**4.4 Comparison between RARE and TrueFISP imaging**

136 Advances in the Biology, Imaging and Therapies for Glioblastoma

Table 1. S/N and contrast for brain and tumor, as a function of the flip angle, α. TE was set

Using a low bandwidth generates images with a high S/N. However, a decrease in bandwidth causes an increase in the sequence's TE. The influence of this was measured over the range between 1.5 and 10 ms (with the flip angle set at 35°). With a low TE (1.5 ms), the contrast appeared better than in the reference image acquired at 2.5 ms, but the S/N was significantly reduced, and the signal from fatty tissue was increased due to its relatively low T1. This resulted in a lower-quality image. Interestingly, at higher TEs (5 and 10 ms), the TrueFISP sequence generated few susceptibility artifacts compared with a FLASH sequence with the same TE (data not shown). However, at TE = 10 ms, the tumor was no longer

As described in the table below, only a TE value of 2.5 ms provides a contrast greater than

Fig. 9. Images acquired using the 3D TrueFISP sequence after correction of banding artifacts

Table 2. S/N and contrast for brain and tumor, as a function of the echo time. The flip

**4.3 Influence of TE on S/N and contrast in TrueFISP imaging** 

visible due to a strongly reduced contrast between brain and tumor.

10, allowing the tumor to be segmented semi-automatically.

based on TE. The flip angle was set to 35°.

angle α was set at 35°.

to 2.5 ms.

The reference method for tumor detection by anatomic imaging is T2-weighted spin-echo imaging. Images were therefore acquired with a T2-weighted 2D RARE sequence. This sequence detects the tumor despite its small size (see Fig. 10d). In addition, as expected from simulations, contrast between the tumor and healthy brain is higher than with the TrueFISP sequence (Fig. 10a). The total acquisition time for the 2D RARE sequence was comparable with that for the 3D TrueFISP sequence. However, as the images were acquired in 2D by multi-slice imaging, reconstruction in other dimensions generates images with poor resolution and low quality (Figs. 10ef) compared with the 3D TrueFISP sequence (Figs. 10bc).

The two imaging methods were next used to measure tumor volume. With the TrueFISP sequence, the volume was estimated at 2.4 µL, while with the RARE sequence it was estimated at 4.3 µL. This overestimation of the volume with the RARE sequence is due to the low spatial resolution of this sequence in the third dimension. Actually, significant partialvolume effects are generated by this sequence and lead to volume-measurement errors.

A 3D RARE sequence validated the 2.4 µL volume determined using the TrueFISP sequence. The 3D RARE image was acquired over 45 minutes.

Fig. 10. 3D TrueFISP (a, b, and c) and multi-slice 2D RARE imaging (d, e, and f) of a mouse with implanted glioma: a) & d) sagittal sections; b) & e) reconstruction of transverse sections; c) & f) reconstruction of coronal sections. The white arrows indicate the tumor.

Volume overestimation based on 2D imaging was confirmed by using the two sequences to image a phantom consisting of a known volume of water. With the 2D RARE sequence, the error in volume measurement was around 20%, while with the TrueFISP sequence the error was only 4%.

3D TrueFISP MRI Provides Accurate Longitudinal Measurements of Glioma Volumes in Mice 139

The tumor was easily detectable 7 days after implantation despite its small volume (2 µL). The images obtained with the TrueFISP sequence are shown from day 9 after implantation. The animal was then followed up for approximately 30 days, with a 2 or 3-day interval between MRI scans. The increase in tumor diameter is visible in the images (Fig. 12). From d25, zones of black signal appear in the tumor's center, these are probably due to necrotic

Fig. 12. Longitudinal follow-up of tumor progression by 3D TrueFISP imaging at 4.7 T. Images ab show the sagittal and transverse sections of the mouse brain at d9 after

is lower, but the tumor can still be manually delimited and its volume measured.

Based on these 3D images (a-h), the tumor can readily be semi-automatically segmented, and the 3D volume could be reconstructed (Fig. 13). In Images i-j, the tumor/brain contrast

Based on these 3D reconstructions, tumor volume progression was reported as a function of time (Fig. 14). Tumor growth could thus be assessed, from its initial approximately 4µL volume up to its final 100µL volume, i.e. approximately one quarter of the animal's total

implantation. Images cj show tumor progression between d11 and d28.

areas.

brain volume.

#### **4.5 Imaging at 9.4T**

Images were acquired with the same 2 sequences at 9.4 T on a mouse model of glioma. To suppress banding artifacts with the TrueFISP sequence, 8 images had to be acquired with 8 different Δφ values (180°, 0°, 90°, 270°, 45°, 135°, 225°, 315°).

Fig. 11. 2D RARE (a) and 3D TrueFISP (b) imaging at 9.4 T. With RARE imaging, the tumor is clearly visible, while TrueFISP provides a lower tumor/brain contrast.

With the RARE sequence, the tumor is clearly detectable and contrast is comparable with that of an image acquired at 4.7 T (Fig. 10a). For the TrueFISP sequence, the contrast is reduced, making it very difficult to distinguish the tumor from the remainder of the brain, except from the presence of structural heterogeneity. In this case, it is not possible to semiautomatically segment the tumor, and it is therefore impossible to measure its volume with good reproducibility. In addition, as demonstrated previously, small tumors may turn out to be undetectable.


Table 3. S/N and contrast for brain and tumor, at 4.7 T and 9.4 T, with RARE and TrueFISP sequences.

#### **4.6 Longitudinal follow up of glioma volume at 4.7T**

The simulations and experiments performed above show that an implanted tumor can be unambiguously distinguished by 3D TrueFISP imaging when a TE of 2.5 ms is combined with a flip angle of 35° for imaging at 4.7 T. These parameters were therefore used for the longitudinal follow-up of tumor volume in an animal implanted with a U87-type glioma. Images were also acquired with the 2D RARE sequence to compare volume measurements.

Images were acquired with the same 2 sequences at 9.4 T on a mouse model of glioma. To suppress banding artifacts with the TrueFISP sequence, 8 images had to be acquired with

Fig. 11. 2D RARE (a) and 3D TrueFISP (b) imaging at 9.4 T. With RARE imaging, the tumor

With the RARE sequence, the tumor is clearly detectable and contrast is comparable with that of an image acquired at 4.7 T (Fig. 10a). For the TrueFISP sequence, the contrast is reduced, making it very difficult to distinguish the tumor from the remainder of the brain, except from the presence of structural heterogeneity. In this case, it is not possible to semiautomatically segment the tumor, and it is therefore impossible to measure its volume with good reproducibility. In addition, as demonstrated previously, small tumors may turn out

Table 3. S/N and contrast for brain and tumor, at 4.7 T and 9.4 T, with RARE and TrueFISP

The simulations and experiments performed above show that an implanted tumor can be unambiguously distinguished by 3D TrueFISP imaging when a TE of 2.5 ms is combined with a flip angle of 35° for imaging at 4.7 T. These parameters were therefore used for the longitudinal follow-up of tumor volume in an animal implanted with a U87-type glioma. Images were also acquired with the 2D RARE sequence to compare volume

**4.6 Longitudinal follow up of glioma volume at 4.7T** 

is clearly visible, while TrueFISP provides a lower tumor/brain contrast.

8 different Δφ values (180°, 0°, 90°, 270°, 45°, 135°, 225°, 315°).

**4.5 Imaging at 9.4T** 

to be undetectable.

sequences.

measurements.

The tumor was easily detectable 7 days after implantation despite its small volume (2 µL). The images obtained with the TrueFISP sequence are shown from day 9 after implantation. The animal was then followed up for approximately 30 days, with a 2 or 3-day interval between MRI scans. The increase in tumor diameter is visible in the images (Fig. 12). From d25, zones of black signal appear in the tumor's center, these are probably due to necrotic areas.

Fig. 12. Longitudinal follow-up of tumor progression by 3D TrueFISP imaging at 4.7 T. Images ab show the sagittal and transverse sections of the mouse brain at d9 after implantation. Images cj show tumor progression between d11 and d28.

Based on these 3D images (a-h), the tumor can readily be semi-automatically segmented, and the 3D volume could be reconstructed (Fig. 13). In Images i-j, the tumor/brain contrast is lower, but the tumor can still be manually delimited and its volume measured.

Based on these 3D reconstructions, tumor volume progression was reported as a function of time (Fig. 14). Tumor growth could thus be assessed, from its initial approximately 4µL volume up to its final 100µL volume, i.e. approximately one quarter of the animal's total brain volume.

3D TrueFISP MRI Provides Accurate Longitudinal Measurements of Glioma Volumes in Mice 141

Fig. 14. Progression of tumor volume over time based on the volumes shown in Figure 13.

The aim of this chapter was to show how the 3D TrueFISP sequence can improve

This sequence is rarely used at high magnetic fields as it generates numerous artifacts known as banding artifacts [10-12]. However, the S/N and contrast provided by this

It has been shown that these artifacts can be easily corrected at 4.7 T and 9.4 T when studying mouse brains, and also for areas which are much more sensitive to movements and susceptibility artifacts, such as the heart [8,13]. To do so, the so-called 'alternating-phase RF pulse' method is used, in combination with reconstruction through calculation of the square root of the sum of squares of the magnitude signal [14]. As shown here, 4 phase steps are necessary to suppress artifacts at 4.7 T, while at least 8 steps are required at 9.4 T [8]. The resulting image appears perfectly homogeneous in terms of signal, and no longer displays artifacts. The requirement for 4 images increases the total acquisition time, but this nevertheless remains much shorter than with a T2-weighted 3D RARE sequence (12 min vs.

The 3D TrueFISP imaging sequence was compared, in terms of S/N and contrast, with the more commonly used gradient-echo and spin-echo sequences. Simulations, which were confirmed by experimental results, indicate a clear superiority of this sequence in terms of S/N. As the sequence can be used with short TRs (like a gradient-echo sequence), it is perfectly adapted to 3D imaging, unlike spin-echo sequences which require long TRs. A high S/N is obtained with the magnetization maintained in the transverse plane, thanks to a zero gradient sum at the end of each TR. This is much better than 'spoiling', which is used

The TrueFISP sequence provides a much higher signal-to-noise ratio than a 2D accelerated spin-echo sequence. This is associated with better spatial resolution in the third dimension. Yet, the RARE sequence can also be used in 3D and also provides very high S/Ns. However, additional spatial encoding for the T2-weighted RARE spin echo has a drastic effect on the

longitudinal follow-up of glioma volume in a given animal model.

sequence can be useful in imaging small animals at high fields.

with the more common gradient-echo or FLASH sequences.

**5. Discussion and conclusion** 

45 min) [8,15].

On the other hand, with the 2D RARE sequence, the tumor is also visible from day 7 after implantation. However, the volume measured based on images acquired with the RARE sequence is always greater than that measured by TrueFISP, and, importantly, very little difference is observed over the first 4 analysis times (d9, d11, d14, d16). This is probably due to a less precise measurement of the tumor volume, which, as indicated above, is due to very low spatial resolution in the third dimension.

Fig. 13. 3D reconstruction of the tumor by semi-automatic segmentation from the images shown in Figure 11. Image a: reconstruction of the whole brain (gray) and the tumor (red) at d9. Images bj: 3D representation of the tumor over time. The same arbitrary scale is used in all images.

Fig. 14. Progression of tumor volume over time based on the volumes shown in Figure 13.

#### **5. Discussion and conclusion**

140 Advances in the Biology, Imaging and Therapies for Glioblastoma

On the other hand, with the 2D RARE sequence, the tumor is also visible from day 7 after implantation. However, the volume measured based on images acquired with the RARE sequence is always greater than that measured by TrueFISP, and, importantly, very little difference is observed over the first 4 analysis times (d9, d11, d14, d16). This is probably due to a less precise measurement of the tumor volume, which, as indicated above, is due to

Fig. 13. 3D reconstruction of the tumor by semi-automatic segmentation from the images shown in Figure 11. Image a: reconstruction of the whole brain (gray) and the tumor (red) at d9. Images bj: 3D representation of the tumor over time. The same arbitrary scale is used in

all images.

very low spatial resolution in the third dimension.

The aim of this chapter was to show how the 3D TrueFISP sequence can improve longitudinal follow-up of glioma volume in a given animal model.

This sequence is rarely used at high magnetic fields as it generates numerous artifacts known as banding artifacts [10-12]. However, the S/N and contrast provided by this sequence can be useful in imaging small animals at high fields.

It has been shown that these artifacts can be easily corrected at 4.7 T and 9.4 T when studying mouse brains, and also for areas which are much more sensitive to movements and susceptibility artifacts, such as the heart [8,13]. To do so, the so-called 'alternating-phase RF pulse' method is used, in combination with reconstruction through calculation of the square root of the sum of squares of the magnitude signal [14]. As shown here, 4 phase steps are necessary to suppress artifacts at 4.7 T, while at least 8 steps are required at 9.4 T [8]. The resulting image appears perfectly homogeneous in terms of signal, and no longer displays artifacts. The requirement for 4 images increases the total acquisition time, but this nevertheless remains much shorter than with a T2-weighted 3D RARE sequence (12 min vs. 45 min) [8,15].

The 3D TrueFISP imaging sequence was compared, in terms of S/N and contrast, with the more commonly used gradient-echo and spin-echo sequences. Simulations, which were confirmed by experimental results, indicate a clear superiority of this sequence in terms of S/N. As the sequence can be used with short TRs (like a gradient-echo sequence), it is perfectly adapted to 3D imaging, unlike spin-echo sequences which require long TRs. A high S/N is obtained with the magnetization maintained in the transverse plane, thanks to a zero gradient sum at the end of each TR. This is much better than 'spoiling', which is used with the more common gradient-echo or FLASH sequences.

The TrueFISP sequence provides a much higher signal-to-noise ratio than a 2D accelerated spin-echo sequence. This is associated with better spatial resolution in the third dimension. Yet, the RARE sequence can also be used in 3D and also provides very high S/Ns. However, additional spatial encoding for the T2-weighted RARE spin echo has a drastic effect on the

3D TrueFISP MRI Provides Accurate Longitudinal Measurements of Glioma Volumes in Mice 143

[1] Bock NA, Zadeh G, Davidson LM, et al. High-resolutionlongitudinal screening with

[2] Moats RA, Velan-Mullan S, Jacobs R, et al. Micro-MRI at 11.7T of a murine brain tumor

[3] Miraux S, Lemiere S, Pineau R, et al. Inhibition of FGF receptor activity in glioma

[4] Sun Y, Mulkern RV, Schmidt K, et al. Quantification of water dif- fusion and relaxation times of human U87 tumors in a mouse model. NMR Biomed 2004;17:399–404. [5] Quesson B, Bouzier AK, Thiaudiere E, Delalande C, Merle M, Ca- nioni P. Magnetization

using a binary spin-bath model. J Magn Reson Imaging 1997;7:1076–1083. [6] Hennig J, Nauerth A, Friedburg H. RARE imaging: a fast imaging method for clinical

[7] Oppelt A, Graumann R, Barfu H, Fischer H, Hartl W, Schajor W. FISP—a new fast MRI

[8] Miraux S, Massot P, Ribot EJ, Franconi JM, Thiaudiere E. 3D TrueFISP imaging of mouse

[9] De Coene B, Hajnal JV, Gatehouse P, Longmore DB, White SJ, Oatridge A, Pennock

[10] Haacke EM, Wielopolski PA, Tkach JA, Modic MT. Steady-state free precession imaging

[11] Zur Y, Wood ML, Neuringer LJ. Motion-insensitive, steady-state free precession

[12] Vasanawala SS, Pauly JM, Nishimura DG. Linear combination steady-state free

[13] Miraux S, Calmettes G, Massot P, Lefrancois W, Parzy E, Muller B, Arsac LM,

[15] Natt O, Watanabe T, Boretius S, Radulovic J, Frahm J, Michaelis T. High-resolution 3D

JM, Young IR, Bydder GM. MR of the brain using fluid-attenuated inversion recovery (FLAIR) pulse sequences. AJNR Am J Neuroradiol. 1992 Nov-

in the presence of motion: application for im- proved visualization of the

precession MRI. Magn Reson Med 2000;43:82–90. 11. Casselman JW, Kuhweide R, Deimling M, Ampe W, Dehaene I, Meeus L. Constructive interference in steady state-3DFT MR imaging of the inner-ear and cerebellopontine angle. AJNR Am J

Deschodt-Arsac V, Franconi JM, Diolez P, Thiaudie`re E. 4D retrospective black blood trueFISP imaging of mouse heart. Magn Reson Med 2009; 62: 1099–1105. [14] Vasanawala SS, Pauly JM, Nishimura DG. Linear combination steady-state free

MRI of mouse brain reveals small cerebral structures in vivo. J Neurosci Methods

brain at 4.7T and 9.4T. J Magn Reson Imaging 2008;28:497–503.

magnetic resonance imaging in a murine brain cancer model. Neoplasia 2003;5:546–

model using delayed contrast enhancement. Mol Imaging 2003;2:150–158 [erratum:

implanted into the mouse brain using the tetra- cyclin-regulated expression system.

transfer fast imaging of implanted glioma in the rat brain at 4.7T: interpretation

**6. References** 

554.

Mol Imaging 2003;2(4)].

Dec;13(6):1555-64.

Neuroradiol 1993;14:47–57.

2002;120:203–209.

Angiogenesis 2004;7:105–113.

MR. Magn Reson Med 1986;3:823–833.

sequence. Electromedica 1986;54:15–18.

cerebrospinal fluid. Radiology 1990;175: 545–552.

precession MRI. Magn Reson Med 2000; 43: 82–90.

imaging. Magn Reson Med 1990;16:444 – 459.

total acquisition time: good T2 weighting is required for tumor detection, and can only be obtained with TRs greater than 3,500 ms. This leads to a total acquisition time on the order of 1 hour, whereas around 10 minutes is adequate with the 2D sequence or the 3D TrueFISP sequence.

The TrueFISP sequence provides adequate contrast, allowing even very small tumors to be detected clearly, and semi-automatically segmented using generic image-analysis tools. However, as shown in both simulations and images, the contrast between the healthy brain and the tumor remains lower than that provided by the spin-echo sequence. This is due to TrueFISP-sequence weighting, which relies on the T1/T2 ratio, while the RARE sequence only uses T2 weighting. This has several consequences. In some cases, e.g. when the tumor becomes very large and heterogeneous, semi-automatic segmentation becomes difficult due to inadequate contrast with the healthy brain. Manual segmentation therefore becomes necessary, which may be both tedious and subjective. Nevertheless, the error rate when estimating even large volumes remains low.

Another consequence is that it is almost impossible to detect a glioma in the brain at fields higher than 4.7 T based on natural contrast. This is because the T1 and T2 values for the brain and tumor become almost identical, which limits the contrast at very high magnetic fields. The T2-weighted RARE sequence still works perfectly at these higher fields.

Thus, for this type of application, the TrueFISP sequence at 9.4 T or greater provides little advantage over other, more commonly used sequences. However, at low fields, the S/N and contrast provided by the TrueFISP sequence make it particularly interesting. It has already been shown with specific instruments that it allows tumors to be easily detected in mice based on natural contrast, and that it provides an excellent signal-to-noise ratio at clinical magnetic fields (1.5 T and 3 T) [16,17].

In terms of spatial resolution, the advantages of 3D compared with 2D imaging are obvious. Three-dimensional imaging results in much smaller voxel sizes with comparable acquisition times, while maintaining a high S/N. This requires the addition of a phase-encoding table in the slice direction, i.e. a much greater number of lines read in the Fourier volume. The advantages of sequences with short TRs thus become obvious. As shown, the TrueFISP sequence also provides a much greater precision when measuring tumor volume, in particular for small tumors. With the 2D RARE sequence, it is impossible to precisely measure volume changes in the first days following tumor implantation. With the 3D TrueFISP sequence, this information is readily available.

In addition, as the sequence is relatively rapid in terms of total acquisition time, it can be repeated at very short intervals to follow tumor progression, even in relatively fragile mice. The total examination, between preparation of the animal and imaging, lasts less than 20 minutes. This makes it possible to study animals on a daily basis. This is much more difficult when contrast agents must be injected, or with acquisition times of around 1 hour, as with 3D RARE sequences [15].

To conclude, the 3D TrueFISP sequence can be easily used to follow tumor progression in a small animal model imaged at 4.7 T. Thanks to a particularly high S/N, artifact-free images can be acquired, with excellent spatial resolution and good tumor/healthy brain contrast. The total acquisition time remains under 15 minutes, thus offering precise longitudinal follow-up of tumor volume. Thus, this sequence could be used to noninvasively validate the efficacy of new genetic or pharmacological treatments for glioma.

At lower magnetic fields, the sequence has also demonstrated its efficacy. In contrast, it is currently not applicable at 9.4 T or higher to precisely measure tumor volumes.
