This document describes a technique for temperature mapping on a 0.3-T open MRI system using proton frequency shift. An image-based three-point correction is used to stabilize successive temperature maps by eliminating spatial and offset fluctuations due to hardware instability. The technique was tested by monitoring temperature changes in a ham during laser ablation over 150 seconds. Temperature maps generated before and after the three-point correction showed that the correction eliminated image distortions and yielded stable temperature distributions.
This document describes a technique for temperature mapping on a 0.3-T open MRI system using proton frequency shift. An image-based three-point correction is used to stabilize successive temperature maps by eliminating spatial and offset fluctuations due to hardware instability. The technique was tested by monitoring temperature changes in a ham during laser ablation over 150 seconds. Temperature maps generated before and after the three-point correction showed that the correction eliminated image distortions and yielded stable temperature distributions.
This document describes a technique for temperature mapping on a 0.3-T open MRI system using proton frequency shift. An image-based three-point correction is used to stabilize successive temperature maps by eliminating spatial and offset fluctuations due to hardware instability. The technique was tested by monitoring temperature changes in a ham during laser ablation over 150 seconds. Temperature maps generated before and after the three-point correction showed that the correction eliminated image distortions and yielded stable temperature distributions.
T. Takabashi,’ K. Komura,’ M. Dohi; J. Harada* ‘R&D Center, Hitachi Medical Corp., Chiba, Japan. ‘Dep. of Radiology, Tbe Jikei University School of Medicine, Kashiwa Hospital, Chiba, Japan
Introduction temperature shift of the materials used in the magnetic
Temperature mapping is used to monitor laser and RF ablation circuit. The three-point correction eliminated the distortion processes, microwave therapy, and so on [l], and is usually done by and yielded the homogeneous temperature distribution continually evaluating Tl weighted signal intensity, proton expected. Transaxial, sagittal and coronal temperature frequency shift or a diffusion constant [2]. There are many reports maps were stabilized by using the correction. of temperature mapping on medium-field and high-field superconducting-magnet MRI systems 131,but a few of temperature Experimental results mapping on low-field and medium-field open MRI systems. And A ham of 10 cm in diameter was heated by a laser fiber inserted to a for permanent-magnet open MRI systems, there are only depth of 3 cm. The laser intensity was 5, 8, or 10 W/pulse. Each preliminary reports relating changes in signal intensity change to pulse lasted 1 s and there was 1 s between pulses. The duration of object temperature [4]. Here we report a temperature mapping the laser-pulse irradiation was 150 s. MR images ware acquired technique that is based on proton frequency shift and is optimized continually during and after laser irradiation and temperature maps for an open MRl system with a 0.3-T permanent magnet. ‘Ibis were obtained every 15 s. technique uses image-based three-point-correction for successive Figure 2 shows a temperature map obtained during laser ablation. temperature map in order to stabilize one. The temperature change during 5-W laser irradiation is displayed in Fig. 3, which shows both the temperature increase during Method irradiation and the temperature decrease after irradiation. The A Hitachi AIRIS-II MRI system was modified for dynamic standard deviation of the temperature fluctuation after the cooling temperature mapping by installing signal-processing and color- (from the 25th image to the 51st image) was 2.3 degree. map-display software on its control workstation. The signal- processing software provided complex image subtraction and Conclusions transformation from a phase map to a temperature map. Temperature mapping based on proton frequency shift was Tbe interimage signal fluctuation due to hardware instability was optimized for a 0.3-T permanent-magnet open MRI system, and an reduced by using an image-based three-point-correction that image-based three-point-correction was used. eliminated spatial 1St-order fluctuation and offset fluctuation. The true temperature change AT,,,(x,y) was obtained by using the References measured temperature AT,&x, y) and a fluctuation component [I] R Steiner et aI., Radiology, vol. ZOS, pp. 803-810, 1998. AL&~ Y) : [2] H.E. Cline et al., 4th ISMRM, p. 1754,1996. [3] Y. Ishihara et al., 11th SMRM, p. 4803,1992. 4,,, (KY) = AT,,,& Y) - ATnuct(x,Y) (1) [4] 7: J. Russer et al., 5th ISMRM, p. 1958,1997. The AT&x, y) distribution, defined in Eq. (2) was obtained from the temperature measured at least three points that were not heated : . AT,,,, (x, Y) = ax + by + c, (2) where a, b and c are constants. Tbe three reference points on a reference image were selected by an operator before the mapping, and the ATDuct(x, y) value was updated continually during the mapping. Thus the measured temperatures could be successively corrected using updated AT&x, y) values and a corrected _-_ 17n temperature color map could be displayed in real time. (a) without correction (b) with correction We acquired temperature images using a gradient echo Fig. 1. Temparature distribution of a water phantom. sequence ( TR/TE = 80 ms/30 ms, flip angle = 60 degree, FOV = 200x200 mm, slice thickness = 8 mm, matrix size = ham 128x128, data acquisition = 1, and imaging time = 10.2 s). Fig. 2. laser fiber This sequence is almost same as the one for temperature mapping using the change in signal intensity on a 0.3-T Temperature map MRI system [4]. Long TR of the sequence, results in a during laser ablation. better signal-to-noise ratio. And because of long TE, the phase change due to proton frequency shift becomes larger. This compensates the smaller phase change due to the - 30.0 resonance frequency in 0.3-T MRI being lower than that in e25.0 MRI at 0.5T-1.5T. Figure 1 shows examples of corrected % G 20.0 and uncorrected temperature distributions of a water 2 o 15.0 phantom that was not heated. These distributions were E generated from 51 successive images obtained over 12 min. 2B 10.0 As shown in Fig. l(a), before the three-point-correction the g 5.0 images were distorted. The distortion varied image-by- g 0.0 image because of hardware condition, such as the 0 10 20 30 40 50 60 image number Fig. 3. Temperature change during and after laser ablation.