Stiffness distribution in the ablated zone after radiofrequency ablation for liver: An ex-vivo study with a tissue elastometer

Abstract

OBJECTIVE:

To investigate the stiffness distribution in the ablated zone after radiofrequency ablation (RFA), we used a device called tissue elastometer based on gross liver samples.

MATERIALS:

AND METHODS: Twelve freshly excised porcine livers were subject to RFA under a same setup to form elliptic ablated samples. Each sample was cut open for gross examination, and then the surface of the section plane was sliced into one piece for Young’s modulus test using the tissue elastometer. Five test points along the long- and short-axis on each piece were selected to evaluate stiffness distribution respectively. Among them, four points distributed equidistantly from center to boundary in the ablated zone and one was in the unablated zone.

RESULTS:

In the ablated zone, we found the Young’s moduli were significantly different among the four test points both in long- (F = 99.04, p <0.001) and short-axis (F = 79.47, p <0.001) directions. The Young’s modulus showed a downtrend in each direction, and was linearly related to the distance from the center to the test point (for long axis, R² = 0.968; for short axis, R² = 0.984, both p <0.001). A more significant downtrend was observed in short-axis direction. The Young’s moduli gained from the inner edge of ablated zone were comparable and significantly higher than those from the outer edge for both directions. The maximum value of 24.71kPa for Young’s modulus was the appropriate threshold to ensure the tissues were necrotic completely.

CONCLUSION:

The stiffness inside the ablated zone represented a radial distribution with downtrend, following a linear law. The stiffness at the inner edge of ablated zone is stable and significantly higher than that at the outer edge. The maximum value of 24.71 kPa close to the inner edge of Wz may be used as the standard of complete ablation.

Keywords

Stiffness distribution radiofrequency ablation tissue elastometer Young’s modulus

1 Introduction

In recent years, thermal ablation including radiofrequency ablation (RFA) and microwave ablation (MWA) offers a minimally invasive alternative way for treating solid tumors such as liver malignancy [1 –3]. All these ablations are aimed to necrotize the cancerous and some surrounding normal tissues. A safety margin of 5 mm or greater in width is necessary to cover the whole tumor [4]. Otherwise, it is highly possible to leave untreated viable tumor tissue around the periablational zone, which may result in subsequent local tumor progression [5]. Therefore, it is imperative to monitor the volume of necrosis and evaluate the treatment response as accurate as possible [6].

Ideally, once any residual tumors are detected, it is more reasonable to perform complementary ablations during one single ablation session. However, current contrast enhanced imaging modalities including contrast-enhanced ultrasound (CEUS), CT and MRI cannot provide accurate feedback of the extent and range of tissue damage during ablation immediately [7 –10]. Namely, if necessary, complementary ablation has to be delayed into another ablation session, resulting in additional medical cost and damage to the patient. In addition, administration of contrast agent also increases medical cost, and some patients might be contraindicative to contrast agent due to allergic reaction, kidney function failure or others [11]. Therefore, it is particularly important to develop new method to visualize tissue damage real-timely during ablation.

Over the last decade, ultrasound (US) elastography, a method for imaging the stiffness properties of tissue has been widely studied with an aim to visualize the necrotic zone [12]. In theory, the heat leads to protein denaturation and tissue dehydration, changing the stiffness of the ablated tissue [13]. Alternatively, by revealing the differences of stiffness between the ablated and unablated zones, it may be a useful way to monitor the treatment during ablation.

However, it was reported that the size of ablation zone depicted by US elastography was underestimated comparing to gross pathology [14 –16]. The misinterpretation of the ablation boundary might have serious consequences, like incomplete ablation. On the other hand, current studies with US elastography were carried out using US machines from different manufacturers, giving different elasticity thresholds to predict necrosis [16 –18]. Moreover, the trend of stiffness from the center of the ablated zone to the periphery has not been investigated comprehensively. In many elastograms, the ablation zone usually presents a mottled appearance [19, 20], so that it is unable to reflect the inner stiffness distribution accurately.

In this study, we attempted to map the stiffness distribution and determinate the stiffness margin between the ablated and normal tissues using a special third-party device. Furthermore, the stiffness value of the necrotic threshold will be investigated to provide a gold-standard reference for US elastography techniques from different manufacturers. Moreover, the stiffness changing trends inside the ablated tissues in different directions were also evaluated.

2 Materials and methods

2.1 Liver samples

Freshly excised porcine livers were obtained from a local slaughterhouse for this study. The interval between the death of each porcine and the beginning of the sample analysis was less than 6 hours. Each liver sample was cut into a cube roughly about 15 cm×10 cm×5 cm (width×length×thickness) in size and placed into a plastic dish. In order to fix the liver cube, additional liver tissues were poured into the same dish and around the sample. Ablation was performed in a single section of each sample.

2.2 Ablation procedure

A bipolar radiofrequency ablation system (Celon AG Medical Instruments, Teltow, Germany), including a 15 cm-long, 14-gauge bipolar electrode with an active length of 3 cm, was employed to perform ablation procedures. This electrode was set to provide an initial output power of 40 W. During ablation, the output power would keep decreasing slowly along with the increase of the impedance of ablated liver to avoid carbonization. To increase the ablation efficiency, an internal liquid circulation was conducted at room temperature for device cooling. And the deliver rate was set to 30 ml/min using normal saline solution.

Before inserting the electrode, The MyLab Touch US system (Esaote, Milan, Italy) which was equipped with a linear transducer (SL1543, frequency range: 4–13 MHz) was applied to scan the liver and find an appropriate section for ablation. The section was at least 5 cm in thickness of liver parenchyma and without any large vessels and bile ducts around. Subsequently, the electrode was inserted laterally in parallel with the US transducer, and placed into the center of above-mentioned section under US guidance. One image was taken before starting ablating and then every 5 min until the total energy achieved 30 kJ. After ablation procedure, each sample was cut in half along the long axis of the electrodes.

2.3 Measurement

In this study, the Young’s modulus (in kPa) was used to represent the stiffness value and obtained using a device called Tissue Elastometer (TE) TE-v2.0 (ARTANN Laboratories, Inc.) (Fig. 1). This device is specifically designed to measure the Young’s modulus of soft tissue samples based on stress vs. strain dependence [21].

Fig.1

Tissue Elastometer TE-v2.0 (white arrow) and its control computer with installed specialized operational software (red arrow).

Before stiffness measuring, gross examination on sample was performed as soon as cutting open the sample. Macroscopically, the ablated areas presented as a “white zone” (Wz) [22], which were subsequently photographed and measured. Then, the surface of each sample’s section plane including the whole Wz and surrounding unablated tissue was sliced into one piece carefully for further test (Fig. 2). According to the instructions of the TE device, the thickness of each piece was limited to no thicker than 3 mm to ensure the precision of stiffness measurement.

Fig.2

Tissue Elastometer TE-v2.0 is testing Young’s modulus for one piece of sample (white arrow).

The Wz observed by gross pathological examination indicates lethal damage to tissue [23]. With this knowledge, in each slice, five fixed points (marked as y0 to y4) along with the long-axis of Wz were selected to obtain the values of Young’s modulus (Fig. 3). Among them, the point y4 was positioned at the center of the Wz, y1 at the inner edge of Wz, y3 and y2 at the inner and outer tertile points between y4 and y1. Additionally, the point y0 was positioned at the outer edge of Wz. Similarly, another five points (marked as x0 to x4) along with the short-axis of Wz were selected according to above-mentioned plans for measurement (Fig. 3). The distance from each test point to the center of Wz was also measured.

Fig.3

The distribution of test points.

Each point was tested for seven trials to take the average. A personal computer was used to control and record the measurement procedure. When all measurements were done, each slice was preserved in 10% formalin solution for histological analysis.

2.4 Statistical analysis

Statistical analyses were performed using SPSS 20.0 (SPSS, Inc, Chicago, IL, USA). The mean Young’s moduli of all test points and their respective standard deviations were calculated. Along one of the two directions, the differences in Young’s modulus among 4 test points in Wz were evaluated by one-way analysis of variance (ANOVA) test. When differences among them were found to be statistically significant (p < 0.05), each Young’s modulus of the test point was compared with every other one using least significant difference (LSD) test. Kolmogorov-Smirnov test was applied for evaluation of normal distribution of the Young’s modulus of each test point. Univariate linear regression analysis was used to determine the association of Young’s moduli with the distance within the range of Wz. Finally, unpaired t-test was used to analyze the differences in Young’s modulus between two testing points which were close to the edge of Wz (y0 vs. y1, x0 vs. x1, y0 vs. x0, y1 vs. x1, respectively).

3 Results

A total of 12 ablations were performed on 12 samples. However, only 8 complete sample slices were obtained. Breakages occurred in other 4 slices when being sliced (the number of samples was 3, 5, 7, 8, respectively), and subsequently significant biases were observed when testing Young’s modulus along short-axis. Therefore, a total of 12 and 8 sets of data along long and short-axis respectively were enrolled for analysis. Histological analysis demonstrated that the cells in Wz were markedly abnormal, illustrating irreversible cellular damage.

Table 1 shows the average Young’s modulus and distance of each point. Due to a fixed power and time setting of ablation, all the Wz were with a comparatively similar size, about 4.8 cm long and 3.0 cm wide. Therefore, the distance of each testing point to the center of Wz were nearly the same. Thus, we set that all corresponding test points were at the same location and use the average distance when performing univariate linear regression analysis.

One-way ANOVA showed significant differences in Young’s moduli among the 4 points in long (F = 99.04, p < 0.001) and short-axis (F = 79.47, p < 0.001) directions respectively. Further analysis with LSD showed that between every two adjacent testing points, the Young’s modulus was significantly higher for the inner one. In the aspect of boundary, the difference of Young’s moduli between the points x1 vs. y1 (p = 0.177), and x0 vs. y0 (p = 0.390) were comparable. However, significantly difference were observed between x1 vs. x0 (p = 0.013), and y1 vs. y0 (p = 0.016). In order to ensure tissue meets necrosis standards, we selected the maximum value of 24.71 kPa as the threshold standing for the necrosis instead of the mean value. It should be mentioned that this most strictly threshold value may raise the identification standard to ensure necrosis.

Linear regression analyses were undertaken with distance as the independent factor and Young’s modulus as the dependent variable. The correlation coefficient between Young’s modulus and distance in both directions had a high level of statistical significance (R2 were in the range 0.89∼0.99) for each sample (Fig. 4). And the correlations could be expressed by the following linear regression equations in (Table 2).

Table 1
The average Young’s modulus and distance

Direction Test point Average Young’s modulus (kPa) Average distance (mm)

Long-axis y4 165.18±33.32 0.12±0.01

y3 102.01±29.44 8.03±0.08

y2 47.93±11.02 15.8±0.11

y1 14.35±3.26 23.5±0.11

y0 2.98±0.96 23.7±0.12

Short-axis x4 170.72±38.64 0.11±0.01

x3 106.18±16.41 5.12±0.08

x2 40.43±13.78 10.1±0.13

x1 14.38±5.26 15.2±0.10

x0 3.84±1.27 15.6±0.10

Direction	Test point	Average Young’s modulus (kPa)	Average distance (mm)
Long-axis	y4	165.18±33.32	0.12±0.01
	y3	102.01±29.44	8.03±0.08
	y2	47.93±11.02	15.8±0.11
	y1	14.35±3.26	23.5±0.11
	y0	2.98±0.96	23.7±0.12
Short-axis	x4	170.72±38.64	0.11±0.01
	x3	106.18±16.41	5.12±0.08
	x2	40.43±13.78	10.1±0.13
	x1	14.38±5.26	15.2±0.10
	x0	3.84±1.27	15.6±0.10

Table 2

Linear regression equations of each sample for long and short-axis directions

The number of samples	Long-axis direction		Short-axis direction
	linear equations	R²	linear equations	R²
1	y =–9.10x + 222.79	0.992	y =–8.74x + 143.38	0.994
2	y =–5.76x + 138.79	0.934	y =–7.95x + 123.05	0.957
3	y =–5.69x + 141.40	0.984	/	/
4	y =–4.64x + 119.89	0.981	y =–14.68x + 206.3	0.901
5	y =–7.67x + 177.68	0.916	/	/
6	y =–6.15x + 147.41	0.972	y =–9.52x + 148.21	0.886
7	y =–7.48x + 178.41	0.967	/	/
8	y =–7.50x + 188.38	0.986	/	/
9	y =–5.96x + 147.09	0.974	y =–9.46x + 155.38	0.981
10	y =–4.69x + 113.16	0.972	y =–13.43x + 191.33	0.888
11	y =–8.10x + 206.42	0.987	y =–12.21x + 192.9	0.995
12	y =–5.20x + 134.25	0.986	y =–9.08x + 152.20	0.965
Average	y =–6.51x + 159.55	0.984	y =–10.63x + 164.08	0.968

Furthermore, based on the linear equations, we found a sharper slope in short-axis direction, illustrating a more significant downtrend of Young’s modulus (Fig. 5).

Fig.4

The pots representing Young’s moduli measured from test points in long- (A) and short-axis (B) directions were marked and connected for each sample respectively.

Fig.5

Correlation between the average Young’s modulus and distance. The red line indicates their correlation in long-axis, and the blue line indicates that in the short-axis.

4 Discussion

An imaging modality that provides real-time visual feedbacks of necrosis could potentially benefit the thermal ablation procedure [24]. Thus, US elastography may help visually monitor the evolution of stiffness changes caused during thermal ablation. Before applying it, the key is to find quantitative stiffness data of tissue and their quantitative relation to pathology. However, the scarcity of reliable experimental data on tissue elasticity limits the clinical use of elasticity imaging. One of the reasons is the absence of an adequate way to directly measure the property of stiffness on fresh samples.

In previous studies, many attempts had been made to indirectly determine a universal threshold that predicted the formation of necrosis. In some of them, the selection of necrosis boundary usually relied on conventional US or reconstructed elastography imaging. For instance, in Eyerly et al’s study using US-based elastography techniques to monitor tissue stiffness at ablation sites, the necrosis area was defined as: where stiffness increased 75% of the postablation images or equivalent to two times the stiffness prior to the ablation [25]. Unfortunately, this definition of necrosis was not based on actual pathology results. Whether the boundary of the stiffened area in image could be used to detect the threshold of necrosis was questionable.

For some other studies, gross pathology images were registered to elastography imaging for finding an accurate boundary. However, the registration of ablation dimension between elastography imaging and gross pathology was not performed side by side. Mariani et al. performed “semiautomatic rigid registration” using a homemade software program and rebuilt new elastographic maps to meet these registrations [18]. They selected some integer values to test, and finally recommended a threshold of 20 kPa with the best compromise between sensitivity (0.8) and positive predictive value (0.83) to predict necrosis. However, this threshold was also based on indirect measurements in elastography imaging. The accordance of elastography imaging and gross pathology was confirmed only by the value of area size, while neglecting the differences of shape. And this threshold was an optimal value merely among some integer values which were selected artificially. So were DeWall et al., they tried to visualize radiofrequency and microwave ablation zones using electrode vibration elastography. When comparing elastography imaging and gross pathology, the registration was performed by the method of “maximizing the similarity” [19]. During those procedures, errors were resulted from manual operation and visual inspection.

On the other hand, it is well known that necrosis occurs when temperature is greater than 46°C. According to the setting of our ablation protocol, the temperature inside the Wz should be high enough to cause tissue necrosis. Contrary to the uniformity of pathological status, based on our results, the stiffness distribution inside Wz was observed to decrease in a linear fashion from the center to the boundary, and like a concentric ellipse. DeWall et al. believed that this phenomenon could occur because of the way of thermal deposition of RFA [20]. The ablated region was formed by electrical (ablation core) and thermal conduction. In this way, the temperature radially decreased from the ablation core. For each point inside the Wz, the local temperature undoubtedly was related to its distance to the ablation core. Therefore, similar to the temperature gradients, the stiffness also represented a radial distribution of a downtrend. So in theory, concentric elliptical should be the final form of the stiffness distribution for RFA. As a comparison, in Zhang et al’s high intensity focused ultrasound (HIFU) ablation study [26], and Au et al’s irreversible electroporation ablation (IRE) study [27], a more uniform stiffness was respectively observed. For HIFU, it is probably because the high intensity insonification forms a relative even distribution of temperature [28]. For IRE, it inducted cell membrane permeability via an electric field to cause cell death without temperature increase. Why and how temperature affects the stiffness distribution should be further studied.

With respect to the visualized elastogram of RFA, unfortunately, it usually presented as an irregularly appearance which was far from a concentric ellipse, and the boundary of the ablated region mainly manifested an indented pattern. It may be attributed to that the resolution of stiffness is not high enough to be distinguished. Moreover, the needle or its tracks may also influence indirect estimation [15]. Clinically, the heat sink effect [29] may provide more variability to the elastogram. Our direct results without considering above-mentioned variability demonstrate that the measured Young’s modulus values on both sides of the ablation boundary were stable and statistically different. However, comparing to other tissues [30], the difference of stiffness between the two sides of the ablation boundary may be too tiny to be differentiated using elastogram at now stage. Along with the underestimation of ablation zone comparing to gross pathology, we believe the current level of elastogram may not reflect the ablation region accurately.

The main strength of this study is the direct measurement protocol. Thanks to the TE device, we investigated the property of stiffness directly on fresh gross sample, and collected Young’s modulus from each target region of interest. By this way, the Young’s modulus was strictly registered to the gross pathology sample. Second, we investigated the internal stiffness distribution of the ablated region, and found the value of stiffness decreased from the center to the edge in a linear law.

There are several limitations of this study. Firstly, this study was performed in vitro, thus the biologic cell death caused by apoptosis in transition zone was failed to be included [10, 31]. As we know, delayed thermal effect may further contribute to a larger ablation region. Secondly, we only performed RFA according to one single parameter setting. Whether adjusting these parameters could influence the stiffness distribution was not clear. Third, we had not measured the tissue stiffness using US machines before using TE. A further direct comparison between US machine and physical measurements device should be conducted.

5 Conclusion

The theoretical stiffness distribution of ablated region may be a concentric ellipse, with a downtrend from the inside out in a linear law. Additionally, a more obvious downtrend was observed along the short axis. The values of Young’s moduli measured close to the inner edge of ablation boundary are stable and significantly higher than those measured close to the outer edge. The maximum value of 24.71 kPa close to the inner edge of Wz may be used as the standard of complete ablation.

Footnotes

Acknowledgments

Supported in part by Grants SHDC22015005 and 16CR3061B from Shanghai Hospital Development Center, Grant 14441900900 from Science and Technology Commission of Shanghai Municipality, and National Natural Scientific Foundation of China (Grants 81371570, 81671695 and 81725008).

References

Ahmed

, Solbiati

, Brace

, et al. Image-guided tumor ablation: Standardization of terminology and reporting criteria–a 10-year update. Radiology. 2014;273(1):241–60.

, Xu

, Lu

, et al. Treatment efficacy and safety of ultrasound-guided percutaneous bipolar radiofrequency ablation for benign thyroid nodules. Br J Radiol. 2016;89(1059):20150858.

Bruix

, Reig

, Sherman

. Evidence-based diagnosis, staging, and treatment of patients with hepatocellular carcinoma. Gastroenterology. 2016;150(4):835–53.

Nakazawa

, Kokubu

, Shibuya

, et al. Radiofrequency ablation of hepatocellular carcinoma: Correlation between local tumor progression after ablation and ablative margin. AJR Am J Roentgenol. 2007;188(2):480–8.

Horiike

, Iuchi

, Ninomiya

, et al. Influencing factors for recurrence of hepatocellular carcinoma treated with radiofrequency ablation. Oncol Rep. 2002;9(5):1059–62.

Solbiati

, Ierace

, Tonolini

, et al. Radiofrequency thermal ablation of hepatic metastases. Eur J Ultrasound. 2001;13(2):149–58.

Bruenn

, Beyer

, Haimerl

, et al. Comparison of computed tomography (CT) and contrast-enhanced ultrasound (CEUS) for the quantitative evaluation of an ablation defect following radiofrequency ablation of malignant liver lesions. Clin Hemorheol Microcirc. 2017;67(3-4):445–51.

Goldberg

, Walovitch

, Straub

, et al. Radio-frequency-induced coagulation necrosis in rabbits: Immediate detection at US with a synthetic microsphere contrast agent. Radiology. 1999;213(2):438–44.

, Yu

, Liang

, et al. Contrast-enhanced ultrasound in the characterization of hepatocellular carcinomas treated by ablation: Comparison with contrast-enhanced magnetic resonance imaging. Ultrasound Med Biol. 2013;39(9):1571–9.

10.

, Wilkins

, Ziats

, et al. Real-time monitoring of radiofrequency ablation and postablation assessment: Accuracy of contrast-enhanced US in experimental rat liver model. Radiology. 2014;270(1):107–16.

11.

van der Molen

, Reimer

, Dekkers

, et al. Post-contrast acute kidney injury. Part 2: Risk stratification, role of hydration and other prophylactic measures, patients taking metformin and chronic dialysis patients: Recommendations for updated ESUR Contrast Medium Safety Committee guidelines. Eur Radiol. 2018;28(7):2856–69.

12.

Varghese

, Techavipoo

, Liu

, et al. Elastographic measurement of the area and volume of thermal lesions resulting from radiofrequency ablation: Pathologic correlation. AJR Am J Roentgenol. 2003;181(3):701–7.

13.

Bharat

, Techavipoo

, Kiss

, et al. Monitoring stiffness changes in lesions after radiofrequency ablation at different temperatures and durations of ablation. Ultrasound Med Biol. 2005;31(3):415–22.

14.

Correa-Gallego

, Karkar

, Monette

, et al. Intraoperative ultrasound and tissue elastography measurements do not predict the size of hepatic microwave ablations. Acad Radiol. 2014;21(1):72–8.

15.

DeWall

, Varghese

, Brace

Quantifying local stiffness variations in radiofrequency ablations with dynamic indentation. IEEE Trans Biomed Eng. 2012;59(3):728–35.

16.

Wiggermann

, Brünn

, Rennert

, et al. Monitoring during hepatic radiofrequency ablation (RFA): Comparison of real-time ultrasound elastography (RTE) and contrast-enhanced ultrasound (CEUS): First clinical results of 25 patients. Ultraschall Med. 2013;34(6):590–4.

17.

, Li

, Xu

, et al. 2D shear-wave ultrasound elastography (SWE) evaluation of ablation zone following radiofrequency ablation of liver lesions: Is it more accurate. Br J Radiol. 2016;89(1060):20150852.

18.

Mariani

, Kwiecinski

, Pernot

, et al. Real time shear waves elastography monitoring of thermal ablation: In vivo evaluation in pig livers. J Surg Res. 2014;188(1):37–43.

19.

Dewall

, Varghese

, Brace

Visualizing ex vivo radiofrequency and microwave ablation zones using electrode vibration elastography. Med Phys. 2012;39(11):6692–700.

20.

DeWall

, Varghese

, Madsen

Shear wave velocity imaging using transient electrode perturbation: Phantom and ex vivo validation. IEEE Trans Med Imaging. 2011;30(3):666–78.

21.

Egorov

, Tsyuryupa

, Kanilo

, et al. Soft tissue elastometer. Med Eng Phys. 2008;30(2):206–12.

22.

Goldberg

, Gazelle

, Compton

, et al. Treatment of intrahepatic malignancy with radiofrequency ablation: Radiologic-Pathologic correlation. Cancer. 2000;88(11):2452–63.

23.

Gemeinhardt

, Poch

, Hiebl

, et al. Comparison of bipolar radiofrequency ablation zones in an in vivo porcine model: Correlation of histology and gross pathological findings. Clin Hemorheol Microcirc. 2016;64(3):491–9.

24.

NPB

, Beyer

, Hottenrott

, et al. Efficiency of contrast enhanced ultrasound for immediate assessment of ablation status after intraoperative radiofrequency ablation of hepatic malignancies. Clin Hemorheol Microcirc. 2017;66(4):357–68.

25.

Eyerly

, Vejdani-Jahromi

, Dumont

, et al. The Evolution of tissue stiffness at radiofrequency ablation sites during lesion formation and in the peri-ablation period. J Cardiovasc Electrophysiol. 2015;26(9):1009–18.

26.

Zhang

, Zhang

, Wan

, et al. A fast tissue stiffness-dependent elastography for HIFU-induced lesions inspection. Ultrasonics. 2011;51(8):857–69.

27.

, Kingham

, Jun

, et al. Irreversible electroporation ablation of the liver can be detected with ultrasound B-mode and elastography. Surgery. 2013;153(6):787–93.

28.

Zhang

, Castaneda

, Christensen

, et al. Real-time sonoelastography of hepatic thermal lesions in a swine model. Med Phys. 2008;35(9):4132–41.

29.

Frericks

, Ritz

, Albrecht

, et al. Influence of intrahepatic vessels on volume and shape of percutaneous thermal ablation zones: In vivo evaluation in a porcine model. Invest Radiol. 2008;43(4):211–8.

30.

Chen

, Novakofski

, Jenkins

, et al. Young’s modulus measurements of soft tissues with application to elasticity imaging. IEEE Transactions on Ultrasonics Ferroelectrics & Frequency Control. 1996;43(1):191–4.

31.

Wiersinga

, Jansen

, Straatsburg

, et al. Lesion progression with time and the effect of vascular occlusion following radiofrequency ablation of the liver. Br J Surg. 2003;90(3):306–12.