponding graphs show SNR (d), CNR (e), and relative contrast (RC) (f) according to field strength. Kid = kidney.
|
|
Table 2 shows the mean and median quantitative scores for each field strength in each location. Differences in SNR between the three field strengths were statistically significant (P < .05) in all locations except between 1.0 and 1.5 T in the knee. Differences in CNR between the three field strengths were statistically significant (P < .05) in all locations except between 1.0 and 1.5 T in the abdomen. There were no significant differences in relative contrast between the field strengths in the brain, spine, or abdomen; differences in relative contrast between the three field strengths were statistically significant (P < .05) only in the knee.
Qualitative Analysis
Table 3 shows the qualitative scores by field strength for the various locations. The Kruskal-Wallis test showed a statistically significant difference between scores in the brain (P = .015) and in the knee (P = .008); however, although scores were higher at high field strengths, the Wilcoxon test showed no significant differences between the field strength scores in these locations. Differences between field strength scores were not significant in the spine (P = .077) or abdomen (P = .073).
|
DISCUSSION
|
|---|
As was expected from the physics of MR imaging (stronger magnetization and higher precession rates at high field strengths) and from the results of prior studies (6), quantitative analysis showed higher SNR and CNR at high field strengths with a steep increase between 0.5 and 1.0 T and a less steep slope between 1.0 and 1.5 T (Figs 1–6). In theory, SNR increases linearly with magnetic field strength as the square of the field strength. A higher field strength should align a larger number of the protons of the tissue along the main magnetic field and thus increase longitudinal magnetization and increase precession rates in the same proportions.
Nonetheless, additional factors, such as radio-frequency penetration, radiative loss, and quality aspects of the coils, combine to reduce this theoretic SNR benefit at high field strengths; this fact partly explains the nonlinear relationship we found with flattening of the SNR and CNR curves between 1.0 and 1.5 T. One of the major factors in this reduction appears to be lengthening of the T1 at high field strength: With identical repetition times, as in our study, faster magnetization recovery due to lower T1 values at lower field strengths proportionally increases signal on T1-weighted images. This phenomenon explains why the difference in SNR was more pronounced on T2-weighted images than on T1-weighted images.
We used a constant bandwidth by sequence on the three units. This constant bandwidth represents a potential difference from typical clinical conditions, under which the minimum bandwidth for each unit is generally used. However, under the conditions of our study, equivalent performances of the gradients on the three units and equivalent acquisition parameters necessitated an identical bandwidth on the three units for a given sequence.
Besides assessing the SNR and CNR, we also assessed the relative contrast between tissues. This parameter represents differences in contrast between two tissues (eg, white matter vs gray matter, liver vs spleen) but does not take noise into account. The relative contrast corresponds to what the eye effectively sees in an image; it can thus be assumed that this parameter reflects the ability of an imaging sequence to demonstrate a lesion within an organ. We found that relative contrast was less dependent on field strength than was CNR or SNR with no significant differences between the quantitative scores of the field strengths for the brain, spine, and abdomen. Only in the knee did relative contrast clearly increase with field strength. In the abdomen, scores at 0.5 T were equivalent to those at other field strengths and superior to those at 1.5 T on T2-weighted images. This result is probably due to a higher sensitivity to susceptibility artifacts at high field strengths, a tendency that becomes more apparent in tissues with strong susceptibility differences such as the abdomen.
Unlike in other studies, in which the 0.5-T unit was usually a downgraded version of a higher-field-strength unit, the three units in our series were as similar as possible, as were the sequences. The principle of using strictly the same parameters for each sequence on each unit produced no important differences in image appearance except for the axial T1-weighted GRE with breath-holding sequence used in the abdomen (Fig 6). The 3.4-µm difference in resonance frequency between water and fat (72 Hz at 0.5 T, 145 Hz at 1.0 T, 218 Hz at 1.5 T) meant that with the 4.2-msec echo time, the water and fat signals were in quadrature (90°) at 0.5 T, out of phase (180°) at 1.0 T, and in phase at 1.5 T, a result that explains the dissimilar appearances of the images.
As for qualitative analysis, the superiority of high-field-strength imaging (1.0 and 1.5 T) was confirmed with higher scores in the subjective evaluation of global image quality than at 0.5 T. Only with one T1-weighted sequence in the abdomen were 0.5-T images judged equivalent to 1.0-T images and superior to 1.5-T images. However, differences were statistically significant only in the brain and knee with no significant differences in the spine or abdomen. Furthermore, the paired comparisons did not show any significant differences between field strength scores. This result was probably partly due to the small number of readers (four) and the need for the Bonferroni correction for multiple comparisons, which lowers the level of significance to .025. Nonetheless, in the abdomen, 0.5-T imaging performed as well as imaging at the other field strengths and even better when T2-weighted sequences were used. A reason for this result may be the higher sensitivity at high field strengths to motion artifacts, which occur in the abdomen but not in the other locations.
The results of our study confirm those of previous studies, which showed that high-field-strength imaging is generally considered to produce high-quality images that are preferred by readers in image quality assessment studies (7). However, production of high-quality images does not generally translate into greater diagnostic accuracy in a variety of situations: studies of multiple sclerosis (8), the brain (9), internal derangement of the knee (10), tears of the anterior cruciate ligament (11), the spine (7), or the liver (12,13). Therefore, low-field-strength imaging has been proposed as a cost-effective alternative to the more expensive high-field-strength imaging (14). Another advantage of low-field-strength imaging is that for the same bandwidth, the shift between water and fat signals is three times lower at 0.5 T than at 1.5 T; this lower shift can improve image quality in regions where a low shift between water and fat signals is preferable (eg, the spine and knee).
Nonetheless, because we imaged a healthy subject, we cannot draw definite conclusions about imaging of patients. The next step in a more complete assessment of the effect of field strength on diagnostic accuracy will be to image a series of patients with the same protocol and assess the differences due to field strength in various pathologic conditions. Furthermore, our study included only commonly used diagnostic imaging sequences; many other sequences could have been used. Specifically, MR angiography, contrast material–enhanced MR angiography, and diffusion and perfusion MR imaging need to be compared at different field strengths and under clinical conditions.
In summary, excellent image quality was obtained with all three field strengths. Although SNR and CNR increased nonlinearly with field strength, relative contrast was not as dependent on field strength. Image quality was judged to be equivalent at 1.0 and 1.5 T and significantly higher than at 0.5 T with the notable exception of sequences performed in the abdomen.
|
Acknowledgments
|
|---|
Many thanks to the medical and technical teams at Fondation Rothschild, Paris, France, and Centre Hospitalo Universitaire Archet II, Nice, France.
|
Footnotes
|
|---|
**. Multiple body systems 
Abbreviations: CNR = contrast-to-noise ratio
GRE = gradient echo
SE = spin echo
SNR = signal-to-noise ratio
|
References
|
|---|
-
Hoult DI, Cher CN, Sank VJ. The field dependence of NMR imaging. II. Arguments concerning an optimal field strength. Magn Reson Med 1986; 3:730-746.
-
Bilaniuk LT, Zimmerman RA, Wehrli FW. Cerebral magnetic resonance: comparison of high and low field strength imaging. Radiology 1984; 153:409-414.[Abstract/Free Full Text]
-
Bradley WG. Future cost effective MRI will be at high field. JMRI 1996; 6:63-66.
-
Wiener JI. MR field strength: unanswered questions (letter). AJNR 1996; 17:189-190.[Medline]
-
Marshall D, Hailey D, Menon D. Magnetic field strength issues in magnetic resonance imaging. Can Assoc Radiol J 1994; 45:180-184.[Medline]
-
Hart HR, Jr, Bottomley PA, Edelstein WA, et al. Nuclear magnetic resonance imaging: contrast-to-noise ratio as a function of strength of magnetic field. AJR 1983; 141:1195-1201.[Abstract/Free Full Text]
-
Jack CR, Jr, Berquist TH, Miller GM, et al. Field strength in neuro-MR imaging: a comparison of 0.5 T and 1.5 T.. J Comput Assist Tomogr 1990; 14:505-513.[Medline]
-
Lee DH, Vellet AD, Eliasziw M, et al. MR imaging field strength: prospective evaluation of the diagnostic accuracy of MR for diagnosis of multiple sclerosis at 0.5 and 1.5 T.. Radiology 1995; 194:257-262.[Abstract/Free Full Text]
-
Orrison WW, Jr, Stimac GK, Stevens EA, et al. Comparison of CT, low-field-strength MR imaging, and high-field-strength MR imaging—work in progress. Radiology 1991; 181:121-127.[Abstract/Free Full Text]
-
Rutt BK, Lee DH. The impact of field strength on image quality in MRI. JMRI 1996; 6:57-62.
-
Vellet AD, Lee DH, Munk PL, et al. Anterior cruciate ligament tear: prospective evaluation of diagnostic accuracy of middle- and high-field-strength MR imaging at 1.5 and 0.5 T.. Radiology 1995; 197:826-830.[Abstract/Free Full Text]
-
Steinberg HV, Alarcon JJ, Bernardino ME. Focal hepatic lesions: comparative MR imaging at 0.5 and 1.5 T.. Radiology 1990; 174:153-156.[Abstract/Free Full Text]
-
Reinig JW, Dwyer AJ, Miller DL. Liver metastases: detection with MR imaging at 0.5 and 1.5 T.. Radiology 1989; 170:149-153.[Abstract/Free Full Text]
-
Parizel PM, Dijkstra HA, Geenen GP, et al. Low-field versus high-field MR imaging of the knee: a comparison of signal behaviour and diagnostic performance. Eur J Radiol 1995; 19:132-138.[Medline]
This article has been cited by other articles:
|
|
|
|
 
R. G. Steen, T. Emudianughe, G. M. Hankins, L. W. Wynn, W. C. Wang, X. Xiong, and K. J. Helton
Brain Imaging Findings in Pediatric Patients with Sickle Cell Disease
Radiology,
July 1, 2003;
228(1):
216 - 225.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
E. Salomonowitz
MR Imaging-Guided Biopsy and Therapeutic Intervention in a Closed-Configuration Magnet: Single-Center Series of 361 Punctures
Am. J. Roentgenol.,
July 1, 2001;
177(1):
159 - 163.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
C. A. Binkert, M. Zanetti, C. Gerber, and J. Hodler
MR Arthrography of the Glenohumeral Joint: Two Concentrations of Gadoteridol versus Ringer Solution as the Intraarticular Contrast Material
Radiology,
July 1, 2001;
220(1):
219 - 224.
[Abstract]
[Full Text]
[PDF]
|
|
|
Top
Abstract
INTRODUCTION
