Direction of Tissue Contraction after Microwave Ablation: A Comparative Experimental Study in Ex Vivo Bovine Liver

Abstract

Objective

This study aimed to investigate the direction of tissue contraction after microwave ablation in ex vivo bovine liver models.

Materials and Methods

Ablation procedures were conducted in a total of 90 sites in ex vivo bovine liver models, including the surface (n = 60) and parenchyma (n = 30), to examine the direction of contraction of the tissue in the peripheral and central regions from the microwave antenna. Three commercially available 2.45-GHz microwave systems (Emprint, Neuwave, and Surblate) were used. For surface ablation, the lengths of two overlapped square markers were measured after 2.5- and 5-minutes ablations (n = 10 ablations for each system for each ablation time). For parenchyma ablation, seven predetermined distances between the markers were measured on the cutting plane after 5- and 10-minutes ablations (n = 5 ablations for each system for each ablation time). The contraction in the radial and longitudinal directions and the sphericity index (SI) of the ablation zones were compared between the three systems using analysis of variance.

Results

In the surface ablation experiment, the mean longitudinal contraction ratio and SI from a 5-minutes ablation using the Emprint, Neuwave, and Surblate systems were 28.92% and 1.04, 20.10% and 0.53, and 24.90% and 0.45, respectively (p < 0.001). A positive correlation between longitudinal contraction and SI was noted, and a similar radial contraction was observed. In the parenchyma ablation experiment, the mean longitudinal contraction ratio and SI from a 10-minutes ablation using the three pieces of equipment were 38.60% and 1.06, 32.45% and 0.61, and 28.50% and 0.50, respectively (p < 0.001). There was a significant difference in the longitudinal contraction properties, whereas there was no significant difference in the radial contraction properties.

Conclusion

The degree of longitudinal contraction showed significant differences depending on the microwave ablation equipment, which may affect the SI of the ablation zone.

INTRODUCTION

Tissue contraction is clinically significant, particularly for image-guided thermal therapies, such as microwave ablation (MWA). Any changes in tissue dimensions should be considered when planning and assessing treatment [1]. At high temperatures, tissue contraction is positively correlated with local tissue dehydration, protein denaturation, and collagen contraction [2]. Accordingly, the post-treatment ablation zone may underestimate the extent of the original tissue included within the coagulation zone [3].

Many recent studies have described tissue contraction during thermal ablation [2,3,4,5,6,7,8,9]. Such a phenomenon may affect the assessment of the therapeutic efficacy and safety of unwanted heating of any abutting vital structures [5]. Therefore, the phenomenon of tissue contraction should be considered for accurate predictions and assessments of ablation outcomes [10]. Farina et al. [9] reported that a substantial asymmetric contraction of the ablated tissue volume as well as an initial expansion phenomenon occurs during MWA. Subsequently, they evaluated the extent of tissue contraction using three commercial MWA devices [5]. In the radial direction, similar trends in contraction were observed at the margin of the thermally ablated areas, whereas in the longitudinal direction, one device induced different contraction kinetics with respect to the other two [5]. However, in the experimental setup, the relatively small cubes of ex vivo liver in a variety of sizes embedded in the agar phantom restricted the measurements of the ablation dimensions.

Liu et al. [4] evaluated tissue contraction by measuring the displacement of a fiducial marker using computed tomography imaging during the ablation of the ex vivo liver. They discovered that the strength of contraction in the radial direction transverse to the antenna was relatively stronger than that in the longitudinal direction of the antennal axis. Thus, ablations appeared to be more elongated than the initial tissue because of the strong trend of contraction in the transverse plane.

The abovementioned studies [4,5,9] focused on tissue contraction in the radial direction from the antennal axis and reported a trend of relatively low contraction in the longitudinal direction. Most MWA devices appear to inflict an ellipsoidal ablation with low predictability with respect to size [11,12,13]. Recently, a new MWA device that can create a more spherical ablation zone was introduced in the clinical setting [14,15]. However, the mechanism underlying MWA has not been clearly evaluated based on the properties of tissue contraction.

This study aimed to objectively assess tissue contraction in the radial and longitudinal directions associated with MWA using three different systems in ex vivo bovine liver models and to elucidate the correlation between tissue contraction and sphericity of the ablation zone across the systems.

MATERIALS AND METHODS

MWA Equipment

Commercial MWA apparatuses operating at 2.45 GHz, the Emprint™ (Medtronic) with a 13G water-cooled antenna, NEUWAVE™ (J&J Medical Devices) with a CO₂-cooled 17G PR probe, and SurBlate™ (Vison Medical USA) with a 15G water-cooled antenna were used for this experimental study. In all experiments, the microwave power was set according to the manufacturer’s recommendations, namely 100 W for the Emprint, 65 W for the NEUWAVE, and 100 W for the SurBlate.

Preparation of Tissue Samples

Freshly excised normal bovine livers weighing an average of 7 kg each were obtained from a local slaughterhouse (Hyupjin Center) on the day of each experiment and were cut into rectangular blocks (13 × 11 × 8 cm [length × width × height]) at room temperature.

Two ex vivo Bovine Experimental Models

Weiss et al. [16] analyzed the planar strain of ex vivo liver during MWA and postulated that rapid enlargement within approximately 3 minutes occurs due to the thermal expansion of tissue surrounding the antenna, combined with an increase in the vapor pressure of its fluids. Once denaturation of the proteins and desiccation start to occur, the tissue begins to contract. This result led us to assign ablation times of 2.5 and 5 minutes for surface contraction and 5 and 10 minutes for parenchyma contraction because the contraction was assumed to occur after approximately 3 minutes of ablation. To observe the contraction of the central and peripheral regions from the antenna, the parenchyma and surface ablation experiments were conducted with the same 5-minutes ablation time.

Surface Ablation

The surface of the liver was pigmented with two overlapped squares of size 1.95 × 1.95 cm and 2.95 × 2.95 cm, respectively, using a grid frame and SEPIA MicroPigment (The Standard), which is an endoscopic marker used for polyps and lesions in the gastrointestinal tract. Then, an antenna was placed at the center of the square marker, so that the end of the antennal tip was located at the bottom of the square (Fig. 1). Given that the ablation diameter in the ex vivo experiment was approximately 2–5 cm, depending on each device [12], the square size was assigned to be similar to the ablation diameter. Using each MWA device, two different ablation times of 2.5 and 5 minutes were applied, and 10 ablation zones were investigated at each ablation time.

Fig. 1. Schema of surface ablation.

The outer square (gray/2.95 × 2.95 cm) and inner square (white/1.95 × 1.95 cm) markers were overlapped on the surface of the liver. The left and right sides of the squares are the indices of contraction in the longitudinal direction parallel to the antennal axis, whereas the top and bottom sides of the squares are the indices of contraction in the radial direction orthogonal to the antennal axis. H1 and H2 = index of contraction in the radial direction in the outer square (gray), h1 and h2 = index of contraction in the radial direction in the inner square (white), V1 and V2 = index of contraction in the longitudinal direction in the outer square (gray), v1 and v2 = index of contraction in the longitudinal direction in the inner square (white).

The lengths of the eight sides of the two overlapped square markers were measured (Fig. 1), and a total of 160 data points were acquired from each device. The lengths of the top and bottom sides of the two squares were assigned as the contraction in the radial direction orthogonal to the antennal axis, whereas the lengths of the left and right sides of the two squares were assigned as the contraction in the longitudinal direction parallel to the antennal axis. The contraction ratio was estimated from the mean value of the contraction length in the radial and longitudinal directions. The ablation zone size was also estimated by measuring the radial axis diameter (RAD) perpendicular to the antennal axis and the longitudinal axis diameter (LAD) along the antennal axis using ImageJ 1.42 (http://rsbweb.nih.gov/ij/download.html). The sphericity index (SI) of the ablation zones was calculated using the following formula: SI = RAD²/LAD² (Fig. 2).

Fig. 2. Images after a 5-minutes ablation on the surface of the bovine liver.

The shape of the ablation zone and tissue contraction can be seen in both the radial (gray) and longitudinal (white) directions. LAD = longitudinal axis diameter, RAD = radial axis diameter

Parenchyma Ablation

The parenchyma of the liver was internally pigmented with ten markers using a specially designed fixture with ten needles (19G), each with a sheath (18G). The intervals between the pigmented markers were as follows: H1, 1 cm; H2, 1 cm; H3, 1 cm; H4, 2 cm; H5, 3 cm; V1, 2 cm; and V2, 2 cm (Fig. 3). After puncturing the liver directly with the ten needles, the ten sheaths were left behind by pulling out only the needles, after which ten rods (21G) dipped into the MicroPigment were placed into each sheath. The ten sheaths were then pulled out, while the ten dyed rods were left behind. Finally, an antenna was inserted into the liver, and the ten rods were pulled down. From each MWA equipment, two different ablation times of 5 and 10 minutes were used, and five ablation zones were investigated at each ablation time.

Fig. 3. Schema of parenchyma ablation.

A. Inserted antenna after dipping into the pigmented markers. B. Magnified image of the dotted circle from (A) and the antennal position. The interval lengths of H1, H2, H3, H4, H5, V1, V2 are 1, 1, 1, 2, 3, 2, and 2 cm, respectively. C. Specific fixture in which dot markers and tools, such as the biopsy needle comprising a needle, sheath, and rod for pigmenting were used. D. 1) puncturing by assembling the sheath and needle fixture, 2) moving out the needle fixture, 3) inserting the dyed rods inside the sheaths after inserting the antenna, 4) pulling down the dyed rods to pigment the marker.

The lengths of the seven intervals from the ten pigmented markers on the cutting plane along the antennal axis were measured (Fig. 4), and 70 data points were acquired for each MWA device. The interval length from H1 to H5 was assigned as the contraction in the radial direction orthogonal to the antennal axis, whereas the interval lengths of V1 and V2 were assigned as the contraction in the longitudinal direction parallel to the antennal axis. The contraction ratio was estimated from the mean value of the contraction length in the radial and longitudinal directions. The ablation zone size (RAD, LAD) was estimated, and the SI was calculated. The ablation volume was estimated using the formula for an ellipsoid: V = (π/6) × RAD² × LAD.

Fig. 4. Images after a 10-minutes ablation in the parenchyma of the bovine liver.

Ten pigmented markers indicated with yellow arrows are displayed on the cut section.

Statistical Analysis

Differences in the length of tissue contraction between the three pieces of equipment were analyzed using one-way analysis of variance (R ver. 3.4.4 with RStudio ver. 1.1.442). The Shapiro–Wilk test for normality checks and Bartlett’s test for homogeneity of variances were performed, and the assumptions were met. Post hoc comparisons using Tukey’s honestly significant difference (HSD) test were performed. Statistical significance was set at p < 0.05.

RESULTS

Surface Ablation

For liver surface ablation using each piece of equipment, the tissue contraction ratios with respect to the directions and SI are shown in Table 1. Remarkably, the ablation shape of the Surblate extended out of the marker area along the antennal axis (Fig. 2), which caused inaccurate measurements of the contraction. For the 5-minutes ablation and inner square condition, the mean longitudinal contraction ratio and SI were 28.92% and 1.04, respectively, for the Emprint system, and 20.10% and 0.53, respectively, for the Neuwave system. The Emprint and Neuwave systems showed similar trends in that the SI was correlated with the contraction ratio in the longitudinal direction. The Emprint and Neuwave systems revealed a similarity in contraction in the radial direction and indicated that a low SI is expected when the discrepancy between the contraction in the radial direction and contraction in the longitudinal direction was high. In the 2.5-minutes ablation experiment, no thermal expansion-related phenomena were observed.

Table 1. Comparison of Contraction and the SI between Surface Ablations.

			Emprint (n = 10)	Neuwave (n = 10)	Surblate (n = 10)	P *
2.5 minutes ablation time
	Radial tissue contraction
		Inner square-length, cm	1.42 ± 0.13	1.40 ± 0.12	1.54 ± 0.16	0.004‡§
		Inner square-contraction ratio, %	27.18 ± 6.55	28.26 ± 6.22	21.13 ± 8.16
		Outer square-length, cm	2.59 ± 0.17	2.59 ± 0.15	2.64 ± 0.20	0.643
		Outer square-contraction ratio, %	12.12 ± 5.58	12.20 ± 5.14	10.66 ± 6.64
	Longitudinal tissue contraction
		Inner square-length, cm	1.56 ± 0.09	1.58 ± 0.13	1.65 ± 0.08	0.004§
		Inner square-contraction ratio, %	19.95 ± 4.66	19.10 ± 6.89	15.36 ± 3.89
		Outer square-length, cm	2.64 ± 0.12	2.66 ± 0.12	2.61 ± 0.10	0.370
		Outer square-contraction ratio, %	10.39 ± 3.98	9.96 ± 4.16	11.41 ± 3.29
	Ablation zone
		RAD, cm	2.81 ± 0.14	2.21 ± 0.10	2.04 ± 0.15	< 0.001†‡§
		LAD, cm	2.80 ± 0.10	3.06 ± 0.15	3.48 ± 0.19	< 0.001†‡§
		SI	1.01 ± 0.12	0.52 ± 0.06	0.34 ± 0.07	< 0.001†‡§
5 minutes ablation time
	Radial tissue contraction
		Inner square-length, cm	1.24 ± 0.12	1.28 ± 0.09	1.41 ± 0.19	0.007‡§
		Inner square-contraction ratio, %	36.36 ± 6.11	34.56 ± 4.51	27.64 ± 9.79
		Outer square-length, cm	2.30 ± 0.22	2.37 ± 0.14	2.43 ± 0.25	0.136
		Outer square-contraction ratio, %	22.08 ± 7.60	19.53 ± 4.88	17.53 ± 8.39
	Longitudinal tissue contraction
		Inner square-length, cm	1.39 ± 0.15	1.56 ± 0.14	1.47 ± 0.13	0.001†
		Inner square-contraction ratio, %	28.92 ± 7.54	20.10 ± 7.27	24.90 ± 6.52
		Outer square-length, cm	2.35 ± 0.19	2.55 ± 0.20	2.34 ± 0.15	< 0.001†‡
		Outer square-contraction ratio, %	20.24 ± 6.35	13.42 ± 6.68	20.58 ± 5.03
	Ablation zone
		RAD, cm	3.41 ± 0.20	2.72 ± 0.12	2.63 ± 0.15	< 0.001†‡§
		LAD, cm	3.35 ± 0.17	3.73 ± 0.12	3.92 ± 0.17	< 0.001†‡§
		SI	1.04 ± 0.09	0.53 ± 0.07	0.45 ± 0.04	< 0.001†‡§

Radial tissue contraction inner square-length (h1, h2; 1.95 cm). Radial tissue contraction Outer square-Length (H1, H2; 2.95 cm). Longitudinal tissue contraction inner square-length (v1, v2; 1.95 cm). Longitudinal tissue contraction Outer square-Length (V1, V2; 2.95 cm). Data are mean ± standard deviation. ^*One-way analysis of variance test (differences statistically not significant if p > 0.05), ^†‡§Indicate significant pairwise comparisons according to the Tukey’s HSD post hoc test: ^†Emprint vs. Neuwave, ^‡Neuwave vs. Surblate, ^§Surblate vs. Emprint. LAD = longitudinal axis diameter, RAD = radial axis diameter, SI = sphericity index

Parenchyma Ablation

For parenchymal ablation of the liver with each MWA system, the tissue contraction ratio according to the direction and SI are shown in Table 2. Because the interval length from H1 to H5, regarded as the contraction in the radial direction, was not significantly different in the analysis of each system, the same intervals H1, H2, and H3 were selected to investigate the contraction in the radial direction. In the 10-minutes ablation condition, the mean longitudinal contraction ratio and SI of the three systems were 38.60% and 1.06, 32.45% and 0.61, and 28.50% and 0.50 for the Emprint, Neuwave, and Surblate, respectively. A similar radial contraction was observed, and a low SI was expected when the discrepancy between the radial and longitudinal contraction ratios was high (Fig. 5). There was a significant difference in longitudinal contraction (p < 0.001), whereas there was no significant difference in radial contraction.

Table 2. Comparison of Contraction and the SI between Parenchyma Ablations.

			Emprint (n = 5)	Neuwave (n = 5)	Surblate (n = 5)	P *
5 minutes ablation time
	Radial tissue contraction
		Length, cm	0.68 ± 0.09	0.68 ± 0.09	0.67 ± 0.08	0.924
		Contraction ratio, %	32.13 ± 8.64	32.26 ± 8.88	33.27 ± 7.98
	Longitudinal tissue contraction
		Length, cm	1.22 ± 0.13	1.40 ± 0.11	1.51 ± 0.18	< 0.001†§
		Contraction ratio, %	39.20 ± 6.63	30.05 ± 5.37	24.5 ± 8.99
	Ablation zone
		RAD, cm	3.11 ± 0.04	2.69 ± 0.16	2.71 ± 0.16	< 0.001†§
		LAD, cm	3.15 ± 0.08	3.81 ± 0.18	4.32 ± 0.37	< 0.001†‡§
		SI	0.97 ± 0.04	0.50 ± 0.07	0.39 ± 0.04	< 0.001†‡§
10 minutes ablation time
	Radial tissue contraction
		Length, cm	0.64 ± 0.09	0.65 ± 0.07	0.66 ± 0.11	0.828
		Contraction ratio, %	36.40 ± 8.80	34.87 ± 7.41	34.40 ± 11.29
	Longitudinal tissue contraction
		Length, cm	1.23 ± 0.05	1.35 ± 0.11	1.43 ± 0.09	< 0.001†§
		Contraction ratio, %	38.60 ± 2.69	32.45 ± 5.61	28.50 ± 4.65
	Ablation zone
		RAD, cm	4.18 ± 0.15	3.51 ± 0.08	3.38 ± 0.11	< 0.001†§
		LAD, cm	4.06 ± 0.15	4.51 ± 0.19	4.78 ± 0.17	< 0.001†§
		SI	1.06 ± 0.10	0.61 ± 0.06	0.50 ± 0.06	< 0.001†§

Radial tissue contraction Length (H1, H2, H3; 1 cm). Longitudinal tissue contraction Length (V1, V2; 2 cm). Data are mean ± standard deviation. ^*One-way analysis of variance test (differences statistically not significant if p > 0.05), ^†‡§Indicate significant pairwise comparisons according to the Tukey’s HSD post hoc test: ^†Emprint vs. Neuwave, ^‡Neuwave vs. Surblate, ^§Surblate vs. Emprint. LAD = longitudinal axis diameter, RAD = radial axis diameter, SI = sphericity index

Fig. 5. The Radial and Longitudinal compared with the SI.

The sphericity of each device was sustained in a similar manner, although the ablation time increased. The DIS of the three microwave ablation systems was inversely proportional to the sphericity. DIS = discrepancy between radial and longitudinal contraction ratio, Longitudinal = contraction ratio in the longitudinal direction, Radial = contraction ratio in the radial direction, SI = sphericity index

A trend in contraction with respect to direction in each of the three devices was observed (Fig. 6). The contraction levels in both directions increased with time. Contraction in the radial direction converged at a similar level over time. However, contraction in the longitudinal direction was sustained at different levels with each device. According to the 5-minutes ablation experimental contraction data, a difference in longitudinal contraction between the surface and parenchyma was observed, which indicated a trend of high contraction in the central region from the antenna.

Fig. 6. The contraction trend with respect to direction in each device.

The contraction levels in both directions increased with time. The Radial converged at a similar level with increasing time. The Longitudinal was sustained at different levels in each device. Longitudinal = contraction ratio in the longitudinal direction, Radial = contraction ratio in the radial direction, 1 = inner square/surface ablation (square mark), 2 = parenchyma ablation (circle mark)

The estimated mean ablation volumes for each device were 96.15 cm³ (Emprint), 71.29 cm³ (Neuwave), and 66.18 cm³ (Surblate).

DISCUSSION

Ablation is widely considered to be a safe and curative procedure. The use of MWA has increased in popularity owing to its better ablation performance at high temperatures and shorter ablation times [17]. However, a more refined ablation strategy is required to optimally ablate the index tumor, especially in patients with poor liver function. In that sense, a clear understanding of the ablation zone profile, including tissue contraction, is essential to achieve better outcomes with MWA.

The current experimental study demonstrated that the direction of tissue contraction varies depending on the commercial microwave equipment used, especially along the longitudinal axis. Although many previous studies have proven that radial contraction (shrinkage) is more prominent than longitudinal contraction, no study has demonstrated that the direction of contraction depends on the type of microwave equipment. The results of the present study show that the difference in the degree of contraction between the radial and longitudinal directions can affect the spherical index of the ablation zone. We can predict the real ablation zone and volume based on this difference in the contraction profile depending on the microwave equipment. This will be critical for achieving an optimal ablation volume with adequate ablation margins and avoiding unwanted thermal injury to adjacent vital organs [18,19].

In this study, we examined tissue contraction in the radial and longitudinal directions, and identified the characteristics of the ablation zone shape for each MWA device. The results showed that longitudinal contractions were different, whereas radial contractions were similar across the MWA devices. The spherical ablation zone shape had an almost equivalent tissue contraction ratio in both the radial and longitudinal directions. Each MWA device produced substantially different ablation volumes and shapes because each SI value was different and time independent.

The Emprint system has unique properties allowing it to reliably produce a spherical ablation zone of a particular shape by field and wavelength control [14]. The clinical performance of these controls has been compared with that of two other routinely used conventional MWA systems, revealing that a more spherical ablation zone and a larger minimal ablative margin can be obtained with field and wavelength control [15]. Furthermore, the antenna is a critical component that transfers microwave energy into the tissue to form an active heating zone [20], and its design is important for supplying efficient radiation into the surrounding tissue to maximize energy delivery and to control the radiation pattern to obtain the desired ablation zone geometry [11,21,22,23,24].

Several studies have evaluated tissue contraction in other models. Brace et al. [2] described tissue contraction in liver and lung tissue resulting from microwave and radiofrequency (RF) ablation, which can heat tissue to 100℃ or more, noting an approximately 25% reduction in ablation zone diameter. Sommer et al. [3] described a similar tissue shrinkage and dehydration caused by MWA, noting an approximately 30% underestimation of effective coagulation in the kidneys. Rossmann et al. [7] developed another model based on studies of tissue shrinkage during controlled temperature exposure from 60℃ to 95℃, and suggested 12.3%–21.7% shrinkage after 15 minutes of temperature exposure. Liu and Brace [4,8] described tissue contraction spatially through fiducial marker movement using CT imaging during MWA. From the localized displacement of markers, central tissue contracted by more than 60% and peripheral tissue contracted by only 15% in any direction. The authors simulated temperature-dependent contraction using numerical modeling. While these studies lacked analysis of longitudinal contraction and relied on comparing independent samples, the contraction rate of the present study is similar to that of previous studies and has been verified through comparative studies.

Farina et al. [9] and Weiss et al. [16] reported that thermal expansion was due to the presence of liquid vaporization at high temperatures. Both studies used a container in which the bovine liver was placed to assess the expansion phenomenon. In contrast, surface ablation experiments were conducted in open chambers unsuitable for observing the expansion phenomenon. It was speculated that the evaporation pressure naturally dissipated owing to the surface experimental conditions.

This study had several limitations. First, the current results were only based on ex vivo experimental studies, and should be demonstrated by in vivo and clinical studies. However, there are many technical hurdles to objectively measure tissue contraction using any marker in living animals or humans. Second, a more detailed assessment of tissue contraction in a wider area of the liver was not available because of technical difficulties with regard to the extensive placement of pigmented markers and limited experimental resources. The lack of markers that did not cover the entire ablation zone limited the detailed analysis of contraction in the ablation zone; in particular, the square grid marker in the surface ablation experiment did not cover the entire ablation zone in the data of one of the MWA systems. Therefore, we did not achieve statistically significant results in this study. Finally, further investigation using variable ablation parameters beyond the manufacturer’s recommendation is warranted to fully explore the characteristics of tissue contraction following MWA and to compare them with RF devices.

In conclusion, the degree of tissue contraction did not differ along the radial direction, and it varied along the longitudinal direction depending on the microwave equipment used in our ex vivo experimental model. A similar contraction profile in both directions may help create a more spherical ablation zone, which is ideal for the optimal configuration of the ablation zones and circumferential ablative margins.

Availability of Data and Material

The datasets generated or analyzed during the study are available from the corresponding author on reasonable request.