http://informahealthcare.com/hth ISSN: 0265-6736 (print), 1464-5157 (electronic) Int J Hyperthermia, 2014; 30(7): 447–455 ! 2014 Informa UK Ltd. DOI: 10.3109/02656736.2014.966337

RESEARCH ARTICLE

Fast conductivity imaging in magnetic resonance electrical impedance tomography (MREIT) for RF ablation monitoring Oh In Kwon1, Munish Chauhan2, Hyung Joong Kim2, Woo Chul Jeong2, Hun Wi2, Tong In Oh2 & Eung Je Woo2 Department of Mathematics, Konkuk University, Seoul, Korea and 2Department of Biomedical Engineering, Kyung Hee University, Yongin, Korea

Int J Hyperthermia Downloaded from informahealthcare.com by Yale Dermatologic Surgery on 01/05/15 For personal use only.

1

Abstract

Keywords

Purpose: This study shows the potential of magnetic resonance electrical impedance tomography (MREIT) as a non-invasive RF ablation monitoring technique. Materials and methods: We prepared bovine muscle tissue with a pair of needle electrodes for RF ablation, a temperature sensor, and two pairs of surface electrodes for conductivity image reconstructions. We used the injected current non-linear encoding with multi-echo gradient recalled echo (ICNE-MGRE) pulse sequence in a series of MREIT scans for conductivity imaging. We acquired magnetic flux density data induced by externally injected currents, while suppressing other phase artefacts. We used an 8-channel RF head coil and 8 echoes to improve the signal-to-noise ratio (SNR) in measured magnetic flux density data. Using the measured data, we reconstructed a time series of 180 conductivity images at every 10.24 s during and after RF ablation. Results: Tissue conductivity values in the lesion increased with temperature during RF ablation. After reaching 60  C, a steep increase in tissue conductivity values occurred with relatively little temperature increase. After RF ablation, tissue conductivity values in the lesion decreased with temperature, but to values different from those before ablation due to permanent structural changes of tissue by RF ablation. Conclusion: We could monitor temperature and also structural changes in tissue during RF ablation by producing spatio-temporal maps of tissue conductivity values using a fast MREIT conductivity imaging method. We expect that the new monitoring method could be used to estimate lesions during RF ablation and improve the efficacy of the treatment.

Continuous monitoring, magnetic resonance electrical impedance tomography (MREIT), radiofrequency ablation, temperature distribution, tissue property imaging

Introduction Image-guided thermal interventional treatments such as radiofrequency (RF) ablation, laser ablation, microwave ablation, and cryoablation are less invasive than surgical resections of hepatocellular carcinoma and metastasis [1–3]. RF ablation, especially, has been rapidly accepted in surgical practice and provided reliable outcomes to treat cancers. However, it still suffers from occasional local recurrence after ablation [4]. To improve the efficacy of RF ablation without adverse effects and reduce the local recurrence rate, noninvasive monitoring is desirable for lesion estimation [5]. Thermal conduction during RF ablation is affected by electrical and thermal tissue properties and also structural factors such as vascularity and blood flow [6]. Due to the complicated interplay of these factors, lesion estimation from a set of RF ablation parameters is difficult. It is therefore highly desirable to monitor temperature distributions during RF ablation treatments and evaluate lesions. This will help us to determine optimal electrode positions and estimate the Correspondence: Tong In Oh, PhD, Department of Biomedical Engineering, Kyung Hee University, 1 Seocheon Giheung, Yongin, Gyeonggi, Korea, 446-701. Tel: +82-31-201-3727. Fax: +82-31-2012378. E-mail: [email protected]

History Received 14 April 2014 Accepted 13 September 2014 Published online 20 October 2014

effectiveness of the applied RF power [7,8]. The temperature mapping can also prevent overheating taking place in the surrounding normal tissue. Several imaging techniques including ultrasound, computed tomography (CT), magnetic resonance imaging (MRI), and positron emission tomography (PET) have been used for planning, targeting, monitoring, controlling, and assessing lesions [9–11]. MRI has advantages over other techniques since it is non-invasive and produces images in any orientation with high spatial resolution [12,13]. Temperature mappings using MRI have been tried based on the proton resonance frequency (PRF) shift [14–16], diffusion coefficient [17], magnetisation transfer [18], and temperature-sensitive contrast agent [19]. However, these methods are not widely used yet due to their weak contrast to discern ablated lesions [20,21]. The electrical conductivity is determined by concentration and mobility of ions in intra- and extra-cellular fluids [22,23]. It changes with temperature and also structural factors such as tissue compositions and cellular structures [24–26]. Therefore, we suggest conductivity imaging as a tool to monitor RF ablation treatments. Magnetic resonance electrical impedance tomography (MREIT) is a non-invasive technique to produce conductivity images using internal measurements of induced magnetic flux

Int J Hyperthermia Downloaded from informahealthcare.com by Yale Dermatologic Surgery on 01/05/15 For personal use only.

448

O. I. Kwon et al.

Int J Hyperthermia, 2014; 30(7): 447–455

density data subject to externally injected low-frequency currents [27–29]. Compared with electrical impedance tomography (EIT) using boundary measurements of induced voltage data [30–41], it produces high-resolution conductivity images with about 1-mm pixel size. Though we recently showed the feasibility of temperature mapping using MREIT through phantom experiments [42], there are several technical issues to be addressed before clinical applications. First, we should produce conductivity images during RF ablation with enough temporal resolution. Second, we should validate the relation between the conductivity change and the temperature change. Effects of structural tissue changes must be also properly interpreted. Third, we should develop a lesion estimation algorithm based on the reconstructed conductivity images and link it to numerous RF ablation parameters for best possible outcome. In this paper we address the first and second issues by developing a fast MREIT imaging method for RF ablation monitoring. We describe the technical details of a series of fast MREIT scans during RF ablation. Explaining the experimental set-up using a chunk of biological tissue, we present reconstructed conductivity images of the tissue during its RF ablation. We analyse the images to validate how conductivity images are correlated with temperature changes during RF ablation. We also investigate the effects of structural tissue changes at high temperatures.

Materials and methods MREIT data collection To produce cross-sectional images of the internal conductivity distribution we inject current pulses between a pair of electrodes into the object inside an MRI scanner. The current injection time should be synchronised with a chosen MRI pulse sequence. We used the injected current non-linear encoding with multi-echo gradient recalled echo (ICNEMGRE) pulse sequence to extend the duration of current injection until the end of multiple read-out gradients

Figure 1. Schematic diagram of the ICNEMGRE pulse sequence for fast conductivity imaging using optimised magnetic flux density data.

(Figure 1). We applied crusher gradients at the end of each TR to remove any residual transverse magnetisation. The externally injected current induces a distribution of magnetic flux density B ¼ (Bx, By, Bz). When the main magnetic field of the scanner is in the z direction the z-component Bz disturbs the main field and contributes to additional dephasing of spins. Consequently, extra phases are accumulated in the MR signal. When we use the interleaved phase encoding method integrated with sequential injections of synchronised positive I+ and negative I current pulses as shown in Figure 1, it modifies the k-space data as in Equation 1. Z  S‘ ðn, mÞ ¼ ‘ ðx, yÞeiðx, yÞ eið‘ ðx, yÞ‘ ðx, yÞÞ R2 ð1Þ i2ðDkx nxþDky myÞ dxdy e where ‘ is the MR magnitude image corresponding to echo time T‘ ,  denotes the effects of the background field inhomogeneity, and  ¼ 26:75  107 rad/T sec is the gyromagnetic ratio of proton. With ‘ and ‘ , we denote the phase signals at time T‘ by the temperature-dependent water chemical shift and the induced magnetic flux density Bz, respectively. Note that ‘ ¼ T‘ B‘z and B‘z is the induced magnetic flux density Bz which can be extracted as ‘ =T‘ . The reciprocals of the field-of-view (FOV) in the x- and y-direction are denoted as Dkx and Dky , respectively. To reconstruct the conductivity distribution  whose information is imbedded in Bz, we need to extract ‘ while removing  and ‘ . We compute the discrete inverse Fourier transform of the measured k-space data S ‘ to get the complex images of Equation 2. ‘ ðx, yÞ ¼ ‘ ðx, yÞeiðx, yÞ eið‘ ðx, yÞ‘ ðx, yÞÞ : The phase signals at T‘ are P ‘ ¼  ð þ ‘  ‘ Þ,

α RF Gz Gy Gx Current

Data acq

Signal

ð2Þ

I+

I

-

‘ ¼ 1, . . . , NE

ð3Þ

RF ablation monitoring using MREIT

DOI: 10.3109/02656736.2014.966337

and we can compute ‘ as ‘ ¼

 Pþ ‘  P‘ , 2

‘ ¼ 1, . . . , NE :

ð4Þ

The magnetic flux density B‘z can be extracted from the phase ‘ , which was accumulated by Bz for the duration of T‘ , as B‘z ¼

‘ : T‘

ð5Þ

The noise standard deviation of the measured B‘z can be estimated as

Int J Hyperthermia Downloaded from informahealthcare.com by Yale Dermatologic Surgery on 01/05/15 For personal use only.

sdB‘z

1 ¼ 2T‘ ‘

ð6Þ

where ‘ is the signal-to-noise ratio (SNR) of the MR magnitude image at T‘ [43,44]. The ICNE-MGRE pulse sequence maximises the duration of the injected current until the end of the read-out gradient and decreases the noise standard deviation of the measured Bz data [45]. Optimal algorithm to extract Bz from multi-echo MR phase signals The magnetic flux density Bz can be calculated as a weighted average of the measured data B‘z for ‘ ¼ 1, . . . ,P NE in the NE ‘ ICNE-MGRE pulse sequence. We express B ¼ z ‘¼1 !‘ Bz P NE with !‘ 4 0 and ‘¼1 !‘ ¼ 1 and assume that the measured k-space signals are contaminated with independent identically distributed (IID) complex Gaussian random noise. Then, the noise variance of Bz is VarðBz Þ ¼

NE X

NE   X !2‘ Var B‘z /

!2  ‘ 2 : 2  ‘¼1 T‘ ‘

‘¼1

Time-varying conductivity image reconstruction algorithm Let V be a three-dimensional object with its boundary @. We assume that the conductivity (r, t) at position r ¼ (x, y, z) and time t is isotropic and time-varying. When we inject current I through a pair of electrodes on @, there occur distributions of voltage, current density, and magnetic flux density in V. The induced voltage u(r, t) satisfies the following partial differential equation: r  ððr, tÞruðr, tÞÞ ¼ 0 in 

@

Note that the magnetic flux density Bz includes information on the conductivity  and the spatial change of the voltage u. In MREIT experiments we repeated the same scan twice with two orthogonal injection currents. This produced two data sets of the induced magnetic flux densities Bz1 and Bz2 subject to the horizontal and vertical injection currents, respectively. There are several conductivity image reconstruction algorithms in MREIT using the measured Bz1 and Bz2 data [28,29]. In this paper we used the transversal J-substitution algorithm since it differentiates the measured Bz1 and Bz2 data only once [47]. We first considered a homogeneous conductivity distribution H and numerically solve Equation 9 with H in place of (r, t). Using Equation H 10 we can compute the magnetic flux densities BH z, 1 and Bz, 2 subject to the horizontal and vertical injection currents, respectively. We can then recover the conductivity image as in Equation 11.   P2 ~ H 1 n¼1 r~? Bz, n ðr, tÞ  BH z, n ðrÞ  run ðrÞ H ðr, tÞ ¼  ðrÞ  P2 ~ H ðrÞ, ru ~ H ðrÞi 0 hru n¼1

‘¼1

ðr, tÞruðr, tÞ   ¼ gðrÞ on @ and Z uðr, tÞds ¼ 0

where  ¼ ð1 , 2 , 3 Þ is the outward unit normal vector on @, ds is a surface element, and g is the current density on @. The current density in  is expressed as Jðr, tÞ ¼ ðr, tÞruðr, tÞ and the magnetic flux density Bz is given by the Biot-Savart law as ( ) 0 , tÞ ðx  x0 Þðr0 , tÞ @uðr @y 0 Z , tÞ ðy  y0 Þ @uðr 0 @x Bz ðr, tÞ ¼ dr0 , r 2 : 4 R3 jr  r0 j3 ð10Þ

ð7Þ

An optimal weighting factor !‘ for each pixel can be obtained as in Nam and Kwon [46].  2 T‘2 ‘  : ð8Þ !‘ ¼ N PE 2   2 T‘ ‘

ð9Þ

449

n

n

ð11Þ where r~? f :¼  for a given scalar function f and h, i denotes the scalar inner product. More details of the image reconstruction algorithm are described by Nam and Kwon [46]. @f ð@y ,

@f @xÞ

Experimental methods We prepared a chunk of bovine femoral muscle of about 100 mm in diameter (Figure 2A). For RF ablation at 475 kHz, we used the RF 2000 generator (RadioTherapeutics, Sunnyvale, CA, USA) and a pair of non-magnetic pure silver wire electrodes (Figure 2B). These bipolar RF ablation electrodes, length 200 mm and diameter 2 mm, were insulated with polyolefin tubes except 10 mm of their tips. We inserted the RF ablation electrodes into the tissue so that their tips were positioned at the imaging slice. The distance between the electrode tips was 20 mm. To directly measure tissue temperature we placed a temperature sensor inside the tissue between the tips of the ablation electrodes at the imaging slice. To inject dc current pulses for conductivity imaging we attached two pairs of carbon-hydrogel electrodes (HUREV, Wonju, Korea) at the middle surface of the tissue. Figure 2(C) illustrates the experimental setting. We placed the prepared tissue in a 3-T MRI scanner (Achieva, Philips, Eindhoven, the Netherlands) with an

450

O. I. Kwon et al.

Int J Hyperthermia, 2014; 30(7): 447–455

Temperature probe

RF ablation electrodes

Bovine muscle

40mm 10mm

10mm

(A)

10mm

Int J Hyperthermia Downloaded from informahealthcare.com by Yale Dermatologic Surgery on 01/05/15 For personal use only.

20mm

(B) Temperature probe Bovine muscle

RF ablation electrodes

Imaging electrode

Imaging electrode

Temperature measurement x 20mm

(C)

(D)

Figure 2. Experimental set-up to monitor RF ablation of ex vivo bovine muscle tissue using MREIT. (A) Chunk of bovine muscle tissue with a pair of RF ablation electrodes and a temperature probe for direct temperature measurements. (B) Non-magnetic silver needle (14 gauge) electrodes for RF ablation. (C) Diagram of the electrode and sensor configurations. (D) Cross-section of the bovine muscle tissue after RF ablation with a magnified view of the lesion (inset).

8-channel RF head coil. We used a passive notch filter with its centre frequency of 128.6 MHz to remove the RF generator signals around the Larmor frequency of the 3-T MRI scanner. This prevented RF generator signals from interfering with MR signals. We turned the RF generator on for 20 min. Using the ICNE-MGRE pulse sequence, we acquired artefact-free MR images during and after RF ablation for 30.72 min. Through two pairs of surface electrodes we injected 10 mA current pulses along the horizontal and vertical directions using a custom-designed MREIT current source [48]. Imaging parameters were as follows: TR / TE ¼ 40/2.2 ms, echo spacing ¼ 2.2 ms, number of echoes ¼ 8, FOV ¼ 160  160 mm2, image matrix ¼ 64  64, number of slices ¼ 1, slice thickness ¼ 10 mm, bandwidth ¼ 550 Hz, and NSA ¼ 1. Starting from the beginning at 23.1  C room temperature, we simultaneously measured the tissue temperature at the middle of the ablation electrodes using the installed temperature sensor throughout the experiment. After the experiment we cut the tissue and observed its cross-section to confirm the formation of a lesion (Figure 2D).

Data processing methods We collected 180 image data sets at 10.24 s intervals for the total data acquisition time of 30.72 min. From each image data set we extracted the induced magnetic flux densities B1z and B2z subject to the horizontal and vertical injection currents, respectively. We applied the weighted averaging method to the acquired multi-echo MR phase signals to maximise the SNR in B1z and B2z . We reconstructed conductivity images from the measured B1z and B2z data using (11). From the time series of 180 conductivity images at 10.24 s intervals, we extracted spatiotemporal conductivity profile maps to show how conductivity values in the lesion changed during and after RF ablation. To quantitatively analyse the time-dependent conductivity images, we used the relative conductivity contrast ratio (rCCR) defined as   lesion  background  rCCR ¼  100 ½% ð12Þ background where lesion and background represent the reconstructed conductivity values of the lesion and a background region, respectively [49].

RF ablation monitoring using MREIT

DOI: 10.3109/02656736.2014.966337

accumulate the extra phase – that is, 1 ¼ Bz T1 was small. The image quality also deteriorated for the echoes after the fifth echo due to the effects of T2 decay. We can clearly see in Figure 3(C) that the optimal weights !‘ using Equation 8 take care of the different amounts of noise in each echo and combine all echo signals in the optimal way. Figure 4 illustrates how we could improve the SNR in measured Bz images by using multiple echo signals and the optimal weights !‘ in Equation 8. Figures 4(A–D) show the extracted Bz images subject to the horizontal injection current

Results Each image data set at 10.24 s intervals included 8 echo signals using the ICNE-MGRE pulse  sequence. Figure 3(A) shows the MR magnitude images  ‘  acquired at each echo time T‘ ¼ 2:2  ‘ ms for ‘ ¼ 1, . . . , 8. Figure 3(B and C) are the images of the induced magnetic flux density B‘z in Equation 5 and the weights in Equation 8, respectively. The first image in Figure 3(B) is B1z , which was obtained from the first echo at 2.2 ms echo time. It was noisy because the externally injected currents did not have enough time to

Int J Hyperthermia Downloaded from informahealthcare.com by Yale Dermatologic Surgery on 01/05/15 For personal use only.

(A)

451

100

(1)

(2)

(3)

(4)

(6)

(5)

(7)

(8)

0 Ts

(B) 5E-6

-5E-6

0.4

(C)

0

Figure 3. Example of acquired 8-echo MR images at every 10.24-s intervals. (A) Magnitude images of 8 echoes. (B) Magnetic flux density images of 8 echoes. (C) Optimal weights to be used to combine all images in (B) to compute the optimal magnetic flux density image from all 8 echo signals.

Ts 8E-6

-8E-6

(A)

(B)

(C)

(D)

Ts 8E-6

(E)

(F)

(G)

(H)

-8E-6

Figure 4. Effects of combining multiple echo signals on the quality of Bz data. (A), (B), (C), and (D) are measured Bz data using 2, 4, 6, and 8 echoes, respectively, for the horizontal injection current. (E), (F), (G), and (H) are corresponding Bz data for the vertical injection current.

452

O. I. Kwon et al.

Int J Hyperthermia, 2014; 30(7): 447–455

ROIV

S/m 0.5

ROIH 0.25 0

600

1200

(B) 0

(A)

Temperature map 70 60

Power (W)

50

Temprature (OC)

600

40 30

Time (s)

20

1200

Int J Hyperthermia Downloaded from informahealthcare.com by Yale Dermatologic Surgery on 01/05/15 For personal use only.

Time (s)

10 0 0

200

400

600 800 Time (s)

1000

1200

S/m 0.7

0.25 (C)

(D)

Figure 5. (A) MR magnitude image with two ROIs in the imaging slice where two ablations electrodes were located. (B, C) Spatiotemporal profile maps of the reconstructed conductivity values in two ROIs. (D) Plots of the directly measured tissue temperature and the output power of the RF generator.

using 1, 4, 6, and 8 echoes, respectively. Figure 4(E–H) are the corresponding Bz images subject to the vertical injection current. It was important to use all 8 echo signals to maximise the SNR in Bz images using the optimal weights !‘ in Equation 8. From 180 conductivity images reconstructed at 10.24 s intervals we produced the spatio-temporal profile maps of the reconstructed conductivity values. Figure 5(A) is the MR magnitude image of the slice including the tips of the ablation electrodes. We chose two regions of interest marked as ROIV and ROIH. The ROIV was a vertical region passing through the middle of two ablation electrodes. The ROIH was a horizontal region passing through the horizontal line connecting two ablation electrodes. Figure 5(B and C) illustrate how reconstructed conductivity values changed within ROIV

and ROIH during and after RF ablation. For easier comparison, in Figure 5(D), we plotted the RF output power and the directly measured tissue temperature in the lesion. We found that the results in Figure 5 were consistent with other simulation and experimental results of RF ablation [8,50]. To quantitatively analyse the conductivity changes during and after RF ablation we calculated rCCR values defined in (12) between the lesion and the background region far away from the ablation electrodes. Inside the lesion we chose the pixel at the middle of two ablation electrodes, which was also very close to the position where we placed the temperature sensor for direct temperature measurements. Figure 6(A) plots the computed rCCR during and after RF ablation together with the directly measured tissue temperature and the output RF power.

RF ablation monitoring using MREIT

DOI: 10.3109/02656736.2014.966337 50

40

50 40 30 20

Power (W)

rCCR (%)

70 60

100

Temperature rCCR 80

30

60

20

10

40

0

20

10 0

I 0

II III IV 200

400

600

V

VI

90

III

V

60 50

II

40

VI

30 20 10 0

800 1000 1200 1400 1600 1800 2000

I 20

30

40

50

60

70

Temperature ( C)

Time (s)

Int J Hyperthermia Downloaded from informahealthcare.com by Yale Dermatologic Surgery on 01/05/15 For personal use only.

IV

80 70

rCCR (%)

Power

80

100

(B)

90

Temperature difference ( C)

(A) 100

453

Figure 6. (A) Plots of the rCCR, directly measured tissue temperature, and RF output power. (B) Plot of the rCCR versus the tissue temperature. The six different stages marked as I to VI are defined in Table 1. The rCCR changed with the tissue temperature and also the structural factors such as tissue compositions and cellular structures. The rCCR value after RF ablation was larger than the initial value by about 20% due to permanent structural changes in the lesion by RF ablation.

The RF output power was self-controlled by the RF generator using the two-terminal impedance measurement method. The measured tissue temperature rapidly increased at the beginning following the rapid increase of the output RF power to 11 W. After reaching about 60  C, the RF generator gradually decreased its power and the tissue temperature stabilised round 60  C from 163.84 to 348.16 s. When the RF power dropped to 3 W at 532.48 s, the temperature decreased exponentially and approached room temperature at 1300.48 s. After the RF generator was turned off at 1239.04 s, the tissue temperature remained at room temperature. We divided the 30.72 min of the experimental period into six stages as defined in Table 1. Figure 6(B) plots the rCCR as a function of the directly measured tissue temperature. During stage I, the rCCR increased almost linearly with the tissue temperature up to 60  C. Since structural changes inside the tissue were small in this early stage, the increased conductivity values, which were reflected in the increased rCCR values, stemmed from the increased tissue temperature. The temperature coefficient of the rCCR was 0.67%/  C during this stage, which correlated with the reported values in previous studies [24,25]. In stage II, the temperature was kept at about 60  C by automatically decreasing the RF power to 8 W. However, the rCCR still increased rapidly. This means that the tissue conductivity increase stemmed from the changes in cellular morphology and in the amounts of intra- and extra-cellular fluids [24,25]. Chauhan et al. [7] also reported that the average rCCRs of seven liver phantoms increased in both the coagulation necrosis and hyperaemic rim produced by RF ablation. In stage III the RF power was maintained at 8 W and the tissue temperature increased a little above 60  C. The rCCR was still increasing and we infer that the tissue conductivity increased by both the temperature increase and the structural changes. In stage IV, the RF power dropped to 3 W and the tissue temperature rapidly decreased. However, the rCCR was maintained at about 87.8%. This means that some structural changes still occurred, which tended to increase the tissue

conductivity. This should have been opposed by the temperature decrease, which tended to decrease the conductivity. In stage V, the RF power was maintained at 3 W and the temperature decayed slowly to room temperature. Since the rCCR decreased rapidly we infer that the structural changes stopped and the reduced temperature pulled the rCCR down to lower values. In stage VI, the RF generator was turned off and the tissue temperature remained at room temperature. We can observe that the rCCR decreased slowly to its final value of 20.92%, which means that the tissue in the lesion was structurally different from the original tissue before RF ablation. We infer from this that the structural changes were permanent.

Discussion MR phase imaging with the PRF method can be used to monitor temperature changes in an RF ablation lesion [51]. Since it uses a gradient echo-based MR pulse sequence, it is weak against MR phase artefacts. In this paper we adopted the conductivity imaging method in MREIT to monitor RF ablation. We implemented the ICNE-MGRE pulse sequence with the interleaved data collection protocol, which cancelled out MR phase artefacts such as the background field inhomogeneity effects. To produce conductivity images at every 10.24 s instead of 3 min in our previous study [42], we implemented the fast MREIT conductivity imaging technique. Using 8 echo signals and the optimal weights to average the acquired signals, we could increase the SNR in measured magnetic flux density data, which were used in subsequent conductivity image reconstructions. From the spatio-temporal analyses of the time series of 180 conductivity images we found that the overall change of tissue conductivity in the lesion correlated well with the overall change of tissue temperature at the same locations. We should, however, note that structural changes in tissue during RF ablation also altered the conductivity values considerably. We found that their effects can be separately observed from

454

O. I. Kwon et al.

Int J Hyperthermia, 2014; 30(7): 447–455

Table 1. Definitions of six intervals shown in Figure 6. Stage I

II

III

Int J Hyperthermia Downloaded from informahealthcare.com by Yale Dermatologic Surgery on 01/05/15 For personal use only.

IV

V VI

Duration (s)

Description

0–163.84

RF generator was on and its power increased to 11 W. Temperature increased from 23.1 to 59.2  C. rCCR increased with temperature. 174.08–348.16 RF generator was on and its power decreased from 11 to 8 W. Temperature was maintained at about 60  C. rCCR increased due to structural tissue changes. 358.40–471.04 RF generator was on and its power was maintained at 8 W. Temperature gradually increased to 66.6  C. rCCR increased with temperature. 481.28–808.96 RF generator was on and its power decreased to 3 W. Temperature rapidly decreased to 35.7  C. rCCR was maintained at about 87.8% due to structural changes. 819.20–1208.32 RF generator was maintained at 3 W. Temperature gradually decreased to 23.7  C. rCCR decreased with temperature. 1218.56–1843.20 RF generator was off. Temperature remained at 23.1  C. rCCR slowly decreased to values different from those before ablation.

the temperature effects by dividing the entire RF ablation period into six intervals and interpreting the time-dependent changes within those intervals. In this paper we used the transversal J-substitution algorithm for conductivity image reconstructions. Since it produced conductivity contrast images, we used the rCCR to interpret the conductivity changes. Future studies should include a fast absolute conductivity imaging method, which can lead to absolute temperature mapping. We also suggest implementing dual-frequency conductivity imaging by combining MREIT and MREPT [52]. The dual-frequency method can provide frequency-difference images, which may distinguish temperature effects from structural effects in a better way. The results presented in this paper indicate that real-time RF ablation monitoring is feasible by using the fast conductivity imaging method in MREIT. When combined with information on the RF output power and the directly measured temperature at the tip of the ablation electrode, it should be possible to estimate lesions during RF ablation. It will also help determine the effectiveness of the applied RF power and prevent overheating. Quantitative information about temperature changes and structural changes will be invaluable during RF ablation to improve its efficacy.

Conclusion In RF ablation, it is desirable to control the temperature distribution in a lesion and also in its surrounding area to improve safety and efficiency. Lesion estimation and proper control of RF power is important to reduce local recurrence after RF ablation. Lesion estimation and proper control of RF power is important to reduce local recurrence after RF ablation [53]. We showed the potential of MREIT as a noninvasive monitoring technique for RF ablation through ex vivo

experiments using a chunk of bovine muscle tissue. We showed that fast MREIT imaging is possible during RF ablation to produce conductivity images at every 10.24 s. From a time series of 180 conductivity images for 30.72 min, we could extract spatio-temporal maps of tissue conductivity values in the lesion. We found that electrical conductivity of tissue manifests temperature as well as structural changes in tissue during RF ablation. Analysing the time series of reconstructed conductivity images, we could distinguish six different stages during and after RF ablation and separately interpret temperature-dependent and/or structure-dependent conductivity changes. We suggest future experimental studies with animal and human subjects.

Declaration of interest This paper was supported by the National Research Foundation of Korea grant funded by the Korean government (MEST) (2012R1A1A2008477, 2013R1A2A2A04016066, NRF-2014R1A2A1A09006320) and a grant of the Korean Health Technology R&D Project, Ministry of Health and Welfare, Republic of Korea (HI14C0743). Oh In Kwon was supported by the Konkuk University research support programme. The authors alone are responsible for the content and writing of the paper.

References 1. Ni Y, Mulier S, Miao Y, Michel L, Marchal G. A review of the general aspects of radiofrequency ablation. Abdom Imaging 2005; 30:381–400. 2. Dodd GD, Soulen MC, Kane RA, Livraghi T, Lees WR, Yamashita Y, et al. Minimally invasive treatment of malignant hepatic tumors: At the threshold of a major breakthrough. Radiographics 2000;20:9–27. 3. Lubner MG, Brace CL, Hinshaw JL, Lee FT. Microwave tumor ablation: Mechanism of action, clinical results, and devices. J Vasc Interv Radiol 2010;21:S192–203. 4. Abdalla EK, Vauthey JN, Ellis LM, Ellis V, Pollock R, Broglio KR, et al. Recurrence and outcomes following hepatic resection, radiofrequency ablation, and combined resection/ablation for colorectal liver metastases. Ann Surg 2004;239:818–27. 5. Haemmerich D, Laeseke PF. Thermal tumour ablation: Devices, clinical applications and future directions. Int J Hyperthemia 2005; 21:755–60. 6. Ahmed M, Liu Z, Afzal KS, Weeks D, Lobo SM, Kruskal JB, et al. Radiofrequency ablation: Effect of surrounding tissue composition on coagulation necrosis in a canine tumor model. Radiology 2004; 230:761–7. 7. Chauhan M, Jeong WC, Kim HJ, Kwon OI, Woo EJ. Radiofrequency ablation lesion detection using MR-based electrical conductivity imaging: A feasibility study of ex vivo liver experiments. Int J Hyperthermia 2013;29:643–52. 8. Cosman ERJ, Cosman ERS. Electric and thermal field effects in tissue around radiofrequency electrodes. Pain Med 2005;6:405–24. 9. Goldberg SN, Grassi CJ, Cardella JF, Charboneau JW, Dodd III GD, Dupuy DE, et al. Image-guided tumor ablation: Standardization of terminology and reporting criteria. J Vasc Interv Radiol 2009;7:S377–90. 10. Dalong L, Ebbini ES. Real-time 2-D temperature imaging using ultrasound. IEEE Trans Biomed Eng 2010;57:12–16. 11. Carey RI, Leveillee RJ. First prize: Direct real-time temperature monitoring for laparoscopic and CT-guided radiofrequency ablation of renal tumors between 3 and 5 cm. J Endourol 2007;21:807–13. 12. Quesson B, de Zwart JA, Moonen CT. Magnetic resonance temperature imaging for guidance of thermotherapy. J Magn Reson Imaging 2000;12:525–33. 13. Lewin JS, Connell CF, Duerk JL, Chung YC, Clampitt ME, Spisak J, et al. Interactive MRI-guided radiofrequency interstitial

DOI: 10.3109/02656736.2014.966337

14. 15. 16.

17.

Int J Hyperthermia Downloaded from informahealthcare.com by Yale Dermatologic Surgery on 01/05/15 For personal use only.

18.

19. 20.

21. 22. 23. 24.

25.

26.

27. 28. 29. 30. 31. 32. 33.

thermal ablation of abdominal tumors: Clinical trial for evaluation of safety and feasibility. J Magn Reson Imaging 1998;8:40–7. Hindman JC. Proton resonance shift of water in the gas and liquid state. J Chem Phys 1966;44:4582–92. Ishihara Y, Calderon A, Watanabe H, Okamoto K, Suzuki Y, Kuroda K, et al. Precise and fast temperature mapping using water proton chemical shift. Magn Reson Med 1995;34:814–23. Rempp H, Martirosian P, Boss A, Clasen S, Kickhefel A, Kraiger M, et al. MR temperature monitoring applying the proton resonance frequency method in liver and kidney at 0.2 and 1.5 T: Segment-specific attainable precision and breathing influence. Magn Reson Mater Phys 2008;21:333–43. Delannoy J, Chen CN, Turner R, Levin RL, Le Bihan D. Noninvasive temperature imaging using diffusion MRI. Magn Reson Med 1991;19:333–9. Peng HH, Huang TY, Tseng WY, Lin EL, Chung HW, Wu CC, et al. Simultaneous temperature and magnetization transfer (MT) monitoring during high-intensity focused ultrasound (HIFU) treatment: Preliminary investigation on ex vivo porcine muscle. J Magn Reson Imaging 2009;30:596–605. Rieke V, Pauly KB. MR thermometry. J Magn Reson Imaging 2008;27:376–90. Mori K, Fukuda K, Asaoka H, Ueda T, Kunimatsu A, Okamoto Y, et al. Radiofrequency ablation of the liver: Determination of ablative margin at MR imaging with impaired clearance of ferucarbotran – Feasibility study. Radiology 2009;251:557–65. Limanond P, Zimmerman P, Raman SS, Kadell BM, Lu DS. Interpretation of CT and MRI after radiofrequency ablation of hepatic malignancies. Am J Roentgenol 2003;181:1635–40. Geddes LA, Baker LE. The specific resistance of biological material: A compendium of data for the biomedical engineer and physiologist. Med Biol Eng 1967;5:271–93. Grimnes S, Martinsen OG. Bioimpedance and Bioelectricity Basics. London: Academic Press, 2011. Pop M, Molckovsky A, Chin L, Kolios MC, Jewett MA, Sherar MD. Changes in dielectric properties at 460 kHz of kidney and fat during heating: Importance for radio-frequency thermal therapy. Phys Med Biol 2003;48:2509–25. Haemmerich D, Schutt DJ, Wright AS, Webster JG, Mahvi DM. Electrical conductivity measurement of excised human metastatic liver tumours before and after thermal ablation. Physiol Meas 2009; 30:459–66. Goldberg SN, Gazelle GS, Dawson SL, Rittman WJ, Mueller PR, Rosenthal DI. Tissue ablation with radiofrequency: Effect of probe size, gauge, duration, and temperature on lesion volume. Acad Radiol 1995;2:399–404. Woo EJ, Seo JK. Magnetic resonance electrical impedance tomography (MREIT) for high-resolution conductivity imaging. Physiol Meas 2008;29:R1–26. Seo JK, Woo EJ. Magnetic resonance electrical impedance tomography (MREIT). SIAM Review 2011;53:40–68. Seo JK, Woo EJ. Electrical tissue property imaging at low frequency using MREIT. IEEE Trans Biomed Eng 2014;61:1390–9. Joy MLG, Scott GC, Henkelman RM. In vivo detection of applied electric currents by magnetic resonance imaging. Magn Reson Imag 1989;7:89–94. Scott GC, Joy MLG, Armstrong RL, Henkelman RM. Measurement of nonuniform current density by magnetic resonance. IEEE Trans Med Imag 1991;10:362–74. Seo JK, Yoon JR, Woo EJ, Kwon O. Reconstruction of conductivity and current density images using only one component of magnetic field measurements. IEEE Trans Biomed Eng 2003;50:1121–4. Oh SH, Lee BI, Woo EJ, Lee SY, Cho MH, Kwon O, et al. Conductivity and current density image reconstruction using harmonic Bz algorithm in magnetic resonance electrical impedance tomography. Phys Med Biol 2003;48:3101–16.

RF ablation monitoring using MREIT

455

34. Ider YZ, Birgul O. Use of the magnetic field generated by the internal distribution of injected currents for electrical impedance tomography (MR-EIT). Elektrik 1998;6:215–25. 35. Joy MLG. MR current density and conductivity imaging: The state of the art. Proc 26th Ann Int Conf IEEE EMBS, San Francisco, CA, USA 2004:5315–19. 36. Hamamura M, Muftuler L, Birgul O, Nalcioglu O. Measurement of ion diffusion using magnetic resonance electrical impedance tomography. Phys Med Biol 2006;51:2753–62. 37. Park C, Lee BI, Kwon O, Woo EJ. Measurement of induced magnetic flux density using injection current nonlinear encoding (ICNE) in MREIT. Physiol Meas 2006;28:117–27. 38. Haacke EM, Petropoulos LS, Nilges EW, Wu DH. Extraction of conductivity and permittivity using magnetic resonance imaging. Phys Med Biol 1991;36:723–33. 39. Katscher U, Voigt T, Findeklee C, Vernickel P, Nehrke K, Do¨ssel O. Determination of electrical conductivity and local SAR via B1 mapping. IEEE Trans Med Imag 2009;28:1365–74. 40. Webster JG. Electrical Impedance Tomography. Bristol: Adam Hilger, 1990. 41. Holder D. Electrical Impedance Tomography: Methods, History and Applications. Bristol: IOP, 2005. 42. Oh TI, Kim HJ, Jeong WC, Chauhan M, Kwon OI, Woo EJ. Detection of temperature distribution via recovering electrical conductivity in MREIT. Phys Med Biol 2013;58:2697–711. 43. Scott GC, Joy MLG, Armstrong RL, Hankelman RM. Sensitivity of magnetic resonance current density imaging. J Magn Reson 1992; 97:235–54. 44. Sadleir R, Grant S, Zhang SU, Lee BI, Pyo HC, Oh SH, et al. Noise analysis in MREIT at 3 and 11 Tesla field strength. Physiol Meas 2005;26:875–84. 45. Minhas AS, Jeong WC, Kim YT, Han Y, Kim HJ, Woo EJ. Experimental performance evaluation of multi-echo ICNE pulse sequence in magnetic resonance electrical impedance tomography. Magn Reson Med 2011;66:957–65. 46. Nam HS, Kwon OI. Optimization of multiply acquired magnetic flux density Bz using ICNE-Multiecho train in MREIT. Phys Med Biol 2010;55:2743–59. 47. Kwon O, Woo EJ, Yoon JR, Seo JK. Magnetic resonance electrical impedance tomography (MREIT): simulation study of J-substitution algorithm. IEEE Trans Biomed Eng 2002;49:160–7. 48. Kim YT, Yoo PJ, Oh TI, Woo EJ. Magnetic flux density measurement in magnetic resonance electrical impedance tomography using a low-noise current source. Meas Sci Technol 2011;22: 1–9. 49. Oh TI, Jeong WC, McEwan A, Park HM, Kim HJ, Kwon OI, et al. Feasibility of magnetic resonance electrical impedance tomography (MREIT) conductivity imaging to evaluate brain abscess lesion: In vivo canine model. J Magn Reson Imaging 2012;38:189–97. 50. Senneville BD, Quesson B, Moonen CTW. Magnetic resonance temperature imaging. Int J Hyperthermia 2005;21:515–31. 51. Terraz S, Cernicanu A, Lepetit-Coiffe M, Viallon M, Salomir R, Mentha G, et al. Radiofrequency ablation of small liver malignancies under magnetic resonance guidance: Progress in targeting and preliminary observations with temperature monitoring. Eur Radiol 2010;20:886–97. 52. Kim HJ, Jeong WC, Sajib SZ, Kim MO, Kwon OI, Woo EJ, Kim DH. Simultaneous imaging of dual-frequency electrical conductivity using a combination of MREIT and MREPT. Magn Reson Med 2014;71:200–8. 53. Kim Y, Rhim H, Cho OK, Koh BH, Kim Y. Intrahepatic recurrence after percutaneous radiofrequency ablation of hepatocellular carcinoma: Analysis of the pattern and risk factors. Eur J Radiol 2006; 59:432–41.